The Arabidopsis thaliana circadian clock

As mentioned in section 2.1, the scientific literature in plant circadian rhythms started almost three centuries ago, when the French astronomer De Mairan reported in 1729 that the daily leaf movements of the Mimosa pudica persisted in constant darkness, which suggested an endogenous origin of these rhythms. However he could not exclude that these persistent rhythms were under the influence of other environmental cues such as daily temperatures fluctuations. Nearly a century passed before De Candolle showed that these rhythms were endogenous and not simply responding to environmental timing cues. Indeed in 1832, he determined that the free running period of M.pudica was 22 to 23h, thus significantly shorter than 24h. In addition he demonstrated that this rhythm could be inverted when imposing a light/dark cycle with an opposite phase angle. All the further experiments were done exploiting plant leave movement, the only circadian manifestation known at this time.

Forward genetic analysis to identify components of circadian clocks began in the 1970s. At the same time, it was realized that leave movement was but one overt rhythm driven by the

clock. Other rhythms include germination, rate of hypocotyl and inflorescence stem elongation, enzymes activities, stomatal movement and gas exchange, photosynthetic activity, flower opening, and fragrance emission (Cumming and Wagner, 1968) (Figure 12).

As Arabodipsis was emerging as a prevailing system, the accumulation of several transcripts was discovered to be circadian. The use of the promoter of one of these rhythmic genes, Arabidopsis LHCB*3 (light-harvesting chlorophyll a/b binding protein, also called CAB2), in driving luciferase expression, provided a sensitive assay allowing to identify the first plant clock mutant, timing of cab expression1 (toc1-1) (Millar et al., 1995). Although the combination of components forming fly and mouse clocks (see section 2.4.1) is fairly similar, no obvious ortologs to most known clock proteins have been identified in Arabodipsis.

Figure 12. Circadian cycles of leaf movements in M.pudica and cytosolic Ca2+ in Arabidopsis.

Extended leaves during the day facilitate harvesting of light (left) and retracted leaves during the night protect from dehydration (center). Many clock-controlled rhythms of higher plants are driven by circadian cystosolic signals, including cycles of Ca2+, monitored here in Arabidopsis (right). This Figure is taken from (Hastings et al., 2008).

Nonetheless, even with a different molecular makeup, the interlocked transcriptional/post transcriptional feedback loops seems to operate in a similar manner in metazoans and plants.

The Arabodipsis clock is quite complex, and three interconnected feedback loops have been identified so far (for reviews, see (Harmer, 2009; McClung, 2006)).

expression of two Myb-related transcription factors (Myb, derived from myeloblastosis, is a DNA binding domain with a structure similar to the homeodomain helix-turn-helix motif);

circadian clock associated 1 (CCA1) and late elongated hypocotyl (LHY). CCA1 and LHY proteins bind directly to the toc1 promoter and inhibit its expression (Alabadi et al., 2001).

Loss-of-function mutants of toc1, cca1 or lhy engender a short period whiles their respective over-expression causes altered or dampened circadian gene expression (Alabadi et al., 2002;

Makino et al., 2002; Schaffer et al., 1998; Wang and Tobin, 1998). TOC1 alone seems not to be sufficient for inducing the expression of CCA1 and LHY. Several other genes, including gigantea (gi), early flowering 4 (elf4), and lux, are required for proper CCA1 and LHY expression (Doyle et al., 2002; Hazen et al., 2005; Park et al., 1999) (Figure 13).

More recently, a second loop in which TOC1 is activated by a hypothetical evening-expressed protein that itself is repressed by TOC1 was proposed. Giantea might represent this hypothetical component, since TOC1 expression is dependent upon GI (but not solely), and GI expression is indirectly affected by TOC1. This suggests that GI acts both in series with and in parallel to TOC1 within the central circadian oscillator (Martin-Tryon et al., 2007).

The Arabidopsis genome contains three genes encoding proteins with similarity to TOC1; PRR5, 7 and 9 (pseudo-response regulator). Reverse genetic approaches revealed that they all play a role in the clock. Although the single mutant phenotypes are subtle, combined mutants have stronger phenotypes - prr5/prr7/prr9 triple mutant is essentially arrhythmic - suggesting that they play overlapping and distinctive roles close to (or within) the central oscillator (Nakamichi et al., 2005). Cca1 and lhy transcripts accumulate in prr7 and prr7/prr9 mutants, indicating that PRR7/PRR9 might be negative regulators of cca1and lhy. Prr7 and prr9 expression peaks are strongly reduced in cca1/lhy mutants, and CCA1 is able to specifically bind promoter elements of prr7 and prr9 in vitro, placing CCA1/LHY as potent positive regulators of prr7 and prr9 (Farre et al., 2005). This findings support the participation of CCA1/LHY and PRR5, 7 and 9 in a possible third feedback loop (Figure 13).

Similarly to the circadian systems described above, post-transcriptional controls are also required for proper clock function in Arabidopsis. The stability and translation of some mRNAs are influenced by the circadian clock and light signaling (see below), and the abundance of many clock proteins is under post translational control. For example, the phosphorylation of CCA1 and LHY by CK2 (casein kinase II), as well as the modification of GI by N-acetylglucosamine transferase SPINDLY, seem to be implicated in the oscillation process (Daniel et al., 2004; Tseng et al., 2004). Also, DET1 protein acts to inhibit the proteolytic turnover of the LHY protein. In accordance with this explanation, det1 mutatants exhibit a short period phenotype (Song and Carre, 2005). Furthermore, the stability of the ZTL (zeitlupe) protein, a blue light photoreceptor is critical for the regulation of core clock TOC1 levels. While ztl RNA is constitutively expressed, the level of ZTL protein cycles. ZTL contains a LOV domain (flavin mononucleotide-binding region) that allows ZTL to sense blue light and to confer a blue light-enhanced interaction of ZTL with GI. ZTL is more stable at dusk when associated with GI, and is more rapidly degraded at dawn. In addition, the Figure 13. A Molecular model of the Arabidopsis thaliana circadian oscillator. The core CCA1/LHY/

TOC1 feedback loop is highlighted in green. Shaded area indicates activities peaking in the subjective night, and white area indicates activities peaking during the subjective day. See text for details.

This Figure is issued from (McClung, 2006).

turn ZTL, which is a component of an SCF complex that recruits TOC1 for proteasomal degradation, participates in the high-amplitude rhythm of TOC1 necessary for proper clock function (Kim et al., 2007).

The clock regulates many aspects of plant physiology, and nearly one third of the transcriptome consists of rhythmically transcribed genes in Arabidopsis (Covington et al., 2008). These hundreds of genes have peak phases of expression occurring at all times of the day and night. Their transcription is mediated by at least three phase-regulatory modules, ME/G-box (morning elements), EE/GATA (evening elements), and PBX/TBX/SBX (protein box, enriched in protein synthesis genes/telo-box/starch box, all three of these elements are enriched in midnight peaking genes). To some extent, this phase-dependent cis-acting elements are somewhat similar in function to the three mammalian elements, E-box (CACGTG), D-box (DBP/E4BP4 binding elements), and RRE (RevErbA/ROR binding elements) (Michael et al., 2008). CCA1 and LHY are likely to act as repressors via the EE, but they also have an unexpected positive effect on EE-mediated gene expression (Harmer and Kay, 2005). Such studies provide first insights on the time specific transcriptional regulatory network that results from the complex interplay between thermocycles, photocycles and the circadian core clock, and that partitions biological activities to the appropriate time windows over the day.

Changes in light and temperature can have strong effects on the circadian plant system. As described above, ZTL is a photoreceptor that controls TOC1 stability in a blue light-regulated manner. Both phytochromes and cryptochromes provide also red and blue light inputs to the clock (Tepperman et al., 2001). The exact signaling pathways downstream of these photoreceptors have not yet been elucidated. CCA1 and LHY expression is induced by light (Wang and Tobin, 1998). The bHLH transcription factor PIF3 (phytochrome interacting factor 3) provides perhaps a direct link from light perception to modification of the negative limb of the circadian clock (Martinez-Garcia et al., 2000). ELF3 (early flowering 3)

and TIC (time for coffee) negatively regulate light input, while SRR1 (sensitivity to red light reduced 1) is a positive regulator of signaling in response to red and white light, but their exact roles are yet unknown (Hall et al., 2003; McWatters et al., 2000; Staiger et al., 2003).

Loss-of-function alleles of elf3 elicit conditional arrhythmicity in continuous light but remain rhythmic in the dark. Light also promotes the degradation of cca1 mRNA and increases the translation rate of lhy mRNA (Kim et al., 2003; Yakir et al., 2007). These data show that light signals modulate the expression of many clock genes by acting at diverse regulatory levels.

Like other circadian oscillators plant circadian clocks are temperature compensated.

The expression of cca1, lhy, toc1 and gi is modulated by temperature. The level of lhy mRNA decreases when temperature increases. And this is counterbalanced by increases in toc1 and gi, while cca1 levels change insignificantly. However cca1 levels increase at lower temperatures. At high temperature, circadian function is impaired in lhy and gi mutant plants, while at low temperature cca1 and gi mutations impair circadian rhythms more than lhy mutations. These observations suggest that GI may participate in the stablilization of the period over a wide range of temperatures by counterbalancing the relative expression levels of CCA1 and LHY in a temperature-dependent manner (Gould et al., 2006). Temperature changes can act to entrain or reset the clock. However, temperature signaling to the clock is still unclear. Gene expression and cotyledon movement can be entrained by temperature cycles. In addition to the number of output rhythms that are set by thermocycles, the clock genes cca1, lhy, toc1, prr7 and prr9 can be set to their correct phase by thermocycles.

Prr7/prr9 double mutant fails to efficiently entrain to temperature cycles and do not respond to temperature pulses (Salome and McClung, 2005). In this study, Salomé and McClung also found that the role of PRR7 and PRR9 is not limited to temperature entrainment, because profound effects are also seen in the prr7/prr9 mutant after entrainment to photocycles. They propose thus that PRR7 and PRR9 are clock components necessary for the integration of light and temperature signaling.

Plants and animals manifest obvious rhythms matching the annual seasonal changes in the environment. In plants, seasonal rhythms include, for example, the formation of flowers or the acclimation to cold at the appropriate time of the year. Many plants flower in response to day length (photoperiod). A complex interplay between endogenous oscillators and external light oscillations controls flowering. A specific phase coincidence between these internal and external oscillations promotes flowering. In Arabidopsis, the key factor for flowering is CO (constans). During short days, co mRNA accumulates late in the day and is translated after dusk, but the CO protein is unstable and fails to accumulate in the dark. However, when day length increases, light perception via CRY2 and PHYA stabilizes CO protein during the end of these long days, allowing CO to activate its targets (Valverde et al., 2004). Such interplay may also occur between exogenous temperature variations and internal oscillations, as suggested by Balasubramania et al. (Balasubramanian et al., 2006). These authors show that growth temperatures above a finely tuned threshold can rapidly trigger flowering, bypassing the need for other inductive stimuli such as day length. A gating of low temperature-induced transcription by the circadian clock seems also to be responsible for the activation of mechanisms required for an increase in freezing tolerance, the adaptive response known as cold acclimation (Fowler et al., 2005).

Similarities and differences between phyla

Now that we have a more detailed view of the different circadian systems in diverse organisms, it is clear that a basic theme - the use of transcriptional/post transcriptional feedback loops to generate endogenous cell-autonomous oscillation - remains remarkably similar, although the players are not always conserved and can be mixed and matched in a variety of ways. There is indeed no sequence conservation between the clock components of prokaryotes and eukaryotes. Within eukaryotes, strong homologies are found between clock components of Drosophila and mammals (see next section) suggesting a conservation during

metazoan evolution. The clock of lower eukaryotes such as N.crassa exhibit a conservation with higher eukaryotes limited to the presence of PAS protein-protein interaction. However no ortolog exist between plant and animal clocks. It is therefore possible that these similar regulatory circuits arose by convergent evolution.

In unicellular or diaphanous organisms, the oscillator is directly linked to environmental light and temperature changes for resetting and entrainment. As mentionaed previously, brain and peripheral oscillators can be directly entraned by light in Drosophila that apparently lacks a dominant centralized pacemacker. By contrast, in opaque homoeothermic multicellular organisms, such as rodents and birds, peripheral tissues are not directly entrained by light, and within the body, temperature is not subject to strong oscillations. Therefore, such organisms exhibit a hierarchical organization of the circadian system, and a centralized pacemaker system in the brain seems to be essential for converting photic phase-setting cues into downstream signals that entrain peripheral tissues.