FEBS Letters 584 (2010) 898–902

journal homepage: www.FEBSLetters.org

In vitro regulation of circadian rhythm of cyanobacterial clock protein KaiC by KaiA and KaiB

Masato Nakajima 1,2, Hiroshi Ito 1,3, Takao Kondo *

Division of Biological Science, Graduate School of Science, Nagoya University and CREST, Japan Science and Technology Agency (JST), Furo-cho, Chikusa-ku, Nagoya 464-8602, Japan article info abstract

Article history: Biochemical circadian oscillation of KaiC phosphorylation, by mixing three Kai proteins and ATP, Received 25 September 2009 has been proven to be the central oscillator of the cyanobacterial . In vivo, the intra- Revised 29 December 2009 cellular levels of KaiB and KaiC oscillate in a circadian fashion. By scrutinizing KaiC phosphorylation Accepted 8 January 2010 rhythm in a wide range of Kai protein concentrations, KaiA and KaiB were found to be ‘‘parameter- Available online 16 January 2010 tuning” and ‘‘state-switching” regulators of KaiC phosphorylation rhythm, respectively. Our results Edited by Jesus Avila also suggest a possible entrainment mechanism of the cellular circadian clock with the circadian variation of intracellular levels of Kai proteins. 2010 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved. Keywords: Ó Circadian clock Cyanobacterium KaiC phosphorylation rhythm Entrainment

1. Introduction three Kai proteins and that this rhythm functions as the most basic pacemaker of the cyanobacterial circadian clock. , the occurrence of daily oscillations in various How only three Kai proteins and ATP generate self-sustainable cellular activities, occurs in a wide spectrum of organisms from oscillation has been studied extensively both experimentally and to humans [1]. Among these, cyanobacteria are the theoretically [8,9]. However, the relationship between KaiC phos- simplest organism known to show a circadian rhythm [2–4]. Three phorylation rhythm and the intracellular environment or meta- , kaiA, kaiB and kaiC, were identified to code for central com- bolic activities of the cell remains poorly understood. Specifically, ponents of the cyanobacterial circadian clock, and feedback regula- when cyanobacteria are actively growing in continuous light (LL) tion of kaiBC expression had been assumed to generate circadian conditions, levels of KaiC and KaiB show circadian oscillations, oscillation in cyanobacteria [5]. However, we found that circadian while KaiA level remains constant [10]. Moreover, we recently oscillation of KaiC phosphorylation was persistent under found that circadian rhythm could be maintained independent of prolonged darkness, during which of kaiBC was pro- the KaiC phosphorylation rhythm [11]. Therefore, the significance hibited [6]. Moreover, a temperature-compensated KaiC phosphor- of KaiC and KaiB accumulation rhythm should be reconsidered to ylation rhythm was reconstituted in vitro by mixing KaiA, KaiB and understand the cellular circadian rhythm of cyanobacteria. KaiC in the presence of ATP [7], and several mutations of KaiC con- KaiC phosphorylation rhythm occurs in a cooperation with KaiA sistently modified the period lengths of both in vivo -expres- which enhances KaiC phosphorylation [12], and KaiB which atten- sion rhythm and in vitro KaiC phosphorylation rhythm [7]. These uates the role of KaiA [10]. KaiA and KaiB also regulate the temper- results demonstrated that KaiC phosphorylation rhythm was ature-compensated ATPase activity of KaiC [13]. However, no other autonomously generated by post-translational processes of the biochemical functions of these proteins have been reported despite many studies on their structure [14,15]. In this report, we evaluated the dependency of KaiC phosphor- ylation rhythm on the concentrations of the three Kai proteins. We * Corresponding author. Fax: +81 52 789 2963. found that both the period and amplitude of KaiC phosphorylation E-mail address: [email protected] (T. Kondo). rhythm were sensitive to the concentration of KaiA, but not to that 1 These authors contributed equally to this work. of KaiB, though KaiB was essential for generating KaiC phosphory- 2 Present address: Laboratory for Systems Biology, RIKEN Center for Developmental lation rhythm. Based on these results, we suggest a possible Biology, 2-2-3 Minatojima-minamimachi, Chuo-ku, Kobe, Hyogo 650-0047, Japan. 3 Present address: Division of Advanced Sciences, Ochadai Academic Production, entrainment mechanism of KaiC phosphorylation rhythm by Ochanomizu University, 2-1-1 Ohtsuka, Bunkyo-ku, Tokyo 112-8610, Japan. changing the ratios of these three proteins.

0014-5793/$36.00 Ó 2010 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.febslet.2010.01.016 M. Nakajima et al. / FEBS Letters 584 (2010) 898–902 899

2. Materials and methods A 1

2.1. Purification of recombinant Kai proteins KaiA=6.0 0.8 ↓ 4.8 Recombinant Kai proteins were produced in Escherichia coli ↓ BL21 and purified as previously described [7,16]. 0.6

3.6↓

2.2. Reconstitution of KaiC phosphorylation rhythm in vitro 2.4 ↓ 0.4 ↓ 1.2 The in vitro reconstitution of KaiC phosphorylation rhythm was Ratio of P-KaiC of Ratio 0.6

performed as described [7]. The concentrations of Kai proteins 0.2 ↓ were changed as described in each figure legend. Aliquots (3 ll) ↓ of reaction mixtures were collected every 4 h and subjected to 0.3 0 SDS–PAGE and Coomassie staining. The ratio of phosphorylated 0 102030405060 KaiC was calculated using NIH image software. Incubation time (h)

2.3. Data analysis B 1 To determine period (s) and amplitude (A), KaiC phosphoryla- ↓ KaiB=0.7 tion ratio (Y) after 16 h incubation was fitted to a cosine function 0.8 of time (t) as follows: ↓ 0.875 YðtÞ¼A cosð2pðt aÞ=s þ b 0.6 1.75 where a and b represent the peak of the rhythm (in hours) and non- ↓ oscillatory component, respectively. We calculated square errors of 0.4 3.5

cosine curves with 108 combinations of parameters. The combina- ↓ Ratio of P-KaiC of Ratio

tion of parameters with the least square error was chosen as the 0.2 best fit. ↓ 7.0, 10.5, 17.5 0 3. Results 0 102030405060 Incubation time (h) 3.1. KaiA and KaiB as parameter-tuning and state-switching regulators Fig. 1. KaiC phosphorylation rhythm in the presence of various concentrations of of KaiC phosphorylation rhythm KaiA and KaiB. Mixtures for in vitro KaiC phosphorylation rhythm were prepared as described in Section 2 with varying concentrations of either KaiA (A) or KaiB (B). Accumulation of KaiC and KaiB in cyanobacterial cells oscillates The ratio of phosphorylated KaiC to total KaiC was plotted against incubation time. in a circadian fashion, while the level of KaiA is kept constant [10]. To elucidate the possible effect of in vivo oscillations of Kai protein the in vitro rhythm. These results indicated that KaiB was a state- level on circadian rhythm generation, we measured the rhythm of switching regulator of KaiC phosphorylation rhythm. KaiC phosphorylation in vitro as a function of different concentra- tions of KaiA and KaiB (Fig. 1). KaiC phosphorylation level oscil- 3.2. Period and amplitude dependency on KaiA and KaiB lated with a period of about 23.5 h in standard conditions (Fig. 1; concentrations KaiA, KaiB and KaiC concentrations were 1.2, 3.5 and 3.5 lM, respectively). When only KaiA concentration was changed The in vitro KaiC phosphorylation rhythm observed under vari- (Fig. 1A), KaiC phosphorylation rhythm was observed in the pres- ous concentrations of KaiA and KaiB is summarized in Fig. 2. When ence of 0.6–3.6 lM KaiA, and the oscillation was dampened at con- KaiB concentration was 3.5 lM or higher, amplitude was not chan- centrations outside of this range. A lower limit on concentration ged by KaiB concentration, but was dependent on KaiA concentra- was evident, as a decrease in KaiA concentration below 0.6 lM se- tion, with the highest amplitude from 0.6 to 2.4 lM KaiA (Fig. 2A). verely dampened the in vitro rhythm (Fig. 1A). An upper limit of However, when KaiB concentration was 3.5 lM or lower, the con- KaiA concentration leading to oscillation of KaiC phosphorylation centration of KaiA required to generate the rhythm depended on was not determined with the KaiA concentration range tested. KaiB concentration (Fig. 2A and B). Phosphorylation rhythm was stably observed from 0.6 to 2.4 lM As far as KaiB concentration permitted rhythm generation, KaiA of KaiA, and weakened gradually in the KaiA concentration ranging defined the period but KaiB did not tune the period length (Fig. 2B). from 3.6 to 6.0 lM(Figs. 1A and 2A). As KaiA enhanced phosphor- The period was shortened from about 25 h to 21 h, when KaiA con- ylation of KaiC [12], the average level of KaiC phosphorylation ele- centration increased from 0.6 to 2.4 (Fig. 2B). These results once vated with increasing KaiA concentration (Fig. 1A). The period and again showed parameter-tuning and state-switching functions of amplitude of the rhythm seemed to be continuously influenced by KaiA and KaiB, respectively. KaiA concentration in the effective range. Thus, KaiA is assigned as a parameter-tuning regulator of KaiC phosphorylation rhythm. 3.3. Biochemical simulation of the intracellular oscillation of Kai The effect of KaiB concentration on KaiC phosphorylation proteins rhythm was different from that of KaiA. Oscillation of KaiC phos- phorylation was observed with KaiB concentrations of 1.75 lMor To address the significance of intracellular oscillations in accu- higher, while excess KaiB did not affect period nor amplitude mulation levels of KaiB and KaiC, we measured the in vitro rhythm (Fig. 1B). As is the case with KaiA, there seemed to be a lower limit to simulate the in vivo situation. When the ratio of KaiC and KaiB concentration of KaiB for generating KaiC phosphorylation rhythm, concentration was kept constant (1:1) and KaiA concentration because KaiB concentrations below 1.75 lM drastically reduced was unchanged, both the period and amplitude of phosphorylation 900 M. Nakajima et al. / FEBS Letters 584 (2010) 898–902

(A) Amplitude (B) Period

0.25 30 0.2 )h(doireP 25 0.15

0.1 20 0.05

Amplitude 0 15 14 14 10.6 0.6 10.6 0.6 KaiB 7.0(μM) 1.2 KaiB 7.0(μM) 1.2 3.5 3.5 2.4 2.4 1.75 1.75 0.875 3.6 0.875 3.6 0.35 4.8 KaiA (μM) 0.35 4.8 KaiA (μM) not determined

Fig. 2. Amplitude and period of KaiC phosphorylation rhythms at different concentrations of KaiA and KaiB. Assays of in vitro KaiC phosphorylation rhythm were performed with various concentrations of KaiA and KaiB, while the level of KaiC was constant (3.5 lM). The amplitude (A) and period (B) are shown on the plane of KaiA and KaiB concentrations. The period was not determined for conditions with amplitude below 0.1 (panel B, black squares).

rhythm were dramatically affected by altering KaiC and KaiB con- centrations (Fig. 3A). In cyanobacterial cells, the range of circadian change in accumulation levels of KaiC and KaiB was estimated to A 1 be threefold [10]. In this study, we confirmed that threefold alter- ation of KaiC and KaiB concentrations in vitro changed the period KaiC=0.875 by 6 h (Fig. 3B). This result implied that intracellular oscillation 0.8 ↓ of KaiB and KaiC levels can alter the angular velocity of KaiC phos- phorylation rhythm to enable the entrainment of the circadian 0.6 1.75

clock to the external light/dark (LD) cycle (see Section 4). It was ↓ also implied that the KaiC phosphorylation rhythm in cells was ↓

not independent from feedback regulation of kaiBC expression that 0.4 3.5

caused intracellular oscillations in Kai protein levels. It should also ↓ Ratio of P-KaiC of Ratio ↓ be noted that in vitro KaiC phosphorylation rhythm is most robust 0.2 7.0

when its period is tuned to 22–26 h (Fig. 3B). 10.5 ↓ 14 4. Discussion 0 0102030405060 Incubation time (h) 4.1. Different roles of KaiA and KaiB in defining KaiC phosphorylation rhythm B 32 0.2 Given the autonomous oscillation of KaiC phosphorylation in Amplitude the presence of a constant amount of Kai proteins both in vitro 30 0.18 [7] and in vivo under constant dark (DD) conditions [6], the biolog- ical relevance of oscillation in cellular accumulation levels of KaiC 28 0.16 and KaiB is poorly understood. As shown by determining a poten- tial effect of a temperature pulse in resetting the phase of the KaiC 26 0.14

oscillator [17,18], quantitative measurement of in vitro KaiC phos- Period (h) Period phorylation rhythm is a crucial approach to infer the intracellular 24 period 0.12 Amplitude rhythm of KaiC phosphorylation. By scrutinizing KaiC phosphoryla- tion rhythm in a wide range of Kai protein concentrations, we 22 0.1 verified that KaiA and KaiB were parameter-tuning and state- switching regulators of KaiC phosphorylation rhythm, respectively. 20 0.08 The characteristics presented in this report are essential for in silico 024681012 simulations of KaiC phosphorylation rhythm [19–27]. KaiC (μM)

4.2. A possible entrainment mechanism by oscillation of intracellular Fig. 3. KaiC phosphorylation rhythms to simulate cellular Kai protein dynamics. levels of Kai proteins Assays of in vitro KaiC phosphorylation rhythm were performed with different concentrations of KaiC and KaiB, while the ratio between the two Kai proteins was constant (1:1) in the presence of a constant concentration of KaiA (1.2 lM). (A) The Based on our in vitro observations on the period of KaiC phos- ratio of phosphorylated KaiC to total KaiC was plotted against incubation time. (B) phorylation rhythm (Figs. 1 and 3), it is expected that the angular Amplitude (square) and period (circle) are plotted against the concentration of KaiC. M. Nakajima et al. / FEBS Letters 584 (2010) 898–902 901 velocity of intracellular KaiC phosphorylation rhythm fluctuates in KaiB or KaiC to KaiA can explain entrainment of cellular circadian a circadian fashion under LL conditions (Fig. 4a), as the relative ra- rhythm of cyanobacteria, we formulated a behavior of KaiC phos- tios of KaiA levels to those of KaiB or KaiC in cyanobacterial cells phorylation rhythm under light and dark conditions and examined oscillate in circadian fashion. On the other hand, it would be ex- possible entrainment as described in Supplementary data. pected that the angular velocity of the Kai oscillator is constant As shown in Fig. 4f, our simulation predicted that when the in cells under DD conditions (Fig. 4b), because levels of all three change in angular velocity is assumed to be 10% of the mean level Kai proteins were kept at a constant level [6]. Thus, under light/ (Fig. 4d), the parametric entrainment by a symmetric LD cycle dark (LD) cycle conditions, phase progression is a combination of (L = D) could entrain the rhythm with a period ranging from 22.6 dark and light conditions (Fig. 4c). The distinction of phase pro- to 25.4 h. Similarly, if the magnitude of fluctuation is 20% of mean gression between light and dark conditions could be the basis for level, the range of entrainment could be extended to 21.2–26.8 h entrainment of the circadian clock to the external LD cycle through (Fig. 4e and g). Note that the predicted range of parametric entrain- a parametric model [28], which was based on a relationship be- ment based on experimental data we reported here was consistent tween period length of the rhythm and ambient light fluence rate with the general entrainment range of circadian rhythms to exter- (Aschoff’s rule). To test if the response of the period to the ratio of nal time cues. Therefore, regulation of the velocity of phosphoryla-

(a) LL (b) DD (c) LD

2 2 2 Angular velocity (rad/hour) Angular velocity (rad/hour) t t Angular velocity (rad/hour) t Time (hour) Time (hour) Time (hour)

(d) (e)

2 10% 2 20%

Time (hour) Time (hour) Angular velocity (rad/hour) Angular velocity (rad/hour)

(f) (g) 20 25 30 Dark period (hour) Dark period (hour) 51015202530 51015 5 1015202530 5 1015202530 Light period (hour) Light period (hour)

Fig. 4. Entrainment model of cellular KaiC phosphorylation rhythm by LD cycle. (a–c) Possible time course of angular velocity of KaiC phosphorylation rhythm. (a) In LL conditions, the angular velocity oscillates with a free-running period, s (h). (b) In DD conditions, the angular velocity is kept constant. s0 represents a free-running period under DD conditions. (c) In LD conditions, the time course of angular velocity is a combination of those under light and dark. Grey areas represent 2p phase progression of KaiC phosphorylation rhythm for one cycle. Time of light–dark cycle is indicated by white (light) and black (dark) bars at the top of each panel. (d–g) Relationship between range of entrainment and amplitude of fluctuation of angular velocity. A sine function, Asin(2pt/t)+2p/t, was used as a model of fluctuation of angular velocity in LL conditions. Two values, 10% (d) and 20% (e) for fluctuation of angular velocity (2p/t), were chosen as amplitudes in the sine function to examine the range of entrainment by LD cycles. (f) and (g) show the predicted ranges of entrainments of LD cycles using Eq. (7) in Supplementary data. 902 M. Nakajima et al. / FEBS Letters 584 (2010) 898–902 tion rhythm by Kai protein ratio could quantitatively contribute to [4] Kondo, T., Tsinoremas, N.F., Golden, S.S., Johnson, C.H., Kutsuna, S. and Ishiura, parametric entrainment of the circadian rhythm of cyanobacteria M. (1994) Circadian clock mutants of cyanobacteria. Science 266, 1233–1236. [5] Ishiura, M. et al. (1998) Expression of a kaiABC as a circadian to a wider range of LD cycles. feedback process in cyanobacteria. Science 281, 1519–1523. The entrainment model proposed here should be confirmed by [6] Tomita, J., Nakajima, M., Kondo, T. and Iwasaki, H. (2005) No transcription– in vivo experiments. A predicted asymmetric range of entrainment translation feedback in circadian rhythm of KaiC phosphorylation. Science 307, 251–254. between light and dark duration (Fig. 4f and g) would be a crucial [7] Nakajima, M., Imai, K., Ito, H., Nishiwaki, T., Murayama, Y., Iwasaki, H., Oyama, test to support our model, specifically, a longer duration of dark T. and Kondo, T. (2005) Reconstitution of circadian oscillation of period greatly expanded the range of entrainment (Fig. 4g). Unfor- cyanobacterial KaiC phosphorylation in vitro. Science 308, 414–415. [8] Dong, G. and Golden, S.S. (2008) How a cyanobacterium tells time. Curr. Opin. tunately, however, it is difficult to obtain high precision in testing Microbiol. 11, 541–546. entrainment of cyanobacterial circadian rhythm by LD cycles be- [9] Johnson, C.H., Mori, T. and Xu, Y. (2008) A cyanobacterial circadian clockwork. cause the in vivo rhythm is inevitably accompanied by acute pho- Curr. Biol. 18, R816–R825. [10] Kitayama, Y., Iwasaki, H., Nishiwaki, T. and Kondo, T. (2003) KaiB functions as tosynthetic effects by LD alteration (the so-called masking effect). an attenuator of KaiC phosphorylation in the cyanobacterial circadian clock system. EMBO J. 22, 2127–2134. 4.3. Intracellular regulation of Kai protein ratios for robust KaiC [11] Kitayama, Y., Nishiwaki, T., Terauchi, K. and Kondo, T. (2008) Dual KaiC-based phosphorylation rhythm oscillations constitute the circadian system of cyanobacteria. Genes Dev. 22, 1513–1521. [12] Iwasaki, H., Nishiwaki, T., Kitayama, Y., Nakajima, M. and Kondo, T. (2002) The standard ratio of Kai proteins in the in vitro KaiC phosphor- KaiA-stimulated KaiC phosphorylation in circadian timing loops in ylation system was evidently different from that in vivo. The rela- cyanobacteria. Proc. Natl. Acad. Sci. USA 99, 15788–15793. [13] Terauchi, K. et al. (2007) ATPase activity of KaiC determines the basic timing tive amount of KaiA to KaiC in cyanobacterial cells was about 1/20 for circadian clock of cyanobacteria. Proc. Natl. Acad. Sci. USA 104, 16377– to 1/40 [10], while it was much higher in vitro (1/3). The discrep- 16381. ancy between in vivo and in vitro ratios could be partly addressed [14] Ye, S., Vakonakis, I., Ioerger, T.R., LiWang, A.C. and Sacchettini, J.C. (2004) Crystal structure of circadian clock protein KaiA from Synechococcus elongatus. by the intracellular distributions of Kai proteins. A fraction of KaiB J. Biol. Chem. 279, 20511–20518. and KaiC were found in membrane fractions, while KaiA was pre- [15] Hitomi, K., Oyama, T., Han, S., Arvai, A.S. and Getzoff, E.D. (2005) Tetrameric dominantly localized in the cytosolic fraction [10]. As KaiC and architecture of the circadian clock protein KaiB. A novel interface for intermolecular interactions and its impact on the circadian rhythm. J. Biol. KaiB in membrane fractions may not effect phosphorylation Chem. 280, 19127–19135. rhythm, the effective intracellular ratios could be close to those ob- [16] Nishiwaki, T. et al. (2004) Role of KaiC phosphorylation in the circadian clock served in vitro. Nevertheless, the relative amount of KaiA in the system of Synechococcus elongatus PCC 7942. Proc. Natl. Acad. Sci. USA 101, 13927–13932. cytosolic fraction would be insufficient for stable KaiC phosphory- [17] Mori, T., Williams, D.R., Byrne, M.O., Qin, X., Egli, M., McHaourab, H.S., Stewart, lation rhythm and some unknown factors might be necessary to P.L. and Johnson, C.H. (2007) Elucidating the ticking of an in vitro circadian maintain KaiC phosphorylation rhythm in cells. clockwork. PLoS Biol. 5, e93. [18] Yoshida, T., Murayama, Y., Ito, H., Kageyama, H. and Kondo, T. (2009) Nonparametric entrainment of the in vitro circadian phosphorylation Acknowledgments rhythm of cyanobacterial KaiC by temperature cycle. Proc. Natl. Acad. Sci. USA 106, 1648–1653. We thank members of the Kondo Laboratory, especially H. Kon- [19] Takigawa-Imamura, H. and Mochizuki, A. (2006) Transcriptional autoregulation by phosphorylated and non-phosphorylated KaiC in do and M. Tamura for technical support. This research was sup- cyanobacterial circadian rhythms. J. Theor. Biol. 241, 178–192. ported in part by Grants-in-aid from the Ministry of Education, [20] Emberly, E. and Wingreen, N.S. (2006) Hourglass model for a protein-based Culture, Sports, Science and Technology of Japan (15GS0308 to circadian oscillator. Phys. Rev. Lett. 96, 038303. [21] Kurosawa, G., Aihara, K. and Iwasa, Y. (2006) A model for the circadian rhythm T.K.), and H.I. was supported by a Research Fellowship for Young of cyanobacteria that maintains oscillation without gene expression. Biophys. Scientists from the Japan Society for the Promotion of Science J. 91, 2015–2023. (21010517 to H.I.) [22] Mehra, A., Hong, C.I., Shi, M., Loros, J.J., Dunlap, J.C. and Ruoff, P. (2006) Circadian rhythmicity by autocatalysis. PLoS Comput. Biol. 2, e96. [23] Takigawa-Imamura, H. and Mochizuki, A. (2006) Predicting regulation of the Appendix A. Supplementary data phosphorylation cycle of KaiC clock protein using mathematical analysis. J. Biol. Rhythms 21, 405–416. [24] Miyoshi, F., Nakayama, Y., Kaizu, K., Iwasaki, H. and Tomita, M. (2007) A Supplementary data associated with this article can be found, in mathematical model for the Kai-protein-based chemical oscillator and clock the online version, at doi:10.1016/j.febslet.2010.01.016. gene expression rhythms in cyanobacteria. J. Biol. Rhythms 22, 69–80. [25] Clodong, S., Duhring, U., Kronk, L., Wilde, A., Axmann, I., Herzel, H. and Kollmann, M. (2007) Functioning and robustness of a bacterial circadian clock. References Mol. Syst. Biol. 3, 90. [26] Yoda, M., Eguchi, K., Terada, T.P. and Sasai, M. (2007) Monomer-shuffling and [1] Young, M.W. and Kay, S.A. (2001) Time zones: a comparative genetics of allosteric transition in KaiC circadian oscillation. PLoS One 2, e408. circadian clocks. Nat. Rev. Genet. 2, 702–715. [27] Eguchi, K., Yoda, M., Terada, T.P. and Sasai, M. (2008) Mechanism of robust [2] Kondo, T., Strayer, C.A., Kulkarni, R.D., Taylor, W., Ishiura, M., Golden, S.S. and circadian oscillation of KaiC phosphorylation in vitro. Biophys. J. 95, 1773– Johnson, C.H. (1993) Circadian rhythms in prokaryotes: as a reporter 1784. of circadian gene expression in cyanobacteria. Proc. Natl. Acad. Sci. USA 90, [28] Johnson, C.H., Elliott, J.A., Foster, R.G., Honma, K. and Kronauer, R.E. (2004) 5672–5676. Fundamental properties of circadian rhythms in: : Biological [3] Kondo, T. and Ishiura, M. (1994) Circadian rhythms of cyanobacteria: Timekeeping (Dunlap, J.C., Loros, J.J. and DeCoursey, P.J., Eds.), pp. 67–105, monitoring the biological clocks of individual colonies by bioluminescence. J. Sinauer, Sunderland, MA. Bacteriol. 176, 1881–1885.