Reciprocal Regulation Between TOC1 and LHY/CCA1 Within the Arabidopsis Circadian Clock

See allHide authors and affiliations

Science  03 Aug 2001:
Vol. 293, Issue 5531, pp. 880-883
DOI: 10.1126/science.1061320


The interactive regulation between clock genes is central for oscillator function. Here, we show interactions between theArabidopsis clock genes LATE ELONGATED HYPOCOTYL(LHY), CIRCADIAN CLOCK ASSOCIATED 1(CCA1), and TIMING OF CAB EXPRESSION 1(TOC1). The MYB transcription factors LHY and CCA1 negatively regulate TOC1 expression. We show that both proteins bind to a region in the TOC1 promoter that is critical for its clock regulation. Conversely, TOC1 appears to participate in the positive regulation of LHY andCCA1 expression. Our results indicate that these interactions form a loop critical for clock function inArabidopsis.

Circadian clocks represent a widespread endogenous mechanism that allows organisms to time different processes appropriately throughout the day-night cycle. Among the activities controlled by the circadian clock are the regulation of transcription in cyanobacteria, the rhythmic movement of leaves in plants, and more complex activities such as control of feeding behavior in flies and control of the sleep-wake cycle in humans. The basic molecular mechanisms underlying the generation of circadian rhythms have been significantly deciphered inSynechococcus elongatus, Neurospora crassa,Drosophila melanogaster, and the mouse (1). In these four model systems, the oscillator is based on transcriptional-translational negative feedback loops involving a group of clock genes. In Arabidopsis, three genes have been suggested as components of the oscillator: LHY,CCA1, and TOC1 (2). LHY andCCA1 encode highly conserved single-MYB transcription factors which, when expressed at high and constitutive levels, disrupt the normal functioning of the clock (3, 4). The third gene, TOC1, was initially identified as thetoc1-1 mutant in a screen for plants with altered period length of cab2::luc expression (5).toc1-1 is a short-period mutant with altered clock function through the entire life cycle of the plant. Moreover, the effect is independent of light quantity, suggesting a role for TOC1 in the core of the oscillator (6, 7). The recent cloning ofTOC1 has allowed us a more detailed characterization of its molecular phenotypes (7). Here, we investigate potential interactions between TOC1 and the MYB genesLHY and CCA1, because those interactions are likely to shape the Arabidopsis oscillator.

We investigated the TOC1expression patterns in lhy and CCA1-OX plants, which constitutively overexpress either LHY orCCA1 from the strong promoter CaMV 35S (Fig. 1, A and B). TOC1 mRNA oscillated with high amplitude in the wild-type parental lines in constant light (LL). The TOC1 mRNA level was constant in both lhy and CCA1-OX plants under the same light conditions, as expected for plants in which the oscillator function is mostly disrupted. The transcript level in these mutants was similar to or lower than the trough of wild-type expression. Therefore, high and constant expression levels of either LHY or CCA1result in low and constant levels of TOC1 transcript in LL. LHY and CCA1 appear to be negative regulators of TOC1. The fact that TOC1 transcript oscillates 12 hours out of phase with both the LHY and CCA1 transcripts in wild-type plants further supports this idea. The CCA1 protein has been found to cycle with little lag with respect to its transcript (4), indicating that maximum CCA1 level coincides with minimum TOC1 mRNA level, and vice versa.

Figure 1

TOC1 expression is strongly affected inlhy, CCA1-OX, and elf3-1 plants.lhy plants and the parental wild type (Ler) (A), CCA1-OX (line o34) and elf3-1plants, and the parental wild type (Col-0) [(B) and (C), respectively] were germinated and entrained to cycles of 12 hours light, 12 hours dark [LD (12, 12)] for 7 days and then released into constant light (LL). TOC1 RNA levels were analyzed by Northern blot. The dark and white bars at the bottom of each graph represent lights off and on, respectively. Representative data are shown.

elf3-1 plants become arrhythmic due to an arrest of its circadian clock in continuous light (8, 9). We investigated the TOC1 expression profiles in the elf3-1 mutant under LL conditions (Fig. 1C). Contrary to what we observed for the wild type, the transcript level remained elevated and unchanging in the mutant in free-running conditions. The arrhythmic phenotypes observed in elf3-1 and in lhy and CCA1-OXplants in LL are associated with different levels of TOC1transcript. Both the LHY and CCA1 mRNA levels were found to be reduced in elf3-1 plants in LL (3,10), consistent with the elevated TOC1 transcript levels in this mutant arising from a significant reduction in the amount of the negative regulators LHY and CCA1.

If LHY and CCA1 act directly to negatively regulate TOC1expression, they might bind to a functionally important region of theTOC1 promoter. Accordingly, we sought to identify regulatory regions of the TOC1 promoter by generating a series of deletions fused to the luciferase gene and analyzing expression of the reporter in Arabidopsis plants transformed with these constructs (Fig. 2, A and B). Progressive deletions from –2340 (base pairs from the starting ATG) to –834 had only a slight effect on the mean expression level of the reporter and did not affect its oscillation (Fig. 2A) (10). On the other hand, deletion from –834 to –620 abolished rhythmicity. The region from –834 to –620 was sufficient to confer rhythmic expression to the reporter (Fig. 2B). To assess the possibility that LHY and/or CCA1 utilize this region to regulate TOC1 expression, the ability of these two proteins to bind to the –834 to –620 fragment was tested by performing electrophoretic mobility shift assay (Fig. 2, C and D) (11). Addition of extracts from Escherichia coliexpressing either GST-CCA1 or GST-LHY to a probe corresponding to the –834 to –620 region of the TOC1 promoter produced DNA species with retarded mobility (11); bacterial extracts harboring the empty glutathione S-transferase (GST) vector did not produce comparable bands. Competition experiments show that GST-CCA1 and GST-LHY specifically bind to a DNA fragment corresponding to this part of the promoter (Fig. 2, C and D). To delimit further this region, three more deletions were analyzed for the expression of the reporter (Fig. 2B). Removal of the region between –734 and –687 greatly reduced rhythmicity. Therefore, the –734 to –687 region contains cis elements important for clock regulation ofTOC1. Recently, a novel sequence element (evening element or EE) has been identified in the promoter region of 31 cycling genes peaking at the end of the subjective day (12). The importance of that element for circadian regulation in vivo was shown for one of those genes, CCR2. We found one EE in the –734 to –687 fragment of the TOC1 promoter. Mutation of the EE in the context of the –834 to –620 fragment caused strong reduction in rhythmicity (Fig. 2B). This indicates that the EE is important for clock regulation of TOC1. Given the high similarity between the EE (AAAATATCT) and the motif that CCA1 binds (AAAAATCT) (13), we investigated whether the EE was the actual target used by LHY and CCA1 to regulate TOC1. Figure 2, C and D, show the specific interaction of both transcription factors with the EE. Taken together, these results strongly suggest that LHY and CCA1 negatively regulate TOC1 expression by binding to the EE in its promoter region.

Figure 2

Identification of a region in the TOC1 promoter that is essential for its circadian oscillation and is a target of LHY and CCA1. (A andB) Analysis of the TOC1 promoter. The different deleted versions of the promoter were fused to the firefly luciferase+ gene and the resulting constructs were introduced intoArabidopsis by Agrobacterium-mediated DNA transfer (21). Primary transgenic plants were selected under LD (12, 12) cycles for 8 days, transferred to media without selection, and imaged and analyzed as described (7). In the mutant promoter [m– 834/–620 in (B)], the last six nucleotides of the EE were changed to CTGCAG. Error bars in (A) represent the standard error of the mean. The numbers in the upper left corner of each graph (A) and in the x axis (B) indicate the fragment of the TOC1 promoter analyzed. n, number of independent T1 seedlings analyzed. In (B), seedlings with a relative amplitude error lower than 0.6 were scored as rhythmic; values close to 1.0 indicate weak rhythms (9). Competition experiments reveal that the GST-CCA1 (C) and GST-LHY (D) proteins specifically bind to the EE in the –834/–620 region of theTOC1 promoter. No competitor DNA was added in lanes a and j. The unlabeled competitor DNA corresponded to the –834/–620 region ofTOC1 (–834/–620 TOC1; lanes b through e) or the 379 to 481 region of the pUC18 (379/481 pUC18; lanes f through i). In lanes k through n, the unlabeled competitor DNA corresponded to the –734/–687 region of TOC1 (–734/–687 wt EE), and the competitor DNA in lanes o through r was the same as in lanes k through n but with the last six nucleotides of the EE changed to CTGCAG (–734/–687 mut EE). Black triangles represent increasing amount of competitor, which was present at 1× (lanes b, f, k, and o), 10× (lanes c, g, l, and p), 100× (lanes d, h, m, and q), and 200× (lanes e, i, n, and r) molar excess of the labeled probe. The arrowhead indicates the protein-DNA complexes.

To investigate whether the regulation is reciprocal, we analyzedLHY and CCA1 expression patterns in theTOC1 alleles toc1-1 and toc1-2 (Fig. 3) (5, 7). Both theLHY and CCA1 messages cycled with a shorter period length in the two mutants than in wild-type plants, consistent with the short period observed for the expression of other genes (5, 7). The peak levels of LHY andCCA1 expression were reduced by more than 50% in thetoc1-2 background, and, consequently, the amplitude of oscillation for both transcripts was reduced. The different effects of the toc1-1 and toc1-2 mutations on the expression levels of LHY and CCA1 likely reflect the different nature of these alleles. toc1-1 is semidominant and appears to retain some of the TOC1 function, whereastoc1-2 is recessive and presumably represents a loss-of-function allele (5, 7). The low expression level of both genes might be the cause of the short-period phenotype oftoc1-2. It could also explain the same phenotype in theCCA1 loss-of-function allele cca1-1(14). The semidominant nature of toc1-1suggests distinct biochemical abnormalities, such as altered protein-protein interaction, for its defect in period length. The reduction of the transcript peak level for both LHY andCCA1 in toc1-2 strongly suggests a positive effect of TOC1 on the expression of both genes. Several features of TOC1 protein support this idea: It is localized in nuclear speckles that may represent transcriptional complexes (7); it has an acidic-rich region at the COOH-terminal end of the protein, which is commonly found among several types of eukaryotic transcriptional activators (7, 15); and it bears homology to the CONSTANS protein of Arabidopsis, which has been implicated in transcriptional regulation (16). The ability of two Arabidopsis response regulators to activate transcription has been described (17). The fact that both LHY and CCA1 transcripts continue to oscillate in toc1-2 indicates that TOC1 is not the only factor controlling expression of both genes. It has also been suggested that the basic helix-loop-helix transcription factor PIF3 acts as a positive element in LHY and CCA1 expression, at least in etiolated seedlings in response to light pulses (18). The possibility remains that TOC1 function is, to some degree, redundant with some of the TOC1-like genes found inArabidopsis, one of which oscillates with a phase very similar to TOC1 (19). The low level of TOC1 function expected in CCA1-OX could explain, in part, the low level of endogenous CCA1 and LHY messages found in these plants (4). However, the intermediate levels of endogenous LHY transcript found in thelhy allele lhy TN104 (3) suggest that positive elements other than TOC1 may still be active; an alternative, but not mutually exclusive explanation would be that the effects of lhy and lhy TN104 onTOC1 expression are different.

Figure 3

LHY and CCA1 expression is reduced in toc1-2 but not in toc1-1. toc1-1(A and B), toc1-2 (C andD), and wild-type plants (C24) were grown in LD (12, 12) cycles for 7 days and then released into LL. LHY andCCA1 RNA levels were analyzed by Northern blot. Representative data are shown.

As mentioned above, high levels of TOC1 expression correlate with low levels of LHY and CCA1 transcripts in the elf3-1 mutant in LL (Fig. 1C) (3, 10). However, the fact that the level of both transcripts remains low suggests that TOC1 fails to regulate positively their expression in the absence of the ELF3 function. Otherwise, the steady-state levels ofTOC1, LHY, and CCA1 transcripts would reach intermediate values due to the simultaneous action of positive (TOC1) and negative (LHY and CCA1) factors.

The results presented here show interactions among genes involved in the Arabidopsis oscillator. The data reveal different roles for LHY and CCA1 versus TOC1. The MYB factors act as negative elements that repress TOC1 expression, and, conversely, TOC1 appears to be a positive element for LHY and CCA1expression (Fig. 4). LHY and CCA1 appear to have a dual role in the Arabidopsis clock. It has been suggested that CCA1 acts as a positive regulator of CAB gene expression (4, 13, 14). Given the overall similarity betweenCCA1 and LHY, it is likely that the LHY protein is also a positive regulator of CAB genes. Furthermore, both of these proteins also act as negative regulators of one evening-phased gene, TOC1. We predict that LHY, CCA1, or both may also negatively regulate other genes having EE in their promoter regions, which include, for example, GIGANTEA and CCR2(12). Similar to a mechanism proposed forDrosophila (20), we suggest a model in which these MYB factors simultaneously regulate genes that are phased to different times of the day. The analysis of TOC1 biochemical properties and the identification of TOC1-interacting proteins will help us to close the loop. Our results suggest that the interactive regulation between these genes define the basic framework for the clock mechanism in Arabidopsis.

Figure 4

Model showing the proposed interactions betweenTOC1 and the MYB genes LHY and CCA1within the Arabidopsis circadian clock. Light activatesLHY and CCA1 expression at dawn (10). Posttranslational regulation by light has been suggested for CCA1 (4). Hence, light reinforces the function of LHY and CCA1 at dawn. CCA1, and probably LHY, activate CAB expression (4, 13, 14); they may activate other genes that cycle with a phase similar to CAB. LHY and CCA1 simultaneously repressTOC1 and potentially other evening genes. Progressive reduction of LHY and CCA1 expression levels during the day (3, 4) allows TOC1 transcript levels to rise and reach maximum levels toward the end of the day, whenLHY and CCA1 expression levels are lowest (3, 4, 7). TOC1, either directly or indirectly, appears to augment the expression of LHY and CCA1, which reach maximum levels at dawn, starting the cycle again.

  • * These authors contributed equally to this work.

  • Present address: Division of Biological Science, Graduate School of Science, Nagoya University, Furo-cho, Chikusa-ku, Nagoya, Aichi 464-8602, Japan.

  • To whom correspondence should be addressed. E-mail: stevek{at}


View Abstract

Navigate This Article