Phosphorylation at Thr432 induces structural destabilization of the CII ring in the circadian oscillator KaiC

Transcription

1 Phosphorylation at Thr432 induces structural destabilization of the CII ring in the circadian oscillator KaiC Katsuaki Oyama 1, Chihiro Azai 2, Jun Matsuyama 1 and Kazuki Terauchi 1,2 1 Graduate School of Life Sciences, Ritsumeikan University, Kusatsu, Japan 2 College of Life Sciences, Ritsumeikan University, Kusatsu, Japan Correspondence K. Terauchi, College of Life Sciences, Ritsumeikan University, Kusatsu, Shiga , Japan Fax: Tel: terauchi@fc.ritsumei.ac.jp (Received 19 October 2017, revised 28 November 2017, accepted 9 December 2017, available online 29 December 2017) doi: / Edited by Miguel De la Rosa KaiC is the central oscillator protein in the cyanobacterial circadian clock. KaiC oscillates autonomously between phosphorylated and dephosphorylated states on a 24-h cycle in vitro by mixing with KaiA and KaiB in the presence of ATP. KaiC forms a C 6 -symmetrical hexamer, which is a double ring structure of homologous N-terminal and C-terminal domains termed CI and CII, respectively. Here, through the characterization of an isolated CII domain protein, CII KaiC, we show that phosphorylation of KaiC Thr432 destabilizes the hexameric state of the CII ring to a monomeric state. The results suggest that the stable hexameric CI ring acts as a molecular bundle to hold the CII ring, which undergoes dynamic structural changes upon phosphorylation. Keywords: ATPase; circadian clock; hexamer; KaiC; phosphorylation Cyanobacteria are the simplest organisms known to exhibit a circadian rhythm [1]. The circadian rhythm of KaiC phosphorylation in the cyanobacterium Synechococcus elongatus PCC 7942 proceeds in the absence of any transcription translation feedback [2], and the oscillation of KaiC phosphorylation can be reconstituted in vitro by mixing KaiA, KaiB, and KaiC in the presence of ATP [3]. KaiC is a duplicate P-loop ATPase consisting of the N-terminal and C-terminal domains, CI and CII, respectively (Fig. 1A), and forms a C 6 -symmetrical hexamer in the presence of ATP [4,5]. Thus, the KaiC hexamer is regarded as a double ring of CI and CII hexamers (Fig. 1B). KaiC shows a very low level of ATPase activity that correlates with a period of circadian oscillation [6]. In addition, KaiC exhibits autokinase and autophosphatase activities on two residues, Ser431 and Thr432, in the CII domain [7 9]. These contiguous residues that can be phosphorylated are located proximal to the interfaces between the protomers in the hexameric CII ring [5]. The relationship between the activities of ATPase and autokinase/autophosphatase is one of the main targets of current studies to understand the molecular mechanism underlying the circadian oscillation of KaiC. To investigate the role of ATPase activity in the CI domain, a KaiC variant, E77Q/E78Q-KaiC, in which conserved catalytic glutamate residues in the CI domain were substituted for glutamine residues, was characterized [10 13]. This variant impaired the ATPase activity of CI as expected, and resulted in the loss of the circadian oscillation of phosphorylation; nonetheless, the autokinase and autophosphatase activities were fully maintained [11]. The ATPase activity of the CI domain thus modulates the autokinase and autophosphatase activities of the CII domain to generate the circadian oscillation of phosphorylation of KaiC. However, tight and complex coupling of the structural changes between the CI and CII domains makes it difficult to determine the individual roles of each domain from the reaction course in the full-length KaiC protein. Abbreviation DTT, dithiothreitol; nlc/ms/ms, nano-scale high-performance liquid chromatography coupled with tandem mass spectrometry. 36 FEBS Letters 592 (2018) ª 2017 Federation of European Biochemical Societies

2 K. Oyama et al. Phosphorylation in KaiC CII ring Fig. 1. KaiC structure. (A) Schematic illustrations of the overall structure, consisting of duplicated domains of KaiC, the N-terminal CI domain, and the C-terminal CII domain. CI KaiC and CII KaiC consist of a single domain, CI or CII, respectively. P-loops, catalytic carboxylates (Glu) and two phosphorylation sites (Ser and Thr) are indicated. (B) KaiC hexamer with one protomer highlighted. Each protomer consists of the CI and CII domains. Both domains bind nucleotides, but only the CII domain is phosphorylated at residues Ser431 and Thr432. Six protomers of KaiC assemble into a hexamer, and the CI and CII domains form a CI and CII ring, respectively. In this study, to reveal the role of the CI domain in the oscillation of phosphorylation of KaiC, we characterized two truncated variants of KaiC, named CI KaiC and CII KaiC, which consist solely of the CI domain or CII domain, respectively. We found that CII KaiC retained autokinase activity while losing autophosphatase activity, and that a hexameric structure of the CII domain is destabilized upon phosphorylation at Thr432. Materials and methods Plasmids construction for CI KaiC and CII KaiC proteins Two DNA fragments corresponding to the CI and CII domains covering Thr2 to Phe247 and Leu249 to Ser519 of KaiC, respectively, were amplified by PCR from the Synechococcus elongatus PCC 7942 kaic gene. The PCR fragments were cloned into the BsaI site of pask-iba-5plus (IBA, St. Louis, MO, USA). In this overexpression system, the CI and CII domain proteins (CI KaiC and CII KaiC, respectively) were expressed as fusion proteins containing the Strep-tag sequence (MASWSHPQFEKGA) at the N termini (Fig. 1A). Plasmids for the overexpression of the two CII KaiC variants, CII DE KaiC and CII AA KaiC, were constructed independently of the plasmid for CII KaiC as follows. The region corresponding to the KaiC CII domain (Leu249-Ser519) was amplified by PCR using two plasmids as templates that were used for the overexpression of KaiC-DE and KaiC- AA proteins in the previous work [13]. The PCR fragments were cloned into the BsaI site of pask-iba-5plus to overexpress CII DE KaiC and CII AA KaiC as N-terminal Strep-tag fusion proteins. CII DE KaiC contained the substitutions Ser to Asp and Thr to Glu, which correspond to the phosphorylation sites Ser431 and Thr432, respectively, in KaiC. CII AA KaiC contained Ala substitutions at the same Ser and Thr residues. All oligonucleotide primers are listed in Table S1. Purification of recombinant Kai proteins KaiA, KaiB, and KaiC with N-terminal Strep-tag were expressed and purified as previously described [13]. CI KaiC was expressed and purified as for KaiC. All steps were performed on ice or at 4 C. The method was modified for CII KaiC purification. Following transformation of Escherichia coli DH5a cells with the expression plasmid, the cells were cultured in LB medium containing 100 lgml 1 ampicillin at 37 C. Overexpression of CII KaiC was induced by the addition of 0.2 lgml 1 anhydrotetracycline, and cells were collected 3 h after induction. Cells were suspended in 20 mm Tris-HCl (ph 8.0) buffer containing 150 mm NaCl, 10 mm ATP, 5 mm MgCl 2,2mM dithiothreitol (DTT), and 0.2 mm phenylmethylsulfonyl fluoride and were then disrupted by sonication (SONICS, VCX-130 or BRANSON, Model 450DA). The obtained homogenates were centrifuged at g and the supernatants were applied to a Strep-Tactin Sepharose column (IBA). After adsorption of CII KaiC, the columns were washed with 20 mm Tris-HCl (ph 8.0) buffer containing 150 mm NaCl, 1 mm ATP, 5 mm MgCl 2 and 2 mm DTT. CII KaiC was eluted with the buffer containing 2.5 mm D- desthiobiotin. For further purification and removal of D- desthiobiotin, the purified CII KaiC fraction was applied to a HiPrep 16/60 Sephacryl S-200 HR column (GE Healthcare, Uppsala, Sweden) controlled by an AKTA prime system (GE Healthcare) for gel filtration chromatography. All chromatographic steps were performed at 4 C. Protein concentrations were determined according to the Bradford method using protein assay reagent (Wako, Osaka, Japana) with bovine serum albumin (Bio-Rad, Hercules, CA, USA) as standard. Assay for phosphorylation of KaiC by SDS/PAGE KaiC and CII KaiC were incubated in 20 mm Tris-HCl (ph 8.0) buffer containing 150 mm NaCl, 5 mm MgCl 2 in the presence of ATP (3 or 10 mm). KaiC and CII KaiC samples were subjected to 10%T (0.07%C) and 15%T (0.1%C) FEBS Letters 592 (2018) ª 2017 Federation of European Biochemical Societies 37

3 Phosphorylation in KaiC CII ring K. Oyama et al. SDS/PAGE, respectively, followed by staining with Coomassie Brilliant Blue. Relative quantities of phosphorylated KaiC to total KaiC in each sample were determined using densitometric analyses with IMAGEJ software [14]. Western blot analysis Six hundred nanograms of each protein were separated using 15%T (0.1%C) SDS/PAGE and transferred onto a Immobilon-P transfer membrane (Merck Millipore, Temecula, CA, USA). The membranes were washed with PBST (phosphatebuffered saline containing 0.1% Tween 20), then incubated with 3% BSA at room temperature for 1 h. Next, they were washed with PBST followed by incubation with Strep-Tag II monoclonal antibody (Novagen, Madison, WI, USA; 1 : 5000) for 1 h at room temperature, followed by a final wash with PBST. Rabbit anti-mouse IgG H&L (HRP) (Abcam, Cambridge, UK) was used at a dilution of 1 : for detection, and the Strep-tagged proteins were detected using Chemi-Lumi One Super chemiluminescent substrate (Nacalai Tesque, Kyoto, Japan). Chemiluminescence was measured using a ChemiDoc XRS+ imaging system (Bio-Rad). Assay of KaiC oligomerization by gel filtration chromatography CII KaiC was applied to a COSMOSIL 5Diol-300-II Packed Column (Nacalai Tesque) equilibrated with 20 mm Tris- HCl (ph 8.0) buffer containing 150 mm NaCl, 10 mm ATP, and 5 mm MgCl 2. Proteins were separated at room temperature or 30 C with a flow rate of 1 mlmin 1 and detected by absorbance at 280 nm, using an HPLC system (LC-10AVP; Shimadzu, Kyoto, Japan). Results Preparation of CII KaiC To examine whether the CII domain exhibits autokinase and autophosphatase activities without the CI domain, a truncated KaiC protein variant, CII KaiC, consisting of only the CII domain corresponding to residues of KaiC (Fig. 1A), was prepared. In addition, the CI KaiC protein consisting of the CI domain corresponding to residues of KaiC was also prepared. Both CI KaiC and CII KaiC were overexpressed in E. coli as Strep-tagged fusion proteins, with their molecular masses with the N-terminal Strep-tag II predicted to be 29.1 and 31.7 kda, respectively. Both proteins were purified by Strep-tag affinity chromatography followed by gel filtration chromatography. CI KaiC eluted at around the 44.5-mL volume from the gel filtration column. The apparent molecular mass of this fraction was 162 kda, which was in good agreement with the predicted molecular mass of a hexamer (172.6 kda), indicating that CI KaiC forms a hexamer in the solution (Fig. 2A; upper panel). CII KaiC eluted in two peaks at 45 and 64 ml (Fig. 2A, lower panel), with apparent molecular masses of 157 and 31 kda, respectively. The apparent molecular mass was with 157 kda lower than that predicted for hexameric CII KaiC (190.2 kda), while the other value, 31 kda, was in good agreement with the monomeric form (31.7 kda), indicating that CII KaiC formed an oligomer with a molecular mass of about 157 kda while partially existing as a monomer. Although 157 kda is rather close to that corresponding to a pentameric form (158.5 kda), we regarded the oligomer of CII KaiC as a hexamer because even the hexamer of full-length KaiC behaved as a protein with a molecular mass lower than that predicted in similar gel filtration chromatography assays [13]. CI KaiC and the two CII KaiC fractions were analyzed by SDS/PAGE. Figure 2B shows that the CII KaiC in the monomeric fraction migrated as a doublet of upper and lower bands corresponding to approximately 35 kda, which is consistent with the predicted molecular mass. The CII KaiC in the oligomeric fraction migrated as a single band that was accompanied by a very faint upper band, which possibly corresponded to the upper band of the monomeric fraction (Fig. 2B). Western blot analysis with an anti-strep-tag antibody confirmed that all the bands on SDS/PAGE bore the Strep-tag (Fig. 2C). Analogous to full-length KaiC, which shows a doublet band of phosphorylated and nonphosphorylated polypeptides on SDS/PAGE [9], the upper and lower bands of monomeric CII KaiC on SDS/PAGE were expected to correspond to the phosphorylated and nonphosphorylated polypeptides, respectively. To confirm this, purified CII KaiC was treated with a phosphatase (the Mn 2+ -dependent protein phosphatase of k bacteriophage). The upper band of CII KaiC as well as that of full-length KaiC disappeared following treatment, and were accompanied by an increase in the lower band intensity (Fig. S1). Therefore, we concluded that the upper CII KaiC band corresponded to the phosphorylated form of CII KaiC. Coupling between phosphorylation and dissociation of CII KaiC hexamer The upper band of monomeric CII KaiC was thicker than that of the lower band (Fig. 2B), indicating that monomeric CII KaiC was highly phosphorylated. The phosphorylation reactions of hexameric KaiC occur on each protomer interface, in which the ATP-binding loops and catalytic residues for the autokinase activity are located [11,15]. We hypothesized that hexameric 38 FEBS Letters 592 (2018) ª 2017 Federation of European Biochemical Societies

4 K. Oyama et al. Phosphorylation in KaiC CII ring Fig. 2. Purification of CII KaiC. (A) CI KaiC (upper panel) and CII KaiC (lower panel) were purified by gel filtration chromatography using Sephacryl S-200 HR columns. Arrowheads at the top of the panel indicate void volumes and the apparent molecular mass of standards: b amylase, 200 kda; alcohol dehydrogenase, 150 kda; albumin, 66 kda; and carbonic anhydrase, 29 kda. Purified CI KaiC and CII KaiC were subjected to SDS/PAGE (B) and western blot analysis with anti-strep-tag antibody (C). Upper and lower bands in each lane correspond to phosphorylated (P-CII KaiC ) and nonphosphorylated (NP-CII KaiC ) forms. CII KaiC retains phosphorylation activity and dissociates into monomers upon phosphorylation. To confirm this, the hexameric CII KaiC fraction was incubated in the presence of 10 mm ATP at 30 C to promote phosphorylation, and analyzed by SDS/PAGE and gel filtration chromatography (Fig. 3A, B). SDS/PAGE analysis showed that the lower band decreased and had disappeared by 24 h and that the upper band increased in a reciprocal manner to be almost a single band by 24 h, indicating that hexameric CII KaiC was phosphorylated during incubation with ATP. In contrast to CII KaiC, SDS/PAGE analysis of full-length KaiC showed that the lower band increased while the upper band decreased, indicating that the ratio of phosphorylated KaiC to total KaiC decreased at 30 C as previously reported (Fig. 3A) [2]. The gel filtration chromatography revealed that hexameric CII KaiC converted into a monomer during the 12-h incubation at 30 C, unlike KaiC, which remained in its stable hexameric form (Fig. 3B). These data strongly suggest that phosphorylation and hexamer dissociation of CII KaiC are closely coupled. We analyzed two CII KaiC variants, CII AA KaiC and CII DE KaiC, that carry two substitutions at the phosphorylation residues (Fig. 3C). CII DE KaiC is a phosphorylated mimic in which the Ser and Thr residues corresponding to the phosphorylation sites of KaiC were replaced with Asp and Glu residues, respectively. CII AA KaiC is a nonphosphorylated mimic in which both the Ser and Thr residues were replaced with two Ala residues. In gel filtration chromatography, CII- KaiC DE eluted only in one fraction with an apparent DE molecular mass of 26 kda, indicating that CII KaiC existed as a monomer. In contrast, CII AA KaiC eluted much earlier than CII DE KaiC in a fraction with an apparent molecular mass of 141 kda, which is in good agreement with that of a pentamer or hexamer. These results suggested that phosphorylated CII KaiC is a monomer that does not form a hexamer. Phosphorylation of CII KaiC may destabilize the hexamer to dissociate it into monomers. To identify which of the phosphorylation sites in CII KaiC is related to hexamer dissociation, we analyzed the phosphorylated polypeptide by nlc/ms/ms. First, CII KaiC hexamer incubated for 24 h at 30 C was applied to Phos-tag PAGE, in which phosphorylated polypeptides migrate much slower than nonphosphorylated ones in the presence of MnCl 2 [16] (Fig. S2, panel A). The well-resolved band of phosphorylated CII KaiC was excised and digested by a combination of two site-specific proteases, trypsin and AspN, and the digested peptides were subjected to nlc/ms/ms. Only two phosphorylated peptides were observed with high intensities: DSHISpTIT (retention FEBS Letters 592 (2018) ª 2017 Federation of European Biochemical Societies 39

5 Phosphorylation in KaiC CII ring K. Oyama et al. Fig. 3. Phosphorylation and multimeric structure of CII KaiC. (A) Phosphorylation profiles of CII KaiC and KaiC. CII KaiC and KaiC (0.4 mgml 1 ) were incubated at 30 C in the presence of 10 mm ATP. Aliquots of the reaction mixture were subjected to SDS/PAGE (upper panel: CII KaiC, middle panel: KaiC). The upper and lower bands in each lane correspond to the phosphorylated (P-CII KaiC, P-KaiC) and nonphosphorylated (NP-CII KaiC, NP-KaiC) forms. The bands were analyzed by densitometry and the ratio of the phosphorylated forms to total protein of CII KaiC and KaiC were plotted as symbols and, respectively (lower panel). (B) Gel filtration chromatography of CII KaiC and KaiC. CII KaiC and KaiC before (black line) and after (red line) 12-h incubation were subjected to gel filtration chromatography using a column packed with silicabased resin (Fig. S3). Arrowheads at the top of the panel show elution volumes corresponding to the hexamer and monomer forms. (C) Gel filtration chromatography of CII KaiC AA and CII KaiC DE. CII KaiC AA (black line) and CII KaiC DE (red line) were purified by gel filtration chromatography using Sephacryl S-200 HR column. Arrowheads at the top of the panel are indicated as in Fig. 2A. time = 1470 s, m/z = ) (Fig. S2, panel B) and DSHISpTITDTIILLQYVEIR (retention time = 2530 s, m/z = ) (Fig. S2, panel C). The latter peptide fragment might have been produced by partial digestion by AspN. The phosphate group occurred on the same residue, Thr, corresponding to Thr432 of KaiC. We could not observe any peptides with nonphosphorylated Thr432 or with phosphorylated Ser431. These data strongly suggest that hexameric CII KaiC is phosphorylated at the specific Thr residue at 30 C, but that no further phosphorylation at other Ser or Thr residues occurs in CII KaiC. Therefore, the destabilization of hexameric CII KaiC would be caused by phosphorylation of the Thr residue corresponding to KaiC Thr432. Concentration-dependent dissociation of hexameric CII KaiC When the hexamer fraction of CII KaiC (Fig. 2A, lower panel) was diluted to a protein concentration of 0.4 mgml 1 to be reanalyzed by gel filtration chromatography, a significant portion of it eluted in the monomer fraction in addition to the hexamer fraction (Fig. 3B), implying that dilution causes dissociation of the hexameric form. To confirm this, isolated hexameric CII KaiC as well as KaiC and CI KaiC were diluted to concentrations of 0.1, 0.2, and 0.4 mgml 1 at 30 C and then immediately analyzed by gel filtration chromatography (Fig. 4A). Hexameric CII KaiC yielded peaks corresponding to the hexamer and monomer forms at all three protein concentrations. The ratio of monomer to hexamer at 0.1 mgml 1 was the highest of the three concentrations, and decreased at higher concentrations. In contrast, KaiC and CI KaiC conservatively eluted as a single hexameric peak at all the concentrations examined (Fig. S4). The unique behavior of CII KaiC in gel filtration chromatography suggested that the hexameric structure of CII KaiC is much more unstable than those of KaiC and CI KaiC. The concentration of CII KaiC was likely responsible for the equilibrium between the hexamer and monomer forms. No significant change in the chromatographic profiles was observed in CII KaiC at all concentrations after incubation at 4 C for 12 h (Fig. 4B). However, when CII KaiC was incubated at 30 C, only the monomer peak was detected for all concentrations. The 40 FEBS Letters 592 (2018) ª 2017 Federation of European Biochemical Societies

6 K. Oyama et al. Phosphorylation in KaiC CII ring phosphorylation reactions at the different protein concentrations during incubation at 30 C were also examined for CII KaiC and KaiC (Fig. 4C, D). Fulllength KaiC showed the same kinetics of the ratio of phosphorylated form irrespective of protein concentration (Fig. 4C). In contrast, the ratio of phosphorylated CII KaiC increased during incubation, and the initial increasing rate became higher with increasing protein concentration (Fig. 4D), suggesting that CII KaiC does not have an autophosphatase activity or, if it does, it would be much lower than that of the autokinase activity; the CI domain would thus be essential for the autophosphatase activity of the CII domain in fulllength KaiC. As the initial ratio of phosphorylated CII KaiC was approximately 0.2, CII KaiC might even be destabilized by such a low phosphorylation level. Once the hexamer form dissociated into monomers, the rate of phosphorylation in monomeric CII KaiC was dependent on protein concentration (Fig. 4D), presumably because the nonphosphorylated CII KaiC monomer behaves not only as an enzyme but also as a diffusible substrate. Irreversible phosphorylation of CII KaiC To examine whether CII KaiC has the autophosphatase activity or not, we examined the reversibility of CII KaiC phosphorylation. The ratio of phosphorylated KaiC is the result of equilibration between the phosphorylation and dephosphorylation reactions, whose rates are Fig. 4. Concentration-dependent phosphorylation of CII KaiC. (A) Gel filtration chromatograms of CII KaiC at different concentrations. CII KaiC was prepared at concentrations of 0.1, 0.2, and 0.4 mgml 1 and subjected to gel filtration chromatography. The chromatograms of CII KaiC at concentrations of 0.1, 0.2, and 0.4 mgml 1 are shown as orange, gray, and black lines, respectively. (B) CII KaiC was prepared at concentrations of 0.1, 0.2, and 0.4 mgml 1 and incubated at 4 and 30 C for 12 h in the presence of 10 mm ATP. CII KaiC before (black line) and after 12 h at 4 C (blue line) and 30 C (red line) incubation were subjected to gel filtration chromatography. KaiC (C) and CII KaiC (D) at concentrations of 0.05, 0.1, 0.2, and 0.4 mgml 1 were incubated at 30 C for 24 h in the presence of 1 mm ATP. Aliquots of reaction mixture were subjected to SDS/PAGE. The ratio of the phosphorylated form to total protein was plotted against time. The symbols,,, and M represent 0.05, 0.1, 0.2, and 0.4 mgml 1, respectively. FEBS Letters 592 (2018) ª 2017 Federation of European Biochemical Societies 41

7 Phosphorylation in KaiC CII ring K. Oyama et al. Fig. 5. Stability of phosphorylated CII KaiC and CII KaiC autophosphorylation activity. (A) Stability of phosphorylated CII KaiC at 4 C. CII KaiC and KaiC were incubated at 30 C for 24 h, and subsequently incubated at 4 C for 24 h in the presence of 3 mm ATP. Aliquots were subjected to SDS/PAGE. The ratios of phosphorylated CII KaiC ( ) and phosphorylated KaiC ( ) are plotted. The data represent the means SD from more than three independent experiments. (B) Phosphorylation of CII KaiC and KaiC under different temperatures. CII KaiC and KaiC were incubated at each temperature between 4 and 30 C in the presence of 3 mm ATP for 24 h. The values of the phosphorylated ratio at 24 h subtracted from the starting ratio (DP) are plotted. Phosphorylation assay was not performed at temperature higher than 30 C, as CII KaiC aggregated at these temperatures. determined by the two antagonistic autokinase and autophosphatase activities of KaiC. As shown in Figs 3A and 4C, the levels of phosphorylated KaiC were decreased by dephosphorylation during incubation at 30 C, while the levels of phosphorylated KaiC increased again following incubation at 4 C (Fig. 5A) [17]. The autophosphatase activity of KaiC was thus much higher than its autokinase activity at 30 C, as the ratio of phosphorylated KaiC decreased at this temperature. The autokinase activity became higher at 4 C and vice versa. In contrast to KaiC, the phosphorylation level of CII KaiC increased at 30 C, reaching approximately 0.7 after 24 h; however, it remained constant during incubation at 4 C (Fig. 5A), indicating that phosphorylation of CII KaiC is irreversible between 4 and 30 C. We further analyzed the temperature dependency between 4 and 30 C of phosphorylation of CII KaiC and KaiC. Figure 5B shows the change in the ratio of phosphorylated protein between pre- and post-24-h incubation at each temperature (ΔP). For KaiC, approximately 0.15 of the ΔP value at 4 C decreased, becoming negative at approximately 10 C and reaching approximately 0.26 at 30 C (Fig. 5B). In other words, the autokinase activity is greater than the autophosphatase activity between 4 and 10 C, and the autophosphatase activity becomes higher than the autokinase, gradually increasing to between 11 and 30 C. In contrast, the ΔP value of CII KaiC constantly increased from 0.03 at 4 C to approximately 0.44 at 30 C, being positive over the range of examined temperatures (Fig. 5B). A negative ΔP value, indicating the dominance of the autophosphatase activity, was not observed for CII KaiC at any temperature examined. As the phosphorylation of CII KaiC was irreversible (Fig. 5A), the temperature dependences indicate that the autophosphatase activity is almost completely impaired in CII KaiC. The CII KaiC protein would show the bare autokinase activity of the CII domain in full-length KaiC. Although KaiA enhances KaiC phosphorylation to increase the KaiC phosphorylation level, KaiB inhibits KaiA activity, thereby decreasing the KaiC phosphorylation level [7,8]. KaiA and KaiB repeatedly assemble and disassemble with KaiC to form complex periodically [18 20]. As CII KaiC was not apparently dephosphorylated in the examined temperature range (Fig. 5B), the effects of KaiA and KaiB on the ratio of phosphorylated CII KaiC were examined (Fig. S5). Incubation with either KaiA or KaiB at 30 C had no effect on the CII KaiC phosphorylation level. In addition, even in the presence of both KaiA and KaiB, the CII KaiC phosphorylation level was the same as the control (CII KaiC only). Discussion KaiC is a unique protein that shows circadian oscillation [3]. Although no conventional kinase or phosphatase motifs are present in KaiC, it exhibits autokinase and autophosphatase activities [7,8]. The two phosphorylation residues, Ser431 and Thr432, are located in the CII domain [9]. In this study, we found that the Thr residue of CII KaiC, which corresponds to Thr432, was phosphorylated, indicating that the CII domain is an active autokinase even without the CI domain. This CI domain-independent autokinase activity of the CII domain is consistent with the previous observation that the phosphorylation activities are also fully retained in a KaiC variant in which the ATPase activity of the CI domain was impaired [11 13]. 42 FEBS Letters 592 (2018) ª 2017 Federation of European Biochemical Societies

8 K. Oyama et al. Phosphorylation in KaiC CII ring CII KaiC was found to be partially phosphorylated in its hexameric state (Fig. 2), indicating that hexameric KaiC possesses autokinase activity [11]. During the KaiC phosphorylation cycle, residues Ser431 and Thr432 in the CII domain are phosphorylated and dephosphorylated in the following programmed sequence: S/T? S/pT? ps/pt? ps/t? S/T, where S represents Ser431, ps represents phosphorylated Ser431, T represents Thr432, and pt represents phosphorylated Thr432 [21,22]. In the nlc/ms/ms analysis of CII KaiC, only the Thr residue but not the Ser residue, was phosphorylated (Fig. S2). Transition to the S/pT state appeared to be caused dissociation of the CII KaiC hexamer into monomers (Fig. 3A, B). The ps/pt or ps/t state of CII KaiC could be cryptic in the nlc/ms/ms analysis and serve as a trigger, while the dissociated S/pT monomers never reassembled in a hexamer. We found that the quasi-static ratio of hexameric and monomeric CII KaiC fractions at 4 C was dependent on protein concentration (Fig. 4A, B). These results suggest that nonphosphorylated CII KaiC is in an equilibrated state of hexamer and monomer forms. Phosphorylation at the Thr residue in CII KaiC would destabilize the interaction between the protomers of hexameric CII KaiC and inhibit further hexamer assembly to promote dissociation of the hexamer into monomers. This is presumably due to the negative charge and/or steric effect of the phosphate group. As CII KaiC AA formed more stable hexamers than CII KaiC and CII KaiC DE, which only existed in the monomeric state (Fig. 3C), electrostatic repulsion between the phosphate group at the Thr residue and ATP might be more crucial than the steric effect of a phosphate group. Once hexameric CII KaiC dissociates into monomers, there would be very little chance for further phosphorylation to occur to form ps/pt, which requires a feasible protomer interaction as discussed below. The CII KaiC hexamer would have a structure similar to the KaiC CII domain. In the atomic structure of KaiC (PDB ID: 3DVL), all amino acid residues responsible for autophosphorylation are located on each protomer interface of the C 6 -symetirical CIIdomain ring [5]. At the protomer interface, one protomer provides the ATP-binding loops and the two catalytic glutamate residues, while the adjacent protomer provides the two phosphorylated Ser and Thr residues. In the single protomer, these residues are located far from each other and the phosphorylation seems not to proceed. Indeed, this structural implication is consistent with a previous report indicating that phosphorylation occurred between monomers of KaiC [15]. The hexameric structure of the CII domain would thus be required to locate the phosphorylated residues close to the c-phosphate of ATP on the protomer surface. It is noteworthy that even after the CII KaiC hexamer totally dissociated to monomers in 12 h, the phosphorylated ratio of CII KaiC still increased (Fig. 4). This unexpected phosphorylation of monomeric CII KaiC showed protein concentration dependency, which is unique for CII KaiC and not observed in KaiC (Fig. 4C, D). As both enzyme and substrate should exist in the diffusible CII KaiC monomer form in this situation, phosphorylation is expected to obey a second- or higher-order reaction model in which CII KaiC forms an oligomer as an enzyme substrate complex. As CII KaiC would be in an equilibrated state of hexamer and monomer as discussed above, it is thus possible that each nonphosphorylated CII KaiC monomer in the solution could be assembled transiently into an oligomer, with the protomer interfaces accessible for phosphorylation of the Thr residue. Supporting this hypothesis, both the hexameric CII KaiC fraction and the CII KaiC phosphorylation level increased with increasing protein concentration at 30 C (Fig. 4). The phosphorylation of CII KaiC proceeded and was irreversible under all the conditions examined, with the phosphorylated form being stable at 4 C (Fig. 5, Fig. S5). CII KaiC thus has no or very little autophosphatase activity; autophosphorylation is always dominant in CII KaiC, suggesting that the CI domain of KaiC is essential for dephosphorylation. Unlike reactions mediated by conventional protein phosphatases, the dephosphorylation of KaiC occurs via the reversal of the phosphorylation reaction, which is proposed to occur between protomers in the hexamer [17,23]. This situation is similar to that for the phosphorylation of monomeric CII KaiC, in which the reaction requires close interaction between protomers to form an enzyme substrate complex, as discussed above. Therefore, the loss of autophosphatase activity in CII KaiC can be regarded as a result of destabilization and disassembly of the hexamer by the phosphorylation of the Thr residue. Taken together with the absence of Ser phosphorylation in CII KaiC, a stable hexameric CII ring of KaiC would be necessary to enable the programmed phosphorylation and dephosphorylation sequence to proceed after the S/pT state. Full-length KaiC forms a stable hexamer irrespective of phosphorylation state. In contrast, CII KaiC dissociated easily into monomers during phosphorylation at Thr (Fig. 3). A previous study proposed that the KaiC hexamer consists of a rigid hexameric CI ring, and a flexible CII ring [15]. The flexibility of the CII ring governs the rhythm of KaiC autophosphorylation and FEBS Letters 592 (2018) ª 2017 Federation of European Biochemical Societies 43

9 Phosphorylation in KaiC CII ring K. Oyama et al. autodephosphorylation [24], and during the phosphorylation cycle, the CII ring exhibits global expansion and contraction motions [25]. The results of this study demonstrate that a stable CI ring is crucial to suppress dissociation of the phosphorylated CII ring. Such a bundle, holding the unstable CII ring by the rigid CI ring, may contribute to the robust circadian oscillation of KaiC phosphorylation. Acknowledgements This study was supported by Support Center for Advanced Medical Sciences, Institute of Biomedical Sciences, Tokushima University for nlc/ms/ms analysis. The authors thank Megumi Fujimoto, Koichi Kato, Yoshitaro Sanbayashi, and Hirokazu Yagi, for their technical supports, and Yuichi Fujita and Anika Wiegard for useful discussions. This study was supported by JSPS Grants-in-Aid No , , , 16H00784, 17K19247 (to KT), and 16J08362 (to KO). Author contributions KT, CA, and KO designed research. KO, JM, and KT performed experiments. KT, CA, and KO wrote the manuscript. All authors analyzed the data and discussed the results. References 1 Kondo T, Strayer CA, Kulkarni RD, Taylor W, Ishiura M, Golden SS and Johnson CH (1993) Circadian rhythms in prokaryotes: luciferase as a reporter of circadian gene expression in cyanobacteria. Proc Natl Acad Sci U S A 90, Tomita J, Nakajima M, Kondo T and Iwasaki H (2005) No transcription-translation feedback in circadian rhythm of KaiC phosphorylation. Science 307, Nakajima M, Imai K, Ito H, Nishiwaki T, Murayama Y, Iwasaki H, Oyama T and Kondo T (2005) Reconstitution of circadian oscillation of cyanobacterial KaiC phosphorylation in vitro. Science 308, Hayashi F, Suzuki H, Iwase R, Uzumaki T, Miyake A, Shen JR, Imada K, Furukawa Y, Yonekura K, Namba K et al. (2003) ATP-induced hexameric ring structure of the cyanobacterial circadian clock protein KaiC. Genes Cells 8, Pattanayek R, Wang J, Mori T, Xu Y, Johnson CH and Egli M (2004) Visualizing a circadian clock protein: crystal structure of KaiC and functional insights. Mol Cell 15, Terauchi K, Kitayama Y, Nishiwaki T, Miwa K, Murayama Y, Oyama T and Kondo T (2007) ATPase activity of KaiC determines the basic timing for circadian clock of cyanobacteria. Proc Natl Acad Sci U SA104, Iwasaki H, Nishiwaki T, Kitayama Y, Nakajima M and Kondo T (2002) KaiA-stimulated KaiC phosphorylation in circadian timing loops in cyanobacteria. Proc Natl Acad Sci U S A 99, Kitayama Y, Iwasaki H, Nishiwaki T and Kondo T (2003) KaiB functions as an attenuator of KaiC phosphorylation in the cyanobacterial circadian clock system. EMBO J 22, Nishiwaki T, Satomi Y, Nakajima M, Lee C, Kiyohara R, Kageyama H, Kitayama Y, Temamoto M, Yamaguchi A, Hijikata A et al. (2004) Role of KaiC phosphorylation in the circadian clock system of Synechococcus elongatus PCC Proc Natl Acad Sci USA101, Murakami R, Miyake A, Iwase R, Hayashi F, Uzumaki T and Ishiura M (2008) ATPase activity and its temperature compensation of the cyanobacterial clock protein KaiC. Genes Cells 13, Kitayama Y, Nishiwaki-Ohkawa T, Sugisawa Y and Kondo T (2013) KaiC intersubunit communication facilitates robustness of circadian rhythms in cyanobacteria. Nat Commun 4, Phong C, Markson JS, Wilhoite CM and Rust MJ (2013) Robust and tunable circadian rhythms from differentially sensitive catalytic domains. Proc Natl Acad Sci U S A 110, Oyama K, Azai C, Nakamura K, Tanaka S and Terauchi K (2016) Conversion between two conformational states of KaiC is induced by ATP hydrolysis as a trigger for cyanobacterial circadian oscillation. Sci Rep 6, Schneider CA, Rasband WS and Eliceiri KW (2012) NIH Image to ImageJ: 25 years of image analysis. Nat Methods 9, Hayashi F, Iwase R, Uzumaki T and Ishiura M (2006) Hexamerization by the N-terminal domain and intersubunit phosphorylation by the C-terminal domain of cyanobacterial circadian clock protein KaiC. Biochem Biophys Res Commun 348, Kinoshita E, Kinoshita-Kikuta E, Takiyama K and Koike T (2006) Phosphate-binding tag, a new tool to visualize phosphorylated proteins. Mol Cell Proteomics 5, Nishiwaki T and Kondo T (2012) Circadian autodephosphorylation of cyanobacterial clock protein KaiC occurs via formation of ATP as intermediate. J Biol Chem 287, Kageyama H, Nishiwaki T, Nakajima M, Iwasaki H, Oyama T and Kondo T (2006) Cyanobacterial circadian pacemaker: Kai protein complex dynamics in the KaiC phosphorylation cycle in vitro. Mol Cell 23, FEBS Letters 592 (2018) ª 2017 Federation of European Biochemical Societies

10 K. Oyama et al. Phosphorylation in KaiC CII ring 19 Tseng R, Goularte NF, Chavan A, Luu J, Cohen SE, Chang YG, Heisler J, Li S, Michael AK, Tripathi S et al. (2017) Structural basis of the day-night transition in a bacterial circadian clock. Science 355, Snijder J, Schuller JM, Wiegard A, L ossl P, Schmelling N, Axmann IM, Plitzko JM, F orster F and Heck AJ (2017) Structures of the cyanobacterial circadian oscillator frozen in a fully assembled state. Science 355, Nishiwaki T, Satomi Y, Kitayama Y, Terauchi K, Kiyohara R, Takao T and Kondo T (2007) A sequential program of dual phosphorylation of KaiC as a basis for circadian rhythm in cyanobacteria. EMBO J 26, Rust MJ, Markson JS, Lane WS, Fisher DS and O Shea EK (2007) Ordered phosphorylation governs oscillation of a three-protein circadian clock. Science 318, Nishiwaki-Ohkawa T, Kitayama Y, Ochiai E and Kondo T (2014) Exchange of ADP with ATP in the CII ATPase domain promotes autophosphorylation of cyanobacterial clock protein KaiC. Proc Natl Acad Sci USA111, Chang YG, Kuo NW, Tseng R and LiWang A (2011) Flexibility of the C-terminal, or CII, ring of KaiC governs the rhythm of the circadian clock of cyanobacteria. Proc Natl Acad Sci U S A 108, Murayama Y, Mukaiyama A, Imai K, Onoue Y, Tsunoda A, Nohara A, Ishida T, Maeda Y, Terauchi K, Kondo T et al. (2011) Tracking and visualizing the circadian ticking of the cyanobacterial clock protein KaiC in solution. EMBO J 30, Supporting information Additional Supporting Information may be found online in the supporting information tab for this article: Fig. S1. Phosphatase treatment of KaiC and CII KaiC. Fig. S2. nlc/ms/ms analysis of CII KaiC. Fig. S3. Gel filtration chromatography of CII KaiC separated using a column packed with silica-based resin. Fig. S4. Gel filtration chromatography of KaiC and CI KaiC. Fig. S5. Neither KaiA nor KaiB showed any effect on CII KaiC phosphorylation. Table S1. List of primers for construction of plasmids. FEBS Letters 592 (2018) ª 2017 Federation of European Biochemical Societies 45