Reconstitution of Circadian Oscillation of Cyanobacterial KaiC Phosphorylation in Vitro

See allHide authors and affiliations

Science  15 Apr 2005:
Vol. 308, Issue 5720, pp. 414-415
DOI: 10.1126/science.1108451


Kai proteins globally regulate circadian gene expression of cyanobacteria. The KaiC phosphorylation cycle, which persists even without transcription or translation, is assumed to be a basic timing process of the circadian clock. We have reconstituted the self-sustainable oscillation of KaiC phosphorylation in vitro by incubating KaiC with KaiA, KaiB, and adenosine triphosphate. The period of the in vitro oscillation was stable despite temperature change (temperature compensation), and the circadian periods observed in vivo in KaiC mutant strains were consistent with those measured in vitro. The enigma of the circadian clock can now be studied in vitro by examining the interactions between three Kai proteins.

Circadian rhythms allow organisms to coordinate their lives according to the alteration of their environments by day and by night (1). In most model organisms, transcription-translation-derived oscillatory (TTO) processes based on negative feedback regulation of clock genes are proposed as the core generator of self-sustaining circadian oscillations (1, 2). Cyanobacteria are the simplest organisms that exhibit circadian rhythms. In the cyanobacterium Synechococcus elongatus (PCC 7942), three genes (kaiA, kaiB, and kaiC) are essential components of the circadian clock (3). Negative-feedback regulation of the expression of the kaiBC operon by Kai proteins was proposed as a core loop of prokaryotic TTO (3). In the cyanobacterial TTO model, Kai proteins do not regulate a specific set of circadian-controlled genes, but they regulate genomewide gene expression, including that of the kaiBC operon (4). However, how this transcription-translation feedback loop achieves circadian periodicity and how it stabilizes circadian oscillation against alterations in temperature and metabolic activity (collectively referred to as circadian characteristics) have not been clear. These circadian characteristics are essential for the oscillator to adapt to the environment, so understanding their molecular basis is an important goal in circadian biology.

The phosphorylation state of KaiC robustly oscillates in the cell with a 24-hour period, even under conditions where neither transcription nor translation of kaiBC operon was permitted, raising doubts about the TTO model in cyanobacteria (5). The circadian characteristics of this oscillator suggested that the pacemaker for the cyanobacterial circadian system was not a transcription-translation feedback loop but the KaiC phosphorylation cycle itself. KaiC has both autophosphorylation and autodephosphorylation activities, and KaiA enhances KaiC autophosphorylation (6), whereas KaiB attenuates the effect of KaiA (7, 8). These results imply that, without additional kinases or phosphatases, an autonomous oscillation of KaiC phosphorylation could be generated by cooperation between KaiA and KaiB.

Recombinant KaiC protein was incubated with KaiA and KaiB at a ratio similar to that measured in vivo [KaiA:KaiB:KaiC=1:1:4 (by weight), (7)] in the presence of 1 mM ATP. To our surprise, KaiC phosphorylation robustly oscillated with a period of about 24 hours for at least three cycles without damping (Fig. 1A). The ratio of phosphorylated KaiC to total KaiC cycled between 0.25 and 0.65 (Fig. 1C). The amplitude of this in vitro KaiC phosphorylation rhythm was smaller than that observed in vivo under continuous light conditions (5). Also the total amount of KaiC remained constant during the incubation (Fig. 1B), which indicated that neither phosphorylated nor unphosphorylated KaiC was degraded during the reaction. Thus, oscillation of KaiC phosphorylation generates autonomously with a circadian period by cooperation of three Kai proteins.

Fig. 1.

In vitro oscillation of KaiC phosphorylation. (A) Recombinant KaiC proteins (0.2 μg/μl) were incubated with KaiA (0.05 μg/μl) and KaiB (0.05 μg/μl) in the presence of ATP (1 mM) (17). Aliquots (3 μl each) of the reaction mixtures were collected every 2 hours and subjected to SDS-polyacrylamide electrophoresis (SDS-PAGE) and Coomassie Brilliant Blue staining. The upper and lower bands correspond to phosphorylated (P-KaiC) and unphosphorylated KaiC (NP-KaiC), respectively (6). (B and C) NIH image software was used to perform densitometric analysis of data (5) in (A). The relative densities of total, phosphorylated, and unphosphorylated KaiC are plotted in (B), and the ratios of P-KaiC to total KaiC are plotted in (C).

One characteristic of circadian rhythms is that the free-running period remains stable for a relatively broad range of temperatures, referred to as “temperature compensation” of the period (1). Such temperature compensation of the in vitro oscillation of KaiC phosphorylation was observed with period lengths of about 22, 21, and 20 hours at 25°, 30°, and 35°C, respectively (Fig. 2A). The thermal sensitivity (Q10 coefficient) of the period (Fig. 2B) was about 1.1, consistent with that reported for in vivo gene expression rhythm (9). The rates of in vitro KaiC autophosphorylation and autodephosphorylation show a high degree of compensation for changes in temperature (5). Thus, temperature compensation of the in vitro rhythm is likely due to these reactions.

Fig. 2.

Temperature compensation of the period of the in vitro oscillation of KaiC phosphorylation. (A) Three Kai proteins were incubated as in Fig. 1A at 25°, 30°, and 35°C. SDS-PAGE and densitometric analyses were carried out as in Fig. 1. (B) The period of the oscillation of the in vitro KaiC phosphorylation state was plotted against the incubation temperature. Results from two independent experiments were plotted.

Many mutations of kaiC shorten or extend the period of circadian rhythm of Synechococcus (3). To assess whether the in vitro oscillation of KaiC phosphorylation was affected by such mutations, we incubated recombinant mutant KaiC with KaiA and KaiB. Three KaiC mutants that display high-amplitude bioluminescence rhythms were examined. The period lengths of expression rhythm of kaiBC promoter, monitored with a bioluminescence reporter (3), were 17, 21, and 28 hours, respectively, in mutant stains with amino acid substitutions of Tyr for Phe470 (F470Y), Pro for Ser157 (S157P), and Ser for Thr42 (T42S). The KaiC phosphorylation profiles, obtained when the mutant proteins were assayed in vitro, were consistent with those observed in vivo (Fig. 3A). We also confirmed that the bioluminescence profiles of these mutant strains were consistent with the in vitro oscillation of phosphorylation for each of the respective mutant KaiC proteins (Fig. 3, A and B). These results indicate that oscillation of KaiC phosphorylation is the molecular timer for the circadian rhythm of Synechococcus.

Fig. 3.

Correlation of period length between the in vitro oscillation of the KaiC phosphorylation and the in vivo rhythms from wild-type and mutant strains. (A) Bioluminescence profiles from the P-kaiBC reporter of mutants (F470Y, S157P, and T42S) and wild-type strains were monitored under continuous light conditions at 30°C with a photomultiplier-based assay system (3, 17). The in vitro oscillations of KaiC assayed as described in Fig. 1 and in vivo mutant KaiC phosphorylation profiles were analyzed as described (5, 17). In vitro ratios of P-KaiC to total KaiC (filled circle), in vivo ratios of P-KaiC to total KaiC (open circle), and bioluminescence profiles (open square) are plotted. To compare the period lengths, the phases of in vitro and in vivo oscillations of KaiC phosphorylation are shifted to put first peak of phosphorylation rhythms on that of the corresponding bioluminescence rhythm. Periods of bioluminescence rhythm are shown in parentheses. (B) For wild type and each mutant strain shown in (A), the period length of the in vitro oscillation of the KaiC phosphorylation state is plotted against that of the in vivo rhythms of the strain. Results from two independent experiments were plotted.

Phosphorylation of clock proteins has been reported in various prokaryotic and eukaryotic model organisms. PERIOD and TIMELESS in Drosophila and FREQUENCY in Neurospora degraded and/or translocated to the nucleus according to their circadian rhythms of phosphorylation state (10, 11). Because alterations of the phosphorylation of these clock proteins affect the period length, the phosphorylation processes were assumed to be important components of the TTO models, including those for cyanobacteria (2, 3). In these models, phosphorylation only contributes at a specific phase of the circadian cycle. However, our study demonstrates that the oscillation of KaiC phosphorylation is the pacemaker of the cyanobacterial circadian clock. In addition, the in vitro oscillation is generated in a homogenous system, whereas heterogeneous compartments are assumed in eukaryotic models (2). KaiC forms hexamers (12, 13) and is phosphorylated at Ser431 and Thr432 (14, 15). These results imply that KaiC hexamer has multiple phosphorylation states that may have different biochemical characteristics, including autophosphorylation and autodephosphorylation activities and/or binding preferences with other Kai proteins. In addition, the reaction rates of KaiC phosphorylation and dephosphorylation are quite slow, with rate constants of 10-3 to 10-4 s-1 [(16), SOM text]. Functionally, this feature would reduce the energy needed for timekeeping.

We propose a model of the cyanobacterial clock in which the autonomous oscillation of KaiC phosphorylation controls the expression of relevant genes including the kai genes to generate physiologically functional circadian oscillation (fig. S1). Simultaneously, the in vivo oscillation of KaiC phosphorylation could be amplified by coupling it with periodic changes in the concentrations of Kai protein and/or additional regulatory components of KaiC phosphorylation. The relationship between the phosphorylation of KaiC and the Kai transcription-translation cycle may be similar to that of a pendulum and an escapement mechanism that sustains the pendulum oscillation and transmits time signals to the hands of a wall clock.

References and Notes

View Abstract

Navigate This Article