PSEUDO-RESPONSE REGULATORS 9, 7, and 5 Are Transcriptional Repressors in the Arabidopsis Circadian Clock W OA

An interlocking transcriptional-translational feedback loop of clock-associated genes is thought to be the central oscillator of the circadian clock in plants. TIMING OF CAB EXPRESSION1 (also called PSEUDO-RESPONSE REGULATOR1 [PRR1]) and two MYB transcription factors, CIRCADIAN CLOCK ASSOCIATED1 (CCA1) and LATE ELONGATED HYPOCOTYL (LHY), play pivotal roles in the loop. Genetic studies have suggested that PRR9, PRR7, and PRR5 also act within or close to the loop; however, their molecular functions remain unknown. Here, we demonstrate that PRR9, PRR7, and PRR5 act as transcriptional repressors of CCA1 and LHY . PRR9, PRR7, and PRR5 each suppress CCA1 and LHY promoter activities and confer transcriptional repressor activity to a heterologous DNA binding protein in a transient reporter assay. Using a glucocorticoid-induced PRR5-GR (glucorticoid receptor) construct, we found that PRR5 directly downregulates CCA1 and LHY expression. Furthermore, PRR9, PRR7, and PRR5 associate with the CCA1 and LHY promoters in vivo, coincident with the timing of decreased CCA1 and LHY expression. These results suggest that the repressor activities of PRR9, PRR7, and PRR5 on the CCA1 and LHY promoter regions constitute the molecular mechanism that accounts for the role of these proteins in the feedback loop of the circadian clock.


INTRODUCTION
The circadian clock controls endogenous biological rhythms that allow a wide range of organisms to adapt to 24-h day-night cycles (Young and Kay, 2001). In eukaryotes, a transcriptiontranslation feedback loop connecting clock-associated genes is thought to form the central oscillator (or core) of the clock (Bell-Pedersen et al., 2005). In Arabidopsis thaliana, reciprocal transcriptional regulation between TIMING OF CAB EXPRESSION1 (TOC1; also called PSEUDO-RESPONSE REGULATOR1 [PRR1]) and CIRCADIAN CLOCK ASSOCIATED1 (CCA1) and LATE ELONGATED HYPOCOTYL (LHY) has been proposed as the main feedback loop. CCA1 and LHY proteins are MYB transcription factors and repress TOC1 transcription by binding directly to the TOC1 promoter around dawn (Alabadi et al., 2001;Mizoguchi et al., 2002;Perales and Mas, 2007). TOC1 protein is expressed in the evening and, in turn, activates CCA1 expression, partly by antagonizing a transcriptional repressor of CCA1, CCA1 HIKING EXPEDITION (CHE) of the TCP (for TEOSINTE BRANCHED1, CYCLOIDEA, and PCF) family (Pruneda-Paz et al., 2009). In addition to TOC1, numerous genetic studies have dem-onstrated the importance of other PRR genes in the circadian clock (Eriksson et al., 2003;Kaczorowski and Quail, 2003;Michael et al., 2003;Yamamoto et al., 2003;Farre et al., 2005;Nakamichi et al., 2005;Salome and McClung, 2005;Para et al., 2007;Ito et al., 2009).
The Arabidopsis PRR gene family consists of five members (PRR9, PRR7, PRR5, PRR3, and TOC1), all of which are regulated by the circadian clock, but each of which peaks at different times of the day. PRR9 mRNA levels are greatest at dawn, PRR7 peaks in the morning, PRR5 around noon, and PRR3 and TOC1 in the evening (Matsushika et al., 2000). Not surprisingly, the PRR proteins also peak at different times of the day (Mas et al., 2003;Farre and Kay, 2007;Ito et al., 2007;Kiba et al., 2007;Fujiwara et al., 2008). PRR9, PRR7, and PRR5 have a redundant function, but because they are expressed at different times of the day, they are collectively essential to proper timekeeping. This is demonstrated by the loss of rhythmicity in the prr9 prr7 prr5 triple lossof-function mutant (Nakamichi et al., 2005). CCA1 and LHY are constitutively expressed at a high level, and TOC1 is expressed at a low level in the triple mutant, suggesting that PRR9, PRR7, and PRR5 regulate the circadian clock by downregulating CCA1 and LHY expression and by upregulating TOC1 expression (Nakamichi et al., 2005). Closing the feedback loop, CCA1 and LHY activate transcription of PRR9 and PRR7 by directly binding to their promoters (Farre et al., 2005). Although this regulatory framework connecting PRR9, PRR7, and PRR5 with CCA1, LHY, and TOC1 has been proposed (Farre et al., 2005;Nakamichi et al., 2005;Niwa et al., 2007;Ito et al., 2008Ito et al., , 2009, the molecular mechanism of this interaction remains undetermined. PRR proteins feature a pseudoreceiver domain at the N terminus and a CCT (for CONSTANS, CONSTANS-LIKE, and TOC1) motif at the C terminus Strayer et al., 2000). The pseudoreceiver domain is similar to the receiver domain of a two-component response regulator, but the key Asp residue that accepts a phosphoryl group from a sensory kinase is not conserved . The CCT motif is thought to be involved in protein-protein interactions (Wenkel et al., 2006). Although PRR proteins are localized in the nucleus, their exact molecular function is still unknown Strayer et al., 2000;Matsushika et al., 2007b;Fujiwara et al., 2008).
In this study, we describe the transcriptional repressor activity of PRR9, PRR7, and PRR5. In addition, PRR9, PRR7, and PRR5 associate with the promoter regions of CCA1 and LHY in vivo, coincident with the timing of decreased CCA1 and LHY expression. These results suggest that PRR9, PRR7, and PRR5 proteins are major transcriptional repressors of CCA1 and LHY and are thus essential for proper clock function.

PRR9, PRR7, and PRR5 Downregulate CCA1 and LHY in Arabidopsis Seedlings
To examine whether PRR9, PRR7, and PRR5 regulate CCA1, LHY, and TOC1, at the transcriptional level, we conducted transient assays using reporter plasmids harboring luciferase (LUC) under the control of the CCA1, LHY, or TOC1 promoters (CCA1pro:LUC, LHYpro:LUC, or TOC1pro:LUC) and effector plasmids harboring PRR9, PRR7, or PRR5 genes fused to a gene encoding cyan fluorescent protein (CFP) or a negative control containing CFP only, all under the control of the cauliflower mosaic virus (CaMV) 35S promoter (35Spro:PRR9-CFP, PRR7-CFP, or CFP). We confirmed that the CCA1, LHY, and TOC1 promoters generate rhythmic patterns of luciferase activity in the appropriate phases (see Supplemental Figure  1 online). Reporter and effector plasmids were codelivered into cells of Arabidopsis seedlings by particle bombardment, and relative bioluminescence of the LUC reporter was measured.  Figure 1C), suggesting that PRR9, PRR7, and PRR5 each downregulate promoter activity of CCA1 and LHY but act differently toward TOC1.
CCA1 and LHY Are Immediate Targets of PRR5 To determine whether PRR9, PRR7, and PRR5 directly regulate CCA1, LHY, and TOC1 expression, we generated transgenic plants expressing a chimeric fusion protein of PRR5, the hormone binding domain of mouse glucocorticoid receptor (GR), and CFP under the control of the CaMV 35S promoter (35Spro:PRR5-GR-CFP). Such a GR fusion protein becomes biologically functional in the presence of the glucocorticoid steroid hormone dexamethasone (DEX) in plants (Aoyama et al., 1995;Sablowski and Meyerowitz, 1998). This approach is also used to identify the direct target genes of a GR fusion protein (Sablowski and Meyerowitz, 1998). Transgenic Arabidopsis seedlings overexpressing PRR5 produce shorter hypocotyls than the wild type (Sato et al., 2002). Therefore, we first tested whether PRR5-GR-CFP protein becomes biologically active in the presence of DEX by measuring hypocotyl lengths of PRR5-GR-CFP-overexpressing plants (35Spro:5GC). Two independent 35Spro:5GC lines had significantly shorter hypocotyls under DEX-treated conditions than under DEX-free conditions (Figure 2A), confirming that the PRR5-GR-CFP protein is biologically functional with DEX treatment.
CCA1, LHY, and TOC1 expression in 35Spro:5GC plants grown in constant white light conditions for 2 weeks after germination was measured by quantitative PCR (qPCR) with or without a 4-h incubation in DEX. DEX treatment resulted in a 50% decrease in CCA1 and LHY expression but a 20% increase in TOC1 expression ( Figure 2B). To examine whether PRR5 regulates CCA1, LHY, and TOC1 through de novo biosynthesis of an intermediary protein, 35Spro:5GC plants were cotreated with DEX and the translational inhibitor cycloheximide (CHX). DEX

of 12
The Plant Cell combined with CHX resulted in a 20 to 30% decrease in CCA1 and LHY expression, compared with CHX treatment alone (Figure 2C). However, cotreatment with DEX and CHX did not cause any significant change in TOC1 expression, compared with CHX treatment alone ( Figure 2C, bottom). These results suggest that PRR5 downregulates CCA1 and LHY without de novo biosynthesis of other proteins but upregulates TOC1 indirectly.
To clarify which specific region of these PRRs was responsible for the repressor activity, we introduced into Arabidopsis seedlings several effector plasmids harboring a series of truncated PRR5 clones fused to GAL4DB ( Figure 3B, left). We found that effector plasmids covering the CCT domain (#2) as well as the C-terminal part of the intervening region between pseudoreceiver and CCT (#5) did not cause significant reductions in GAL4pro: LUC reporter activity. However, seedlings expressing effector plasmids that covered a 44-amino acid peptide from the intervening region (#1, 3, 4, and 6) showed reduced reporter activity by about half ( Figure 3B), indicating that this region is sufficient to confer repressor activity on GAL4DB. Comparison of this 44amino acid region with those of PRR7, PRR9, and putative orthologs from other plant species lead to the identification of two relatively conserved motifs, L( E / D )( L / I )S( L / I )( R / K )R and SXXSAF( S / T )( R / Q )( Y / F ). The SXXSAF( S / T )( R / Q )( Y / F ) motif was found in PRR3 and TOC1, whereas L( E / D )( L / I )S( L / I )( R / K )R was not (data not shown). In addition, L( E / D )( L / I )S( L / I )( R / K )R resembles the ETH-YLENE RESPONSE FACTOR (ERF)-associated amphiphilic repression (EAR) motif [( L / F )DLN( L / F )XP] (Ohta et al., 2001).
To determine which motif is responsible for the repressor activity, we generated an effector plasmid expressing the 22amino acid region harboring LDLSLRR, but not SSASAFTRY (plasmid #7), and another effector plasmid expressing the 25amino acid region harboring SSASAFTRY, but not LDLSLRR (plasmid #8). Bombardment with these effectors did not result in significant reduction of GAL4pro:LUC reporter activity compared with GAL4DB alone ( Figure 3D), indicating that the 44-amino acid region harboring both of the motifs is required for the repressor activity of PRR5. Furthermore, effector plasmids that covered corresponding regions from PRR7 and PRR9 (GAL4DB-PRR7#6 and GAL4-DB-PRR9#6) showed reduced reporter activity by about half ( Figure 3E), suggesting that regions harboring both L( E / D )( L / I )S( L / I )( R / K )R and SXXSAF( S / T )( R / Q )( Y / F ) in PRR9, PRR7, and PRR5 are critical for the repression activity.
To assess whether recruitment to DNA is required for the repression activity of PRR5, GAL4DB-PRR5 or PRR5 was coexpressed with the GAL4pro:LUC reporter. Expression of GAL4DB-PRR5 reduced GAL4pro:LUC activity, whereas PRR5 or CFP did not ( Figure 3F). This result suggests that recruitment of PRR5 to a certain DNA region is sufficient to repress transcription from the DNA region.
Amplicons 2, 3, 6, and 7 were significantly enriched in 9pro: F9G plants in comparison to the other amplicons ( Figure 4B). ChIP analyses for 7pro:F7G and 5pro:F5G also indicated a significant enrichment of amplicons 2, 3, 6, and 7 (Figured 4C and 4D). As expected, no significant enrichment was found on any amplicons in the control lines 9pro:FGG, 7pro:FGG, and 5pro:FGG, showing that PRR9, PRR7, and PRR5 specifically associate with CCA1 and LHY promoter regions in vivo. To further test the specificity of the association, protein-DNA complexes from prr5-11 (Yamamoto et al., 2003) and 35S:PRR5 (Sato et al., 2002) seedlings were used in a ChIP analysis, in which nontagged PRR5 protein was immunoprecipitated with anti-PRR5 antibody ( Figure 4E). The same set of amplicons was enriched in 35Spro:PRR5 but not in prr5-11. These results show that PRR9, PRR7, and PRR5 are able to associate with the promoter regions of CCA1 and LHY in vivo. Although several known cis-acting elements (G-box, TBS, and 5A) exist within these regions, the resolution of our ChIP analyses was too low to specify exactly the binding site of PRR9, PRR7, and PRR5. Since PRR9, PRR7, and PRR5 expression is mostly limited to finite but partially overlapping portions of the diurnal cycle, we investigated the levels of association between these proteins and their target promoter regions by analyzing 9pro:F9G, 7pro: (E) GAL4 promoter activity when coexpressed with the GAL4DB fused to corresponding regions in PRR7 and PRR9 described in Figure 3C. (F) GAL4 promoter activity determination after coexpression of PRR5 fulllength or GAL4DB-PRR5. Error bars indicate SD (biological replicates, n = 6). Asterisks indicate values that are statistically different from control (t test; P < 0.05). Each experiment was performed twice with similar results.

of 12
The Plant Cell F7G, and 5pro:F5G plants grown under LD cycles. The plants were harvested at 2-h intervals from ZT0. F9G protein levels peaked from ZT2 to ZT6 ( Figure 5A, left panel), and the protein became associated with CCA1 and LHY promoters during the same timeframe ( Figure 5B, left panels). F7G protein peaked at around ZT8 to ZT12, but its association with CCA1 and LHY promoters was biphasic, with peaks at ZT6 and ZT14 ( Figures 5A and 5B, middle panels). F5G protein both peaked and became associated with the CCA1 and LHY promoters between ZT8 to ZT16 ( Figures 5A and 5B, right panels). These results suggest a correlation between the timing of PRR9, PRR7, and PRR5 expression and the level of association between each PRR protein with the CCA1 and LHY promoters. Furthermore, our results show that, from ZT2 to ZT16 (morning to midnight), at least one of these proteins is present on the CCA1 and LHY promoters. We next measured CCA1 and LHY mRNA levels to quantify the effect of PRR9, PRR7, and PRR5 on CCA1 and LHY transcription in vivo. CCA1 and LHY mRNA expression decreased from ZT0 to ZT6, remained at a low level until ZT18, and started to increase at ZT20 ( Figure 5C). As shown above, PRR9, PRR7, and PRR5 were expressed and associated with the promoter regions of CCA1 and LHY from ZT2 to ZT16 (Figures 5A and 5B). Therefore, the association patterns of PRR9, PRR7, and PRR5 with the CCA1 and LHY promoters are inversely correlated with the expression of CCA1 and LHY, suggesting that PRR9, PRR7, and PRR5 repress CCA1 and LHY expression in vivo.
Sequential Expression of PRR9, PRR7, and PRR5 Is Required for Repression of CCA1 and LHY from Morning until Midnight To understand the significance of the sequential expression of PRR9, PRR7, and PRR5 for regulation of CCA1 and LHY transcription, we analyzed the expression patterns of PRR9 and PRR5 proteins and also monitored CCA1 and LHY levels in prr7 prr5 and prr9 prr7 double mutants (Farre et al., 2005;Nakamichi et al., 2005;Salome and McClung, 2005). For detection of native PRR9, we developed an anti-PRR9 antibody and confirmed its specificity (see Supplemental Figure 4 online).
In prr7 prr5 mutants, PRR9 protein accumulated from ZT2 to ZT10, and CCA1 and LHY mRNA expression was minimal from ZT4 to ZT10 ( Figure 6B). However, at ZT14, PRR9 protein levels were undetectable, and 2 h later (ZT16), CCA1 and LHY expression began to increase. This inverse correlation between PRR9 levels and CCA1 and LHY transcription suggests that PRR9 represses CCA1 and LHY to a certain level in the prr7 prr5 mutant but is not sufficient to repress their expression during nighttime.
In the prr9 prr7 double mutant, PRR5 protein accumulated from ZT10 to ZT18, and CCA1 and LHY expression decreased from ZT12 to ZT18 ( Figure 6C). PRR5 protein was not detectable either from ZT0 to ZT8 or from ZT20 to ZT22. CCA1 and LHY expression levels were relatively high from ZT0 to ZT10 and from ZT20 to ZT22 ( Figure 6C). These patterns suggest that PRR5 alone can account for the repression of CCA1 and LHY transcription during part of the night, but, without PRR9 and PRR7, this effect is not extended to the daytime. Collectively, the sequential expression of PRR9, PRR7, and PRR5 is important to propagate a proper waveform, long-duration repression state of CCA1 and LHY expression.

DISCUSSION
Transcriptional Repressor Activities of PRR9, PRR7, and PRR5 In this study, we identified PRR9, PRR7, and PRR5 as active transcriptional repressors of CCA1 and LHY. An active FLAG-PRR5-GFP (5pro:F5G) seedlings (E) ChIP assays for prr5-11 and 35Spro:PRR5 etiolated seedlings. Fiveday-old etiolated seedlings exposed to white light for 10 h were used. Immunoprecipitation was performed with anti-PRR5 antibody. Error bars indicate SD (technical replicates, n = 3). Each experiment was performed twice with similar results.
Active Repressors in Circadian Clock 5 of 12 transcriptional repressor generally refers to a repressor that contains intrinsic repression domains and has the ability to inhibit transcription via the action of these domains (Ohta et al., 2001). In fact, we found that in PRR9, PRR7, and PRR5, a conserved amino acid sequence between the pseudoreceiver domain and CCT motif ( Figure 3C) harboring both L( E / D )( L / I )S( L / I )( R / K )R and SXXSAF( S / T )( R / Q )( Y / F ) motifs is sufficient for repressor activity ( Figure 3E). Interestingly, the L( E / D )( L / I )S( L / I )( R / K )R sequence resembles the EAR motif [( L / F )DLN( L / F )XP] (Ohta et al., 2001), which is conserved in ERF transcriptional repressors. Under our ex-perimental conditions, however, LDLSLRR of PRR5 alone could not confer repression activity ( Figure 3D), suggesting that flanking sequences of LDLSLRR are also required. Although the PRR proteins lack typical DNA binding domains, and no direct interaction between PRR9, PRR7, or PRR5 and upstream regions of CCA1 or LHY was observed in a yeast onehybrid system (N. Nakamichi, T. Kiba, and H. Sakakibara, unpublished data), our ChIP analysis indicated that PRR9, PRR7, and PRR5 proteins associate with promoter regions of CCA1 and LHY in vivo (Figure 4). We cannot exclude the possibility of direct (A) F9G, F7G, and F5G protein levels in 12-h-light/12-h-dark conditions. 9pro:F9G, 7pro:F7G, and 5pro:F5G plants were grown in 12-h-light/12-h-dark conditions for 2 weeks and cross-linked at 2-h intervals starting at ZT0 (light on). Total protein was immunoblotted by anti-FLAG antibody (top panel). F9G, F7G, and F5G protein amounts normalized with the total protein (bottom). Peak levels were set to 1.0. White and gray areas represent white light and dark conditions, respectively. (B) Percentages of the amplicon 3 of CCA1 region (top) and amplicon 7 LHY region (bottom) coimmunoprecipitated with anti-GFP antibody relative to input DNA in 9pro:F9G, 7pro:F7G, and 5pro:F5G plants were plotted. (C) CCA1 and LHY mRNA expression in 5pro:F5G plants. Expression levels were determined relative to APX3 mRNA. Peak levels were set to 1.0. Error bars indicate SD (technical replicates, n = 3). Each experiment was performed twice with similar results.  (Figures 4 and 5). Furthermore, there is no obvious epistatic interaction on LHY expression between PRR9 or PRR7 and TOC1 , implying that PRR9, PRR7, and PRR5 associate with the LHY promoter in some way other than in the CHE-TOC1 complex. Although the exact binding mechanism of PRR9, PRR7, and PRR5 to target DNA needs to be determined in future studies, some posttranslational regulation of PRR7 and PRR5 might be involved in the promoter binding mechanism, since protein peaks and promoter-association peaks were not exactly the same for PRR7 and PRR5 ( Figure 5). Matsushika et al. (2007b) showed that overexpression of the C-terminal region (containing both the intervening region and the CCT motif) of PRR5 causes downregulation of CCA1 but that neither the intervening nor the CCT motif alone could do so, suggesting that cooperation between the two of them is essential (Matsushika et al., 2007b). Since recruitment of PRR5 to the DNA molecule is required for repression ( Figure 3E), and the CCT motif is implicated in protein-protein interactions (Wenkel et al., 2006), PRR9, PRR7, and PRR5 may be recruited to the promoter region via such protein-protein interactions and subsequently repress target gene transcription with the repression motif. We cannot exclude the possibility of a molecular function other than transcriptional repression for PRR9, PRR7, and PRR5. However, constant high expression levels of CCA1 and LHY and many daypeaked genes in a prr9 prr7 prr5 triple mutant (Nakamichi et al., 2005 suggest that the repressor activity of these proteins is essential to rhythmic expression of their target genes.

PRR9, PRR7, and PRR5 Shape the Waveform of CCA1 and LHY Expression
Sequential expression of PRR9, PRR7, and PRR5 under diurnal conditions has attracted the attention of investigators since its discovery (Matsushika et al., 2000). Our time-course ChIP analysis demonstrated that transcriptional repressors PRR9, PRR7, and PRR5 continuously and sequentially associate with the promoters of CCA1 and LHY from morning till midnight (;16 h), Figure 6. Expression Patterns of PRR9 and PRR5 Proteins, and CCA1 and LHY Expression in prr7 prr5 and prr9 prr7 Mutants.
(A) The wild-type waveforms of PRR9, PRR7, and PRR5 protein levels and CCA1 and LHY mRNA expression. These data were also presented in Figures 5A and 5C. (B) PRR9 protein levels and CCA1 and LHY mRNA expression in the prr7 prr5 double mutant. The amount of PRR9 protein was normalized with total protein, and the amounts of CCA1 and LHY mRNA were normalized with APX3 mRNA (SD; technical replicates, n = 3). Each experiment was performed twice with similar results. (C) PRR5 protein levels and CCA1 and LHY mRNA expression in the prr9 prr7 mutant. The amount of PRR5 protein was normalized with total protein, and the amounts of CCA1 and LHY mRNA were normalized with APX3 mRNA (SD; technical replicates, n = 3). Each experiment was performed twice with similar results. In all panels, peak levels were set to 1.0, and white and gray areas represent white light and dark conditions, respectively.
Active Repressors in Circadian Clock 7 of 12 exactly the duration when CCA1 and LHY are repressed ( Figure  5). Expression of PRR9 and PRR5 coincides with CCA1 and LHY repression in the prr7 prr5 and prr9 prr7 mutants, respectively ( Figure 6). The duration of CCA1 and LHY repression is shorter (;6 h) in these prr double mutants than in the wild type, supporting the notion that sequential expression of PRR9, PRR7, and PRR5 is critical to maintain the repressed state of CCA1 and LHY. In addition, previous studies showed that CCA1 and LHY are expressed at constitutively high levels in the prr9 prr7 prr5 triple mutant (Nakamichi et al., 2005) but at low levels in PRR9-, PRR7-, or PRR5-overexpressing plants (Sato et al., 2002;Farre and Kay, 2007;Matsushika et al., 2007a). These lines of evidence lead us to propose that the sequential expression pattern of PRR9, PRR7, and PRR5 is required to propagate the proper waveforms of CCA1 and LHY.

PRR9, PRR7, and PRR5 in the Arabidopsis Circadian Clock
The central oscillator of the Arabidopsis circadian clock is thought to be a multiple gene transcriptional feedback loop (McClung, 2006). In this loop, CCA1 and LHY activate PRR9 and PRR7 expression, and PRR9 and PRR7 are negative regulators of CCA1 and LHY. This model is supported by both experimental and mathematical approaches, although the exact mechanisms were previously not clear (Farre et al., 2005;Nakamichi et al., 2005;Locke et al., 2006;Zeilinger et al., 2006). In addition, the position of PRR5 in the circadian clock has been a matter of conjecture. In this study, we demonstrated a molecular mechanism that closes the loop between PRR9 and PRR7, and CCA1 and LHY (Figure 7). Furthermore, our results show that PRR5 is also involved in the repression of CCA1 and LHY.
Since the Arabidopsis circadian clock regulates a large number of genes (output genes) (Harmer et al., 2000;Michael et al., 2008;Hazen et al., 2009), it is not surprising that the expression levels and patterns of these genes are altered in the prr9 prr7 prr5 triple mutant. This is the case of genes involved in output pathways, such as flowering time regulation, hypocotyl length regulation, cold stress response, and mitochondrial metabolism (the trichloroacetic acid cycle) that are drastically altered in prr9 prr7 prr5 plants Nakamichi et al., 2007Nakamichi et al., , 2009Fukushima et al., 2009). Our findings that PRR9, PRR7, and PRR5 repress CCA1 and LHY suggest that these proteins could regulate output genes via CCA1 and LHY expression. However, PRR9, PRR7, and PRR5 regulate the flowering pathway and metabolism in mitochondria in a CCA1-independent manner Fukushima et al., 2009), an indication that the molecular links from PRR9, PRR7, and PRR5 to output genes remain a matter of speculation. Therefore, the identification of direct targets of PRR9, PRR7, and PRR5 other than CCA1 and LHY is an exciting future challenge.

Plant Materials and Growth Conditions
Arabidopsis thaliana accession Columbia-0 (Col-0) was used as the wild type in this study. Seedlings were grown at 228C for 14d on Murashige and Skoog (MS) medium (Murashige and Skoog, 1962) supplemented with 2% sucrose under white light (80 to 100 mmolÁm 22 Ás 21 ) either in constant light or in a 12-h-light/12-h-dark cycle. To obtain etiolated seedlings, seeds were sown on MS plates without sucrose. After 2 d of darkness at 48C, seeds were exposed to white light (80 to 100 mmolÁm 22 Ás 21 ) for 5 h at 228C to enhance germination and then incubated for 5d in the dark at 228C. A T87 Arabidopsis cultured cell line was used for real-time bioluminescence assay of CCA1pro:LUC, LHYpro:LUC and TOC1pro:LUC .

Plasmid Construction
To generate reporter plasmids, the promoter regions of CCA1, LHY, and TOC1 were amplified with Prime Star DNA polymerase (Takara) from the Arabidopsis Col-0 genome (primer set for the CCA1 promoter, 59-CTGAAGCTTCATGCATGGTTAGCTTAGC-39 and 59-GGTTCCATG-GTCTCCATCACTAAGCTCCTCTAC-39 to amplify between positions 2854 and 0 relative to the translational start; for the LHY promoter, 59-CACTTAAGCTTCAGCCACTACAATATCACCAC-39 and 59-GGAAC-CATGGTAACAGGACCGGTGCAG-39 to amplify between positions 2909 and 0; for the TOC1 promoter, 59-GCTAAGCTTCCACTGATGATGAGAT-TAACCATC-39 and 59-GTCTTCCATGGACCTCAGCATCTTCATACCC-39 to amplify between positions 21360 to +570). Each promoter region was fused to the firefly LUC gene on modified pSP-luc+ (Promega) between the HindIII and NcoI sites as described previously . These promoters were sufficient to confer rhythmic LUC expression in T87 cells (see Supplemental Figure 1 online). The reporter plasmid in which LUC is transcribed under the GAL4 promoter and the reference plasmid were previously described (Fujimoto et al., 2000).

Transient Transfection Assay in Arabidopsis Seedlings
The reporter, effector, and reference plasmids were delivered to 2-weekold Arabidopsis Col-0 seedlings by particle bombardment (PDU-1000/ He; Bio-Rad) as described previously (Sakakibara et al., 2005). Promoter activity was determined by normalizing luciferase values to Renilla luciferase expressed under the control of the CaMV 35S promoter in reference plasmid pPTRL (Fujimoto et al., 2000). Bioluminescence was detected with a Mithras LB940 (Berthold) as described previously (Yamaguchi et al., 2008).

Real-Time Luciferase Assay
Protoplasts of Arabidopsis T87 cultured cells were transfected with a reporter plasmid by a polyethylene glycol-mediated method (Yamada et al., 2004). Cells were entrained to a coordinate circadian rhythm by incubation in constant dark conditions for 12 h and release into constant white light conditions (time 0). Bioluminescence of cells was measured with a real-time monitoring system (Kondo et al., 1993).

Measurement of Hypocotyl Length
Measurement of hypocotyl lengths of Arabidopsis seedlings under shortday conditions (10 h light/14 h dark) was described previously . Seedlings were grown on MS or MS plus 10 mM DEX (Sigma-Aldrich).

DEX and CHX Treatment
Two independent T3 transgenic seedlings expressing PRR5-GR-CFP (two independent lines, 35Spro:5GC1 and 35Spro:5GC2) were grown on MS (2% sucrose) under constant light conditions for 2 weeks and transferred either to a water control, to 10 mM DEX, to 100 mM CHX (Sigma-Aldrich), or to 100 mM CHX plus 10 mM DEX. After a 4-h incubation, plants were harvested and frozen in liquid nitrogen.

RNA Isolation, Reverse Transcription, and qPCR
For each sample, three to five seedlings were harvested, frozen in liquid nitrogen, and then ground. Total mRNA was extracted with the RNeasy plant mini kit (Qiagen). To synthesize cDNA, 1 mg of each RNA sample was reverse transcribed with SuperScript II (Invitrogen) and oligo (dT20) primer. Real-time qPCR was performed on an ABI PRISM 7000 system (Applied Biosystems) using SYBR Green Extaq II (Takara) and specific primers (see Supplemental Figure 3 online). Reaction conditions were 958C for 5 min followed by 40 cycles of 958C for 15 s and 608C for 31 s. Product sizes were verified by agarose gel electrophoresis and dissociation curve analysis. Amount of a PCR amplicon was quantified by an absolute quantification method using a calibration curve of corresponding DNA.

ChIP Assay
For the ChIP assay, 500 mg of plant sample was cross-linked in 20 mL of 1% formaldehyde solution under vacuum for 40 min. The reaction was stopped by washing twice with ice-cold 0.3 M glycine. Plants were then ground to powder in liquid nitrogen and lysed in 2 mL lysis buffer (50 mM Tris-HCl, pH 7.5, 100 mM NaCl, 0.1% Triton X-100, 50 mM 26S proteasome inhibitor MG132 [Sigma-Aldrich], and complete protease inhibitor cocktail tablets [Roche]). The chromatin complexes were isolated and sonicated, and 50 mL of the chromatin complexes were stored for use as input fractions.
Cross-linking of immunocomplexes or the input fraction was reversed by incubating at 658C overnight followed by digestion with 5 mL of Proteinase-K (#9033; Takara) to remove all proteins. DNA was purified by NucleoSpin (Macherey-Nagel) according to the manufacturer's protocol. The amount of each precipitated DNA and input DNA was determined by real-time PCR using specific primers (see Supplemental Figure 3 online).

PRR9 Antibody
Anti-PRR9 antibody was obtained according to the method of Kiba et al. (2007). A cDNA fragment encoding PRR9 amino acid residues 168 to 411 was cloned into the expression vector pQE30 (Qiagen), which adds an N-terminal 6xHis tag. Recombinant protein was expressed in Escherichia coli BL21 and purified using a nickel-nitrilotriacetic acid agarose column (Qiagen) and then was used to generate polyclonal anti-PRR9 antiserum of rabbit (Cocalico Biological). The antiserum was immunopurified using its cognate antigen immobilized to a nitrocellulose membrane. Anti-PRR9 antibody was confirmed by detecting native PRR9 protein in the wild type and FLAG-PRR9-GFP protein in PRR9pro:FLAG-PRR9-GFP-transformed prr9 (SALK_106072) (see Supplemental Figure 4 online).

Protein Gel Blot Analysis
To detect FLAG-PRR-GFP proteins ( Figure 5), 10 mL of the chromatin complexes was mixed in a 1:1 ratio (v/v) with 23 lithium Active Repressors in Circadian Clock 9 of 12 dodecyl sulfate sample (SM) buffer , boiled at 958C for 5 min, loaded in a Super Sep Ace 10 to ;20% gradient gel (Wako), and blotted onto an Immobilon-P membrane (Millipore). The membrane was incubated with monoclonal anti-FLAG antibody (F3165; Sigma-Aldrich). Goat anti-mouse IgG conjugated with alkaline phosphatase (170-6520; Bio-Rad) was used as the secondary antibody, and protein signals were detected using the NBT/BCIP system (Roche).
To detect native PRR5 and PRR9, frozen plant materials were ground to a fine powder and suspended in a 1:1 ratio (w/v) with 23 SM buffer and incubated for 5 min at 958C. Rabbit anti-PRR5 or anti-PRR9 antibodies were used for the primary antibody. Goat anti-rabbit IgG conjugated with alkaline phosphatase (170-6518; Bio-Rad) was used as the secondary antibody. Quantitation of immunodetected proteins was performed using ImageJ software (http://rsb.info.nih.gov/ij/).

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