A Complex Genetic Interaction Between <em>Arabidopsis thaliana</em> TOC1 and CCA1/LHY in Driving the Circadian Clock and in Output Regulation

Abstract

CIRCADIAN rhythms are self-sustaining biological oscillations that free run under constant conditions with a periodicity close to 24 hr. The rhythmic clock is prevalent and is found in organisms ranging from prokaryotes to eukaryotes and from animals to plants (Dunlap 1999; Baraket al. 2000; Harmeret al. 2001). This clock can be reset according to environmental cues, such as light and temperature (Liuet al. 1998; Collettet al. 2001; Young and Kay 2001; Samach and Wigge 2005; Carret al. 2006). Recently, rapid strides have been made in deciphering the molecular bases of the circadian system. A recognizable pattern that is emerging is the recurring trend of autoregulatory positive/negative feedback loops (Alabadiet al. 2001). Further, clock models have been mathematically derived and the resulting equation principals can be applied (Lockeet al. 2005a,b; Lakin-Thomas 2006). These models explicitly generate hypothesis-driven questions.

In Arabidopsis thaliana, the proposed negative repressors of the oscillator are the morning-acting myb-related factors CIRCADIAN CLOCK ASSOCIATED 1 (CCA1) and LATE ELONGATED HYPOCOTYL (LHY) (Schafferet al. 1998; Wang and Tobin 1998; Green and Tobin 1999; Alabadiet al. 2002; Mizoguchiet al. 2002), which are partially redundant genes encoding similar DNA-binding proteins. They act on the proposed positive-activator termed TIMING OF CAB EXPRESSION 1 (TOC1); it works in the evening and encodes a protein of unknown biochemical activity (Somerset al. 1998a; Strayeret al. 2000; Alabadiet al. 2001). TOC1 belongs to the PRR (PSEUDO-RESPONSE REGULATOR) family of proteins, consisting of five members (PRR9, PRR7, PRR5, PRR3, and PRR1/TOC1) (Matsushikaet al. 2000; Makinoet al. 2002; Erikssonet al. 2003). TOC1 and CCA1/LHY together make up the proposed central circadian loop in Arabidopsis. It is this positive–negative feedback loop between these evening and morning factors that leads to the first genetic model of the plant clock (Alabadiet al. 2001). This regulatory network has consistently and continuously been placed at the core of the molecular oscillator in all published models, but it does not fully describe several experimentally defined features (Schafferet al. 1998; Wang and Tobin 1998; Harmeret al. 2000; Alabadiet al. 2001; Kimet al. 2003). Mathematical approaches drove experimental approaches to refining a simplistic loop where only CCA1/LHY and TOC1 were the sole elements of the clock. An interlocked two-loop clock model was then proposed to describe oscillatory properties, such as entrainment and response to photoperiods (Lockeet al. 2005b). In this model, TOC1 and CCA1/LHY form a central loop, while the flowering-time gene GIGANTEA (GI) works alongside TOC1 to compose a secondary loop. CCA1 and LHY mediate light signal into the clock and GI potentially provides a secondary pathway for light input into the clock (Lockeet al. 2005b). More recently, two groups have extended this to a three/four-loop model that includes PRR9 and PRR7 as morning-acting elements in a tertiary CCA1/LHY loop. We note that none of these studies has tested whether a loop with CCA1/LHY and TOC1 is indeed core to the oscillator (Lockeet al. 2006; Zeilingeret al. 2006).

The circadian clock has been reported to regulate many processes, such as daily biochemical reactions and other general metabolic aspects of the cell. This in turn coordinates most, if not all, physiological processes. These are collectively called the circadian-output pathways (Harmeret al. 2000). For example, both toc1 and cca1 lhy have defects in flowering time and photomorphogenesis, which correlates with respective mutant circadian phenotypes (Somerset al. 1998b; Strayeret al. 2000; Mizoguchiet al. 2002). In toc1, mutant plants have an early flowering phenotype when grown under a short-day photoperiod. It was found that this phenotype is the result of clock-based misinterpretation of photoperiodic information rather than of the direct effects of toc1 on floral-induction pathways (Somerset al. 1998b; Strayeret al. 2000). Both cca1 and lhy also exhibit an early flowering phenotype under short-day conditions, and this was especially marked in the cca1 lhy double mutant; this double mutant is nearly insensitive to photoperiodic sensing (Mizoguchiet al. 2002). Although both toc1 and cca1 lhy have an early flowering phenotype, they have an inverted phenotype regarding early seeding photomorphogenesis, with toc1 displaying a long hypocotyl whereas cca1 lhy displays a short hypocotyl (Maset al. 2003; Mizoguchiet al. 2005).

We sought to provide direct experimental evidence for TOC1 and CCA1/LHY as core-loop elements in the clock and to disclose the genetic relationship between TOC1 and CCA1/LHY in output regulation. For this purpose, we established all the possible double mutants and the triple mutant, tested clock responsiveness under a battery of molecular assays, and performed physiological and molecular analysis of clock outputs. We found that the triple mutant cca1 lhy toc1 often exhibited an arrhythmic phenotype under constant light (LL) conditions, which was consistent with the predictions from current mathematical clock models. Interestingly, the triple mutant displayed some limited rhythmic behavior under certain assays. The implication from this experimental data set is that the latest three/four-loop mathematical model (Lockeet al. 2006; Zeilingeret al. 2006) will need to be further refined. Also, we found that TOC1 and CCA1/LHY participate in photomorphogenesis and flowering-time promotion through distinct epistatic relationships.

Acknowledgments