Intramolecular Regulation of Phosphorylation Status of the Circadian Clock Protein KaiC

Background KaiC, a central clock protein in cyanobacteria, undergoes circadian oscillations between hypophosphorylated and hyperphosphorylated forms in vivo and in vitro. Structural analyses of KaiC crystals have identified threonine and serine residues in KaiC at three residues (T426, S431, and T432) as potential sites at which KaiC is phosphorylated; mutation of any of these three sites to alanine abolishes rhythmicity, revealing an essential clock role for each residue separately and for KaiC phosphorylation in general. Mass spectrometry studies confirmed that the S431 and T432 residues are key phosphorylation sites, however, the role of the threonine residue at position 426 was not clear from the mass spectrometry measurements. Methodology and Principal Findings Mutational approaches and biochemical analyses of KaiC support a key role for T426 in control of the KaiC phosphorylation status in vivo and in vitro and demonstrates that alternative amino acids at residue 426 dramatically affect KaiC's properties in vivo and in vitro, especially genetic dominance/recessive relationships, KaiC dephosphorylation, and the formation of complexes of KaiC with KaiA and KaiB. These mutations alter key circadian properties, including period, amplitude, robustness, and temperature compensation. Crystallographic analyses indicate that the T426 site is phosphorylatible under some conditions, and in vitro phosphorylation assays of KaiC demonstrate labile phosphorylation of KaiC when the primary S431 and T432 sites are blocked. Conclusions and Significance T426 is a crucial site that regulates KaiC phosphorylation status in vivo and in vitro and these studies underscore the importance of KaiC phosphorylation status in the essential cyanobacterial circadian functions. The regulatory roles of these phosphorylation sites–including T426–within KaiC enhance our understanding of the molecular mechanism underlying circadian rhythm generation in cyanobacteria.


Introduction
Circadian clocks are self-sustained intracellular oscillators that regulate daily rhythms of sleep/waking, metabolic activity, gene expression, and many other biological processes. Cyanobacteria are the simplest organisms known to exhibit circadian rhythms [1]. In the cyanobacterium, Synechococcus elongatus PCC 7942, a cluster of three clock genes, named kaiA, kaiB, and kaiC, encodes essential components of the circadian clock [2]; if any of these genes is knocked out, circadian rhythmicity is abolished in S. elongatus cells. The Kai proteins have been crystallized and their threedimensional structures have been determined [3]. Remarkably, a circadian oscillation of KaiC phosphorylation status can be reconstituted in vitro with just the three purified Kai proteins [4]. The relationship of this in vitro oscillator to the entire circadian system in vivo is not defined, but it is clear that the rhythm of KaiC phosphorylation is able to keep circadian time independently of transcription and translation processes in vivo [3,5], suggesting that the KaiABC oscillator is necessary and sufficient as a core oscillator for circadian rhythmicity in cyanobacteria. The circadian pacemaker in S. elongatus choreographs rhythmic patterns of global gene expression, chromosomal compaction, and the supercoiling status of DNA in vivo [6][7][8]. One hypothesis (the ''Oscilloid Model'') suggests that the regulation of gene expression is mediated through the clock regulating chromosomal topology that in turn regulates promoter activities by torsion-sensitive transcription [8,9], while an alternative (and not mutually exclusive) model suggests that the KaiABC oscillator regulates the activity of transcriptional factors such as RpaA that regulate global changes of gene expression [10].
The Kai proteins interact with each other and influence each other's activity [11][12][13][14][15][16]. These interactions lead to the formation of large complexes with KaiC as the core. These complexes mediate the KaiC oscillation between hypophosphorylated and hyperphosphorylated forms in vivo and in vitro [4,5,[17][18][19][20]. In the in vitro system, Kai protein complexes assemble and disassemble dynamically over the KaiC phosphorylation cycle [14,16]. In addition, there is an exchange of monomers among KaiC hexamers that occurs during the KaiC dephosphorylation phase [16,21]. KaiC autophosphorylation is stimulated by KaiA, whereas KaiB antagonizes the effects of KaiA on KaiC autophosphorylation [17,22,23]. On the other hand, dephosphorylation of KaiC is inhibited by KaiA, and this effect of KaiA is also antagonized by KaiB [20]. Therefore, KaiA both stimulates KaiC autophosphorylation and inhibits its dephosphorylation; KaiB antagonizes these actions of KaiA.
The crystal structure of the KaiC hexamer revealed ATP binding, insights into KaiC inter-subunit organization, and a scaffold for Kai-protein complex formation [24]. The KaiC structure also shed light on the mechanism of rhythmic phosphorylation of KaiC by identifying three potential phosphorylation sites at threonine and serine residues in KaiC at residues T426, S431, and T432 [25]. An important role for these residues was shown by the loss of rhythmicity in vivo when any of these residues were mutated to alanine (i.e., T426A, S431A, or T432A; 25). Based on mass spectrometry studies, Nishiwaki and coworkers also showed that S431 and T432 are key phosphorylation sites in KaiC [19]. Moreover, phosphorylation of S431 and T432 occurs in an ordered sequence over the cycle of the in vitro KaiABC oscillator [26,27]. These phosphorylation events are likely to mediate conformational changes in KaiC that allow interaction with KaiB and subsequent steps in the molecular cycle [3].
However, the role of the T426 residue in the phosphorylation/ conformation cycle is unclear. The crystal structure of KaiC revealed that T426 and S431 face each other across a looped region and that an extra electron density that can be attributed to the presence of a phosphate exists between T426 and S431 in some of the subunits [25]. Moreover, mutation of T426 to a non-phosphorylatible alanine abolishes rhythmicity. We therefore speculated that this putative phosphate group shuttles between these two residues, i.e. that S431 and T426 share a phosphate. On the other hand, the mass spectrometry studies did not find T426 to be phosphorylated [26,27]. In the present study, we demonstrate that T426 is a crucial site that regulates KaiC phosphorylation status and is essential for cyanobacterial circadian functions, including period, amplitude, robustness, and temperature compensation. In vivo co-expression of wild-type KaiC with KaiC mutated at T426 establishes dominant-negative and semi-dominant genetic relationships and alters temperature compensation. In vitro, asparagine/glutamate vs. alanine mutations at 426 differentially affect KaiC's biochemical properties-especially the kinetics of KaiC dephosphorylation and the association of KaiC with KaiA and KaiB into complexes. We find that KaiC undergoes an unstable phosphorylation when the S431 and T432 sites are blocked, and crystallographic analyses described in the accompanying paper [28] indicate that this labile phosphorylation status could be at the T426 site. The regulatory roles of these phosphorylation sites (S431, T432, and T426) within KaiC provide further evidence towards understanding the molecular mechanism responsible for circadian rhythm generation in cyanobacteria.

Nomenclature of KaiC mutants
In this paper, KaiC variants are sequentially shown as a superscript format in order of sites 426, 431, and 432 where the wild-type residue is in upper case and mutated residues are shown in lower case, i.e. KaiC xyz , where x = residue at position 426, y = residue at position 431, and z = residue at position 432. For example, KaiC TST is for the wild-type KaiC T426/S431/T432 (i.e. KaiC WT ), KaiC aST for KaiC T426A/S431/T432 , KaiC Tae for KaiC T426/S431A/T432E , etc.

Generation of KaiC mutants at phosphorylation sites
Mutation of KaiC at 426, 431, and 432 was performed by sitedirected mutagenesis based on template plasmids harboring either wild-type kaiABC or kaiC DNA. Briefly, the mutated kaiC strand DNA was synthesized using 12,18 thermal cycles with pfuUltraH Highfidelity DNA polymerase (Stratagene, San Diego, CA) and primers containing the desired mutations. After the nonmutated parental plasmids were digested at 30uC for 1 h with DpnI (New England Biolabs, Beverly, MA), the circular, nicked mutant dsDNA was transformed into E. coli DH5a strain and the constructs were confirmed by DNA sequencing.

Cyanobacterial Strains
For expression of a single copy of kaiC in cyanobacteria with a kaiBCp::luxAB reporter (Figs. 1 and 3C), the wild-type kaiABC cluster was replaced by a kaiABC cluster containing wild-type or mutant kaiC and a spectinomycin selection marker.
For co-expression of KaiC in cyanobacteria, the trcp::KaiC construct (with or without mutations in kaiC) was integrated into the NS II region of a wild-type strain with a psbAIp::luxAB reporter in NS I. The kaiABC-null strain was created by replacement of the kaiABC DNA region with a kanamycin resistance gene [2]. Descriptions of these strains are listed in Table S1. The clock-controlled activation of the kaiBCp::luxAB reporter was monitored by real-time measurement of luminescence as described previously [20].

Immunoblot assays for KaiC phosphorylation
After two LD 12:12 cycles, the cultures were released to constant light (LL). The cells were harvested at different phases, and the protein extracts were subjected to SDS-PAGE separation and immunoblot analyses with a mouse polyclonal antibody directed against KaiC as described previously [20,25]. For comparison of the soluble and insoluble KaiC fractions in cyanobacterial cells after a 6 h induction with IPTG (0.1 mM), cells were harvested and sonicated in a SDS-free buffer (50 mM HEPES pH 7.4, 130 mM KCl, 5% glycerol, 2.5 mM EDTA, 10 mg/ml aprotinin, 10 mg/ml leupeptin, 2 mg/ml pepstatin and 5 mM phenylmethylsulfonyl fluoride) and microcentrifuged for 3 min. The supernatant was collected and the pellet was washed three times in the same buffer. The pellet (insoluble fraction) and supernatant (soluble fraction) were then separately prepared for SDS-PAGE and immunoblotting.

P-labeled KaiC phosphorylation in vitro
Purified proteins (200 ng/ml) of wild-type KaiC TST or mutant KaiCs (aST, Tae, aaa, or aae) were incubated at 30uC in Reaction Buffer (RB; 150 mM NaCl, 5 mM MgCl 2 , 1 mM ATP, 0.5 mM EDTA, 20 mM Tris-HCl, pH 8.0] containing 0.4 mCi/ml [c-32 P]ATP. At each time point, 10 ml aliquots were taken, mixed with 10 ml of 2X SDS-PAGE loading buffer, and stored at 220uC. Before SDS-PAGE, the samples were either (i) not heated above room temperature, (ii) heated to 72uC for 10 min, or (iii) heated to 100uC for 10 min as indicated. Following SDS-PAGE, gels were washed and fixed in 20% MeOH and 7% HOAc overnight, stained with colloidal Coomassie Brilliant Blue (CBB) overnight, destained in MQ-water for several hours, and dried at 68uC for 4 hours. The dried gels were exposed to a phosphorimager screen, which was scanned by a Typhoon Imager (GE Healthcare).

KaiB/KaiA antagonism of KaiC phosphorylation/ dephosphorylation in vitro
Purified wild-type or mutant KaiC proteins (100 ng/ml) were incubated at 30uC in RB with or without 25 ng/ml of the purified KaiA and/or KaiB at 30uC in a time course up to 24 h. The in vitro KaiC phosphorylation/dephosphorylation patterns were examined by SDS-PAGE analysis.

Analysis of stable KaiBC and KaiABC complexes
To analyze the formation of stable complexes between KaiB and KaiC, KaiB (50 ng/ml) and KaiC (200 ng/ml) were mixed and incubated at 30uC in a volume of 100 ml RB. 16 ml aliquots of each mixture were collected at each time point and mixed with 5X non-denaturing buffer (0.312 M Tris-HCl, pH 6.8, 0.05% CBB, 50% glycerol) and frozen immediately at 280uC. Native gel electrophoresis (7.5% polyacrylamide) was performed at 4uC to separate protein complexes. Gels were stained with colloidal CBB, and gel images were digitally captured with the Bio-Rad Gel Doc XR system (Bio-Rad, Hercules, CA).

Results and Discussion
Mutations at T426 that prevent phosphorylation abolish rhythmicity in vivo We previously reported that mutation of T426 to the nonphosphorylatible residue alanine abolished rhythmicity [25]. We confirmed in this study that replacement of wild-type KaiC (i.e. KaiC TST ) with KaiC aST not only abolishes rhythmicity, but also constitutively represses reporter activity in vivo (Fig. 1A). If T426 is replaced with glutamate that could serve as a mimic of a phosphorylated residue (KaiC eST ), the cells are also arhythmic. Therefore, neither constitutive activation at T426 (KaiC eST ) nor constitutive inactivation (KaiC aST ) can support rhythmicity. Kitayama and co-workers reported that a KaiC with glutamate at both 431 and 432 sites (KaiC Tee ) was able to support rhythmicity, albeit with a damped, long period rhythm [29], implying that phosphorylatibility at T426 could be sufficient to allow rhythmicity if both 431 and 432 are constitutively activated by glutamate residues. To test if phosphorylatibility at T426 is sufficient to allow rhythmicity when 431 cannot be phosphorylated, we examined a double mutant KaiC Tae , where 432 is activated by glutamate, but 431 is a non-phophorylatible and non-activated alanine. From the crystal structure, T426 can be phosphorylated in KaiC Tae in vitro ( [28], and see below). However, KaiC Tae cannot support rhythmicity in vivo (Fig. 1A). Because KaiC Tee allows rhythmicity in vivo [29] but KaiC Tae does not, clearly 431 needs to be activated by phosphorylation or as a glutamate. It is likely that 431 must be activated by phosphorylation or as a glutamate in order for KaiB to bind KaiC and initiate the dephosphorylation cycle [3,26,27].
The electron density between S431 and T426 in the crystal structure of KaiC TST [24,25] is placed such that the center of the electron density (the putative phosphorylation) is 1.6 Å from the coxygen of S431, which is appropriate for a covalent bond between S431 and the putative phosphate. The center of this electron density is 4.5 Å from the c-oxygen of T426, which also places the phosphate within hydrogen bonding distance from T426. Therefore, to test if the arhythmicity of KaiC aST is due to the inability of (i) A426 to form a hydrogen bond to the phosphate group attached to S431, or (ii) A426 to be phosphorylated (and therefore that T426 could be a bona fide phosphorylation site), we mutated the 426 residue to asparagine, which can accept or donate hydrogen bonds (also, asparagine is roughly isosteric with threonine), but which cannot be phosphorylated. Therefore, the predictions are: (i) if the function of 426 is to interact with the phosphorylated residue S431 via a hydrogen bond, then KaiC nST should be rhythmic, whereas (ii) if the 426 site needs to be phosphorylatible, then KaiC nST will be dysfunctional as well. Fig. 1A shows that KaiC nST is also arhythmic in vivo; the results with KaiC aST , KaiC eST , and KaiC nST support the conclusion that regulation of phosphorylation at 426 is crucial and that the function of threonine at 426 is not merely to hydrogen bond with a phosphate attached to S431.

Phosphorylation patterns of T426 mutants in vivo
When extracts of S. elongatus cells are immunoblotted with an antibody to KaiC, the phosphorylation states of KaiC can be visualized by mobility shifts that show hyperphosphorylation during the subjective night (hours 12-24 in LL for KaiC TST ) and hypophosphorylation in the subjective day (hours 24-36 for KaiC TST ) [3,26,27]. A 54 h time course of phosphorylation patterns in vivo for KaiC TST vs. mutant KaiCs reveals that none of the KaiC mutations at 426, 431, or 432 (including KaiC aST ) can sustain a rhythm of KaiC phosphorylation (Fig. S1). While none of the KaiC mutations related to residue 426 show rhythms of phosphorylation, they do affect the overall phosphorylation patterns in vivo (Figs. 1B, 1C). KaiC nST and KaiC eST are relatively hyperphosphorylated, suggesting that asparagine or glutamate at 426 might inhibit dephosphorylation at S431 and T432. Note the complete absence of the lowest KaiC band (i.e., non-phosphorylated KaiC) in KaiC eST . On the other hand, KaiC Tae is relatively hypophosphorylated, indicating a weak ability to stably phosphorylate if the key residue 431 is not phosphorylatible (see below). The pattern depicted for KaiC aST in Fig. 1B is particularly interesting. First, KaiC aST clearly shows five bands (implying at least five phospho-states of KaiC in vivo), whereas usually only four bands are observed over the circadian cycle in vitro [26,27]. Second, the KaiC aST pattern shows both hyperphospho-states (upper bands) and hypophospho-states (lower bands) but the middle band is missing or greatly reduced in KaiC aST . It is tempting to speculate that this middle band (third from the top) indicates a phospho-state in which T426 is phosphorylated and which has been captured in the crystal structure of KaiC Tae [28].

Co-expression of 426 mutants with wild-type KaiC in vivo
Another way to test the significance of the T426 site is to coexpress the KaiC mutants at this site with KaiC TST in vivo and assess whether the mutant KaiCs affect or disrupt the rhythmicity imparted by the wild-type KaiC. All versions of KaiC (KaiC TST and mutant versions) show similar levels of expression of soluble KaiC when expressed from an ectopic site under the control of a trc promoter (trcp; 20) (Fig. S2). When wild-type KaiC (KaiC TST ) is expressed in a wild-type strain in which the endogenous kaiABC cluster is present, rhythmicity is preserved with a slightly longer period from ,26.0 h to ,27.2 h at 30uC (Fig. 2). When the KaiC nST , KaiC eST , and Kai Tae mutants are co-expressed with KaiC TST under the same conditions, the period is lengthened very significantly from ,26 h to ,30-34 h (p,0.01; Fig. 2B). Not only the period, but also temperature compensation is affected by the co-expression of these mutants because the Q 10 increases from 1.01 to ,1.3 (Table 1). This result means that co-expression of these 426 site mutants alters the properties of the oscillating system, but does not prevent KaiC TST from maintaining the rhythm. On the other hand, when KaiC aST is co-expressed with KaiC TST under the same conditions, the T426A mutation always exerts a dominant negative effect on the luminescence rhythms ( Fig. 2A) independent of the level of expression of KaiC aST . Therefore, the KaiC aST mutant interferes with the function of the endogenous KaiC TST .
A possible trivial explanation for the data of Figure 2 might be that KaiC TST or KaiC aST are soluble and active when expressed intracellularly, but that KaiC nST , KaiC eST , and Kai Tae aggregate and become inactive. However, this is not the case. All the versions of KaiC have a significant soluble pool when expressed in vivo (the proportion of soluble to insoluble KaiC is essentially the same for the mutants as for KaiC TST , Fig. S2). Also, all the mutant versions of KaiC have an activity in vivo because they either lengthen the period significantly or abolish rhythmicity altogether in the coexpression experiment (Fig. 2). Because rhythmic phosphorylation of KaiC is critical to the clock's operation [3,17,26,27], perhaps the effect of the 426 site mutations can be explained by an effect on phosphorylation patterns. Figure 3 depicts the total KaiC phosphorylation patterns in vivo at two circadian phases in the coexpression experiment. There are significant day/night changes in phosphorylation of the total KaiC pool for coexpression of KaiC TST , KaiC nST , KaiC eST , and Kai Tae . On the other hand, the coexpression of KaiC aST results in constitutive hypophosphorylation of the total KaiC pool (including the endogenous KaiC TST ). Beyond these overall changes in phosphorylation levels, the patterns themselves are not additionally informative. Therefore, the phosphorylation patterns under the KaiC coexpression protocol in vivo are consistent with the data obtained by monitoring the luminescence rhythms (Fig. 2) in that the combinations that show a luminescence rhythm correlate with those that show day/ night differences in phosphorylation patterns. The data of Figs. 1-3 support the conclusion that T426 serves as a crucial site for proper clock function despite the fact that the mass spectrometry approach did not indicate a participation of T426 in the phosphorylation sequence [26,27].

Phosphorylation in vitro of KaiC mutants at residue 426
However, the crystallography results of the accompanying manuscript indicate that T426 can auto-phosphorylate in KaiC Tae crystals, i.e., when 432 is activated by a glutamate and 431 is an alanine and therefore both site 431 and site 432 are unavailable for phosphorylation [28]. In an attempt to demonstrate under other conditions if T426 can be stably phosphorylated, we performed in vitro phosphorylation assays in the presence of ATP with a 32 P radiolabel at the c-phosphate position. As shown in Fig. 4A, KaiC TST , KaiC aST , and KaiC nST robustly label with 32 P at 30uC for 24 h presumably because S431 and T432 are available for phosphorylation in these proteins. The control proteins KaiC aaa and KaiC Taa show very low labeling with 32 P, while the control protein KaiC aae has labeling above background (presumably residues other than 426, 431, and 432 can sometimes phosphorylate). However, KaiC Tae does not show labeling above background by this assay when the samples are heated before loading prior to SDS-PAGE ( Fig. 4A and left panel of Fig. 4B). On the other hand, when the samples are not heated before loading on the SDS-PAGE gel, there is a low but significant labeling with the 32 P radiolabel of the KaiC Tae mutant protein that is slightly above the level in KaiC Taa (Fig. 4B, right panel). The crystallographic results indicate that KaiC Tae can be phosphorylated at T426 under specific conditions [28], but the results of Fig. 4 indicate that other conditions do not promote stable phosphorylation. Moreover, the comparison of results obtained when the samples were heated vs. not heated (Fig. 4B) prior to SDS-PAGE analysis suggest that phosphorylation on residues other than S431 and T432 (possibly at T426) is labile.

Mutations at T426 affect KaiC dephosphorylation kinetics
Our original hypothesis for the role of T426 was that a phosphate could shuttle between T426 and S431, either as a stable phosphorylation or as a labile interaction [25]. The crystal structure of KaiC Tae supports the hypothesis that T426 can autophosphorylate [28], while the data of Fig. 4B supports the hypothesis that the putative phosphate-T426 interaction is unstable. A different possible role for T426, however, is that this threonine acts to stabilize/modulate the phosphorylation on S431 and thereby regulates the phospho-S431 site that is key to the binding of KaiB and subsequent steps in the dephosphorylation sequence [26,27]. Figure 5 provides data that support this alternative hypothesis. Figs. 5A and 5B show that KaiC TST and KaiC aST auto-dephosphorylate at approximately the same rates. Conversely, replacement of threonine with asparagine at 426 (i.e., KaiC nST ) very significantly hinders the rate of dephosphorylation. KaiA stimulates the phosphorylation of all three forms of KaiC, while KaiB alone has little effect on the rate of dephosphorylation for either KaiC TST or KaiC aST (and KaiB has no effect on KaiC nST , which remains hyperphosphorylated). However, when all three Kai proteins are put together, there is a striking difference between KaiC TST and KaiC aST . For KaiC TST , KaiB antagonizes KaiA's stimulation of phosphorylation and therefore KaiC TST dephosphorylates as reported previously [20]. In the cases of  KaiC aST and KaiC nST , however, KaiB is incapable of antagonizing KaiA's effect and these versions hyperphosphorylate in the presence of KaiA and KaiB in vitro (Fig. 5A,B).

Mutations at T426 affect formation of complexes between KaiC and KaiA/KaiB
It might be imagined that KaiB's inability to antagonize KaiA's action on KaiC aST or KaiC nST is because KaiB cannot bind these versions of KaiC. Fig. 5C shows that this could be the case for KaiC nST -by a native gel analysis, KaiC nST does not form a stable complex with KaiB. However, the same analysis indicates that both KaiC TST and KaiC aST form stable complexes with KaiB ( Fig. 5C; the KaiC aST /KaiB complex forms more slowly than the KaiC TST /KaiB complex). When mixed with both KaiA and KaiB, KaiC TST forms stable ABC complexes, but neither KaiC aST nor KaiC nST can form stable complexes with KaiA and KaiB (Fig. 5D). Taken together, the data of Fig. 5 show that KaiB binds to both KaiC TST and KaiC aST , but is incapable of antagonizing KaiA's hyperphosphorylation of the KaiC aST mutant. KaiC nST cannot bind KaiB in stable complexes and is deficient in its ability to auto-dephosphorylate. When both KaiA and KaiB are present, KaiC aST and KaiC nST are incapable of forming stable ABC complexes (Fig. 5D) and are unable to dephosphorylate normally (Fig. 5A & 4B). These data indicate that substitution of threonine at site 426 with non-phosphorylatible residues prevents normal interactions with KaiA and KaiB and also prevents the proper regulation of KaiC phosphorylation status.
Clearly, S431 and T432 are key phosphorylation sites on KaiC [19,[24][25][26][27]. While the crystallographic and mutational analyses support a crucial role for T426 [24,25], e.g., as a potential additional phosphorylation site, mass spectrometry measurements did not detect a stable phosphorylation at the T426 residue [26,27]. However, if phosphorylation at T426 is labile, the phosphate group might have been lost during preparation of the proteins for mass spectrometry. Further crystallographic analyses support the earlier interpretation  that T426 can be phosphorylated [28], although the data shown in Fig. 5D indicates that phosphorylation in the KaiC Tae mutant (possibly at T426) may be an unstable modification. Substitutions at 426 with either alanine or asparagine nonetheless show in vitro that T426 is crucial to the action of KaiB, either its binding (asparagine substitution) or its activity once bound (alanine substitution). In vivo, substitution of threonine with alanine at 426 shows that rhythmicity is abolished when the mutant KaiC is expressed alone (Fig. 1) and disrupts the in vivo rhythm when the mutant KaiCs are co-expressed with wild-type KaiC TST (Fig. 2). Therefore, we demonstrate here that 426 is a crucial site for the regulation of KaiC phosphorylation status and that a threonine residue at this site is essential for proper clock function. Our findings strongly support that these KaiC phosphorylation sites coordinately regulate key circadian properties in cyanobacteria, including period, amplitude, robustness, and temperature compensation.

Supporting Information
Table S1 Synechococcus strains used in this study. ''Strain'' is the name of the strain based on the residues at the 426, 431, and 432 positions and the promoter used. ''Description'' is the full description of the strain. ''Reporter'' is the promoter used in the luxAB fusion and the neutral site in which the reporter is inserted. NSI = neutral site I, NSII = neutral site II. Antibiotic resistances: Spr = spectinomycin, Cmr = chloramphenicol, Kmr = kanamycin.  Figure S2 Comparison of KaiC expression between soluble and insoluble portions from the wild type strain co-expressing wild-type or mutant KaiCs. After 12 h darkness, the cultures of the wild-type strain co-expressing trcp-driven wild-type KaiC (KaiC-TST) or mutant KaiCs (as indicated) were released to LL, and 0.1 mM of IPTG was added. The cells were incubated for 6 h (from LL0 to LL6) before harvesting. The soluble extracts (upper panel), made in the SDS-free buffer, and the insoluble extracts (lower panel), made from the pellets in the SDS-containing buffer, were subjected to immunoblot analysis. Found at: doi:10.1371/journal.pone.0007509.s003 (0.41 MB TIF)