KaiC family ATPases in the nonheterocystous nitrogen-fixing cyanobacterium Leptolyngbya boryana

Abstract

A circadian clock is reconstituted in vitro by incubating three proteins, KaiA, KaiB, and KaiC from the non-nitrogen-fixing cyanobacterium Synechococcus elongatus PCC 7942 in the presence of ATP. Leptolyngbya boryana is a filamentous cyanobacterium that grows diazotrophically under microoxic conditions. Among the aforementioned proteins, KaiC is the main clock oscillator belonging to the RecA ATPase superfamily. Genomic studies have revealed the presence of many genes encoding KaiC family ATPases in archaea and bacteria; however, very few have been analyzed in detail. For example, the L. boryana genome encodes two kaiC homologs designated as LbkaiC1 (LBWT_14830) and LbkaiC2 (LBWT_17950). LbKaiC1 is highly similar to KaiC from S. elongatus PCC 7942 compared with LbKaiC2. LbKaiC1 and LbKaiC2 were purified as Strep-tag fusion proteins. LbKaiC1 formed a hexamer and exhibited autophosphorylation, autodephosphorylation, and ATPase activities. Furthermore, it exhibited circadian phosphorylation rhythm in the presence of KaiA and KaiB from S. elongatus PCC 7942, indicating that LbKaiC1 is the central oscillator of the circadian clock in L. boryana. The temporal separation of nitrogen fixation from photosynthesis may be supported by the circadian rhythm generated by LbKaiC1 in L. boryana. LbKaiC2 had low ATPase activity, which depended on temperature, and its autophosphorylation activity was not detected like a circadian oscillator KaiC. Although the function of LbKaiC2 remains unknown, this work will provide comprehensive understanding of KaiC family ATPases.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-024-81991-x.

Introduction

Circadian rhythms can be observed from prokaryotes to eukaryotes, and they are regulated by an endogenous oscillator called a circadian clock, which coordinates various biological activities with daily environmental changes¹. Cyanobacteria are the simplest organisms exhibiting circadian rhythms. Three clock genes (kaiA, kaiB, and kaiC) are essential for circadian clock activity in the cyanobacterium Synechococcus elongatus PCC 7942². A study has reported that incubating a mixture of purified KaiA, KaiB, and KaiC proteins with ATP at 30°C generates stable oscillations of the phosphorylation states of KaiC³. This oscillator provides a unique model for sub-molecular studies of biological clocks, and it has been used as a self-sustaining biological clock model.

KaiC belongs to the RecA ATPase superfamily⁴; however, it does not participate in DNA recombination. The KaiC ATPase family forms a separate clade in the phylogenetic tree of the RecA superfamily⁴. Based on genomic analyses, proteins homologous to KaiC have been observed in bacteria and archaea⁵, and some variations can be found in the cyanobacterial timing system⁶. In recent studies, KaiC-like proteins are found to play a role in archaeal signaling pathways and in environmental stress in Pseudomonas spp.^7,8.

The N-terminal CI and C-terminal CII domains of KaiC are homologous to each other, and the P-loop ATPase motifs are conserved in both domains⁹. KaiC assembles as a ring-like hexamer with two ATP molecules that are located at each protomer interface¹⁰. The ATPase activity of KaiC is 1 × 10³–1 × 10⁷ times lower than that of other well-known motor proteins^11,12. Apart from low ATPase activity, KaiC exhibits autophosphorylation and autodephosphorylation^9,13,14. The two contiguous residues Ser431 and Thr432 that are located proximal to the protomer interface in the CII domain serve as phosphorylation sites¹⁵. KaiA enhances the autophosphorylation and ATPase activities of KaiC^11,16, whereas KaiB binds to KaiC to inhibit the ATPase and attenuates the effect of KaiA on phosphorylation^13,14. The binding and dissociation of the three Kai proteins occur repeatedly for 24 h^17,18.

Although cyanobacteria are considered as a diverse prokaryotic phylum¹⁹, little is known about the property of circadian clock of other cyanobacteria compared with S. elongatus. The circadian rhythm of bioluminescence was demonstrated in the cyanobacterium Thermosynechococcus elongatus BP-1²⁰, and the circadian rhythm of KaiC phosphorylation was reconstituted in vitro using T. elongatus proteins²¹. The cyanobacterium Synechocystis sp. PCC 6803, which is the most popular model cyanobacterium, carries one kaiA (kaiA1), two kaiB (kaiB1-B2), and three kaiC (kaiC1-C3) homologs in its genome^22,23. kaiA1B1C1, the gene cluster corresponding to the well-characterized kaiABC cluster of S. elongatus, can control the circadian timing in Synechocystis sp. PCC 6803^24,25. The circadian rhythm of the bioluminescence and genome-wide gene expression of the cyanobacterium Anabaena sp. strain PCC 7120 was demonstrated²⁶. KaiC from the cyanobacterium Gloeocapsa sp. PCC 7428 (GlKaiC) was also characterized biochemically, and the result showed that GlKaiC shares basic properties with S. elongatus KaiC²⁷.

Leptolyngbya boryana (formerly known as Plectonema boryanum) is a filamentous, nonheterocystous cyanobacterium that can grow heterotrophically with glucose in the dark²⁸. In addition, L. boryana can grow diazotrophically under microoxic conditions. We previously demonstrated the circadian rhythm of bioluminescence from the reporter gene luxAB in L. boryana²⁹. We also sequenced the genome of L. boryana³⁰ and found two homologous genes (LBWT_14830 and LBWT_17950) exhibiting similar characteristics to kaiC from S. elongatus. Only one kaiC homolog (LBWT_14830) forms a gene cluster with kaiA (LBWT_14850) and kaiB (LBWT_14840). In this study, the recombinant proteins of the two KaiC homologs were purified to elucidate the circadian clock of the diazotrophic cyanobacterium L. boryana, and whether they reconstitute the in vitro circadian clock with KaiA and KaiB from S. elongatus was verified.

Results

Two kaiC homologs in L. boryana

The genomic sequence of L. boryana contains two genes displaying great similarity to kaiC from S. elongatus. The two kaiC-homologous genes, namely, LBWT_14830 and LBWT_17950, were designated as LbkaiC1 and LbkaiC2, respectively. Only LbkaiC1 forms a gene cluster with kaiA (LBWT_14850) and kaiB (LBWT_14840) in the chromosome (Fig. 1A). LbKaiC1 exhibits higher identity (81.5%) to KaiC from S. elongatus than LbKaiC2 (25.7%).

Fig. 1.

(A) Schematic diagram of the gene arrangements around the kaiC homologs in S. elongatus and L. boryana. kaiA, kaiB, and kaiC are shown in orange, blue, and green, respectively. Genes encoding histidine kinases (HK) are also found in L. boryana. (B) Domain arrangement of KaiC. KaiC consists of tandem duplicated domains, namely, the N-terminal CI and C-terminal CII domains. P-loops (P-loop) with a consensus sequence of GXXXXGKT/S (X represents any amino acid), the catalytic carboxylates (two contiguous glutamate residues; catalytic E), and the contiguous phosphorylation residues (phosphorylation sites) are indicated in green, red, and cyan, respectively. The C-terminal extensions are shown in gray. (C) Sequence comparison of SyKaiC, LbKaiC1, and LbKaiC2. Sequence alignment was performed using ClustalW⁵¹. The amino acid residues responsible for the P-loops, catalytic glutamates, and phosphorylation sites in SyKaiC are shown in green, red, and cyan, respectively.

S. elongatus KaiC (SyKaiC) consists of two homologous CI and CII domains, in which P-loop ATPase motifs are conserved (Fig. 1B). The domain arrangements of CI and CII are conserved in LbKaiC1 and LbKaiC2. In particular, LbKaiC1 has conserved P-loop ATPase motifs in the CI and CII domains, which is similar to SyKaiC. In addition, the catalytic E sites E77–E78 and E318–E319 in SyKaiC, which are characterized by consecutive glutamate residues, are conserved in both domains of LbKaiC1 (Fig. 1C). The domain arrangement in LbKaiC2 is also similar to that in SyKaiC, but some important residues are missing. Only one glutamate residue of the catalytic E motif in the CI domain is conserved (E69), whereas the other catalytic E in the CII domain is conserved (E304–E305). The two contiguous phosphorylation residues are also conserved in LbKaiC1 (S431–T432), but LbKaiC2 has only one serine residue (S417). Moreover, SyKaiC has a C-terminal extension domain that is necessary for KaiA binding³¹. This domain is conserved in LbKaiC1 but not in LbKaiC2. Based on the sequence similarity and gene arrangement (Fig. 1C), LbkaiC1 appears to be the kaiC ortholog in L. boryana.

LbKaiC forms a hexamer

LbkaiC1 and LbkaiC2 were overexpressed in E. coli as Strep-tagged fusion proteins and purified via affinity chromatography. Both proteins, as well as SyKaiC, were further purified using gel-filtration chromatography. On the basis of the SDS–PAGE profile (Fig. 2A), the apparent molecular mass of LbKaiC1 (the lower band) was 66 kDa, which was slightly higher than the predicted molecular mass (58,948 Da). The apparent molecular mass of LbKaiC2 (the major band) was 59 kDa, which was consistent with the predicted value (55,474 Da).

Fig. 2.

(A) SDS–PAGE of LbKaiC1 and LbKaiC2 purified via Strep-tag affinity chromatography, followed by gel-filtration chromatography. In resolving proteins, a gel containing 12% T with 2.7% C was used. SyKaiC was loaded as a control. (B) Elution profiles of LbKaiC1, LbKaiC2, and SyKaiC on gel-filtration chromatography. LbKaiC1 (red), LbKaiC2 (green), and SyKaiC (black, control) were subjected to gel-filtration chromatography at 4°C on a Sephacryl S-300 HR column equilibrated with 20 mM Tris–HCl (pH 8.0) buffer containing 150 mM NaCl, 1 mM ATP, 5 mM MgCl₂, and 2 mM DTT. The elution volumes of the molecular mass standards, including apoferritin (443 kDa), β-amylase (200 kDa), and alcohol dehydrogenase (150 kDa), are indicated by the triangles above the elution profiles.

In gel-filtration chromatography, the main elution volume of LbKaiC1 was 54.9 mL, which was identical to that of SyKaiC (Fig. 2B). The apparent molecular mass for a native form of LbKaiC1 and SyKaiC was estimated to be 312 kDa. The hexameric structure of SyKaiC was well characterized¹⁰. Although the apparent molecular mass was slightly less than that of a hexamer (354 kDa), LbKaiC1 exists mainly as a hexamer in solution.

The apparent molecular mass of LbKaiC2 was also estimated via gel-filtration chromatography (Fig. 2B). LbKaiC2 eluted two peaks at 37.2 and 56.3 mL, which correspond to the void volume (> 2000 kDa) and a 270-kDa protein, respectively. The apparent molecular mass of the main peak was consistent with that of a pentamer (277 kDa) rather than a hexamer (333 kDa). However, given that the apparent molecular mass of LbKaiC1 hexamers was underestimated in this gel-filtration chromatography, the main peak of LbKaiC2 was considered to correspond to a hexamer similar to LbKaiC1.

Phosphorylation of LbKaiC

In the SDS–PAGE gel shown in Fig. 2A, purified LbKaiC1 and LbKaiC2 proteins migrated as doublets similar to SyKaiC. In addition, SyKaiC exhibits autophosphorylation, and the upper and lower bands of the doublet on SDS–PAGE in SyKaiC corresponded to phosphorylated and non-phosphorylated forms, respectively³². The doublet bands of LbKaiC1 were similar to those of SyKaiC. LbKaiC2 also migrated as a doublet, but it consisted of a major upper band and a faint lower band.

In obtaining insights into the autophosphorylation of LbKaiC proteins, a radioactive phosphate uptake assay was performed (Fig. 3A). After incubation at 30°C, the phosphorylation level of SyKaiC decreases and stabilizes due to dephosphorylation³³, allowing for the evaluation of the autophosphorylation activity of LbKaiC in this state. KaiC proteins were incubated at 30°C under standard conditions with [γ-³²P] ATP, and aliquots of the reaction mixture taken at 0 and 20 h, followed by SDS–PAGE. Based on the autoradiogram shown in Fig. 3A, phosphate was incorporated into LbKaiC1 and SyKaiC but not into LbKaiC2. Therefore, LbKaiC1 displayed autophosphorylation activity similar to SyKaiC.

Fig. 3.

(A) Autophosphorylation was determined by the transfer of ATP γ-³²P to KaiC proteins. Upper panel: autoradiograph of ³²P-labeled KaiC. Lower panel: Coomassie brilliant blue-stained gel. (B) λ-Phosphatase (λ-PPase) treatment of KaiC. The hexamers of SyKaiC, LbKaiC1, and LbKaiC2 were monomerized by incubating with 0.1 mM ADP buffer at 4°C after replacing the ATP-free buffer³⁵. With or without monomerization treatment, SyKaiC, LbKaiC1, and LbKaiC2 proteins were incubated at 30°C in the absence or presence of λ-PPase for 5 h, followed by the separation of phosphorylated (P-) and non-phosphorylated (NP-) KaiC by SDS–PAGE. (C) Phosphorylation states of KaiC. SyKaiC, LbKaiC1, and LbKaiC2 were incubated at 4°C and 30°C in the presence of ATP for 24 h, followed by SDS–PAGE and Western blot analyses. The KaiC proteins were detected using anti-Strep-tag (upper panel) and anti-SyKaiC (lower panel) antisera. (D) Phosphorylation analysis of LbKaiC1 (filled circles) and SyKaiC (open circles). KaiC proteins were incubated for 24 h at 4°C in the presence of ATP. The ratios of phosphorylated KaiC were estimated using the ratios of the upper-band signals to the total signals in the Coomassie brilliant blue-stained SDS–PAGE profiles. Values are presented as means ± standard deviations from three independent experiments. (E) Dephosphorylation analysis of LbKaiC1 (filled circles) and SyKaiC (open circles). KaiC proteins were incubated for 24 h at 30°C in the presence of ATP. The ratios of phosphorylated KaiC were estimated as described in (D). Values are presented as means ± standard deviations from three independent experiments.

In confirming whether or not LbKaiC1 is phosphorylated, the purified proteins were treated with λ-phosphatase (λ-PPase), which removes phosphate from phosphorylated serine, threonine, and tyrosine residues. Initially, we treated the hexameric LbKaiC1 with λ-PPase and found that the band pattern did not change (Fig. 3B). Then, LbKaiC1 was converted to a monomer and subjected to λ-PPase treatment. After monomerization, LbKaiC1 served as a single non-phosphorylated band, mimicking the properties of SyKaiC. Conversely, λ-PPase treatment did not alter the band pattern of LbKaiC2 regardless of monomerization treatment. These results indicated that the purified LbKaiC1 was phosphorylated but the purified LbKaiC2 was not. Therefore, we concluded that LbKaiC1 exhibits the autophosphorylation activity that gives rise to the shift from lower to upper bands of LbKaiC1 on the SDS–PAGE gel.

Autophosphorylation and autodephosphorylation of LbKaiC1

SyKaiC, LbKaiC1 and LbKaiC2 were incubated at 4–30°C for 24 h, and their phosphorylation states were examined by SDS–PAGE followed by Western blotting (Fig. 3C). Initially, LbKaiC1 and LbKaiC2 proteins were detected by Western blotting using an anti-Strep-tag antiserum. The upper and lower bands of the LbKaiC1 doublet reacted to the Strep-tag antiserum, but only the upper band of LbKaiC2 reacted to the antiserum, which indicated that the faint band was not LBKaiC2. Notably, the antiserum prepared against the SyKaiC protein reacted to LbKaiC1 but not to LbKaiC2 in Western blotting (Fig. 3C).

SyKaiC exhibits autophosphorylation and autodephosphorylation, and the main activity varies depending on the incubation temperature³⁴. Autophosphorylation is dominant at 4°C³⁵. Thus, the phosphorylation level of SyKaiC was increased at 4°C (Fig. 3C, D). By contrast, dephosphorylation is superior at 30°C; thus, the phosphorylation level of SyKaiC was decreased at 30°C (Fig. 3C, E)¹⁴. In addition, the phosphorylation level of LbKaiC1 at 4°C increased, which is similar to that of SyKaiC (Fig. 3D). When incubated at 30°C, the phosphorylation level of LBKaiC1 was decreased slightly slower than that of SyKaiC (Fig. 3E). These results indicated that LbKaiC1 displays autophosphorylation and autodephosphorylation.

LbKaiC1 exhibits a phosphorylation rhythm

In verifying whether LbKaiC displays a circadian phosphorylation cycle in an in vitro reconstitution system, the purified LbKaiC1 was mixed with S. elongatus KaiA and KaiB in the presence of ATP, and the mixture was incubated at 30°C. Based on previous reports, the phosphorylation level of SyKaiC exhibited a clear circadian rhythm with a 24-h period (Fig. 4)³. The phosphorylation levels of LbKaiC1 also oscillated with a slightly longer period (approximately 30 h, Fig. 4). Heterologous combination using KaiA and KaiB from S. elongatus may cause an oscillation of KaiC phosphorylation with longer periods in the reconstituted system. This result indicated that LbKaiC1 is the main circadian clock protein in L. boryana. LbKaiC2 was also mixed with KaiA and KaiB, and the mixture was incubated at 30°C before SDS–PAGE. However, no change in the LbKaiC2 band was observed over 72 h.

Fig. 4.

In vitro reconstitution of circadian oscillation with purified KaiC proteins. (A) SyKaiC and LbKaiC1 were incubated with KaiA and KaiB from S. elongatus PCC 7942 in the presence of ATP for 72 h. Then, the reaction mixtures were subjected to SDS–PAGE every 3 h. (B) The ratios of phosphorylated KaiC to total KaiC were plotted against the incubation time. The data represent the means from three independent experiments. SyKaiC served as the positive control.

LbKaiC1 and LbKaiC2 exhibit low ATPase activity

A previous report indicated that the ATPase activity of SyKaiC is low (approximately 15–16 molecules of ATP per KaiC protomer per day at 30°C) and closely correlated with the circadian period¹¹. The ATPase activities of LbKaiC1 and LbKaiC2 were determined at 30°C (Fig. 5A). The ATPase activity of LbKaiC1 was slightly lower than that of SyKaiC1, but the ATPase activity of LbKaiC2 was even lower than that of LbKaiC1. LbKaiC2 exhibited almost one-third of the activity of SyKaiC, although the P-loop motifs in the CI and CII domains are conserved in LbKaiC2 (Fig. 1B, C).

Fig. 5.

(A) ATPase activities of LbKaiC1 and LbKaiC2 relative to that of SyKaiC at 30°C. ATPase activity was determined by the formation of ADP, which was estimated after separating from ATP by HPLC. (B) The ATPase activities of SyKaiC, LbKaiC1, and LbKaiC2 at 25°C, 30°C, and 35°C are presented as means ± standard deviations from three or more independent experiments. Straight dotted lines indicate linear approximations for SyKaiC, LbKaiC1, and LbKaiC2.

The temperature compensation of KaiC ATPase has been reported^11,36. In examining the temperature dependence of their activity, the ATPase activities of LbKaiC1 and LbKaiC2 were measured at 25°C, 30°C, and 35°C. The ATPase activity of LbKaiC1 was slightly affected by temperatures in the range of 25–35°C, with the activity at 35°C being 1.2-fold higher than that at 25°C. On the contrary, the activity of LbKaiC2 was greatly affected by temperature, being threefold greater at 35°C than at 25°C (Fig. 5B).

Discussion

In this study, LbKaiC1 was identified as the circadian oscillator in the cyanobacterium L. boryana by demonstrating that the phosphorylation of LbKaiC1 displayed circadian rhythms in an in vitro reconstitution system using KaiA and KaiB from S. elongatus instead of their counterparts in L. boryana (Fig. 4). This result indicates that the specific interaction between KaiC and KaiA, as well as between KaiC and KaiB, is conserved in cyanobacteria. Heterologous combination using KaiA and KaiB from S. elongatus may cause an oscillation of KaiC phosphorylation with longer periods in the reconstituted system. Similar to S. elongatus KaiC, LbKaiC1 formed a hexamer and exhibited autophosphorylation, autodephosphorylation, and low ATPase activities (Figs. 2, 3 and 5). GlKaiC from Gloeocapsa sp. PCC 7428 exhibit hexamer formation, phosphorylation, and low ATPase activities²⁷. Therefore, LbKaiC1 shared these properties with GlKaiC and SyKaiC. These features represent the basic properties of the main oscillator of cyanobacterial circadian clocks.

Apart from cyanobacteria, a number of genes encode KaiC-like proteins in the genomes of photosynthetic bacteria and archaea⁷. However, few biochemical studies have investigated KaiC-like proteins other than structural analysis of KaiC-like proteins in photosynthetic bacteria³⁷. In addition to LbKaiC1, the KaiC-like protein LbKaiC2 was also analyzed, and biochemical analysis revealed that LbKaiC2 formed a multimer, either a hexamer or a pentamer (Fig. 2), and showed low ATPase activity in a temperature-dependent manner (Fig. 5). Although previous biochemically analyzed KaiC-like proteins have been reported to be autophosphorylated, no autophosphorylation activity was detected in LbKaiC2 in the present study (Fig. 3). Focusing on whether LbKaiC functions as a circadian oscillator like SyKaiC, autophosphorylation activity was confirmed after 20 h of incubation in our radioactive phosphate uptake assay (Fig. 3A). It will be necessary to assess the activity of LbKaiC2 over a shorter time scale or at 4°C where the autophosphorylation is dominant over autodephosphorylation. Additionally, it would be interesting to examine the phosphorylation of LbKaiC2 in a homologous system using the L. boryana KaiA and KaiB.

LbKaiC2 does not contribute to the circadian clock protein, but it serves as an ATPase with an unknown function. LbKaiC2 may be phosphorylated by other kinases within the cell. Notably, in the chromosome, LbkaiC2 is located upstream of a gene (LBWT_17960) encoding a histidine kinase (HK), which functions as a sensor and signal transducer (Fig. 1). The distance between these two genes is very close, only 3 bases between the start codon of the HK gene and the termination codon of LbkaiC2. It is possible that LbKaiC2 functions in concert with the HK. KaiC-like proteins in archaea may be involved in signal transduction in cells⁷. Thus, our analysis of LbKaiC2 will provide insights into the physiological roles of KaiC-like proteins.

The structures of LbKaiC1 and LbKaiC2 have not yet determined; thus, structural prediction using Alphafold3³⁸ and docking simulation with ATP were performed (Supplementary Fig. S1). The predicted structures were compared with those of KaiC proteins from S. elongatus PCC 7942³⁹, Thermosynechococcus elongatus BP-1 (Thermosynechococcus vestitus BP-1)⁴⁰, and Rhodobacter sphaeroides (Cereibacter sphaeroides)³⁷, which are registered in the Protein Data Bank (PDB). Analysis of the distance between ATP bound in the molecule and the predicted phosphorylation site showed that the γ-phosphate of ATP and the phosphorylation site are farther apart in LbKaiC2 than in KaiC proteins, whose structure has already been elucidated (Fig. S1). This result is consistent with the fact that LbKaiC2 is not autophosphorylated as a circadian oscillator like SyKaiC.

The phylogenetic studies have demonstrated the presence of multiple kaiC homologs in cyanobacterial genomes^24,41. The model cyanobacterium Synechocystis sp. PCC 6803 carries three kaiC homologs in its genome^22,23. Based on the amino acid sequence similarity of the three KaiC proteins in Synechocystis sp. PCC 6803 (kaiC1, slr0758; kaiC2, slr1942; and kaiC3, sll1595), cyanobacterial KaiC homologs have branched into three groups, namely, C1, C2, and C3²⁴. KaiC1 is the authentic circadian oscillator that controls circadian timing with KaiA1 and KaiB1 in Synechocystis sp. PCC 6803 corresponding to the well-characterized KaiABC of S. elongatus. Recently, it was reported that KaiC3 in addition to KaiC1 acts as a circadian clock oscillator and that two circadian clocks are working in the cell of Synechocystis sp. PCC 6803⁴². The arrangement of the gene cluster of LbkaiC1 with kaiA (LBWT_14850) and kaiB (LBWT_14840) and the successful in vitro reconstitution of the circadian rhythms of LbKaiC1 indicate that LbKaiC1 belongs to the C1 group. This group includes the main oscillators displaying circadian rhythms with KaiA and KaiB. The amino acid sequence similarity (identity) between LbKaiC2 and three KaiC proteins; KaiC1, KaiC2, and KaiC3 from Synechocystis sp. PCC 6803 is 25.7%, 21.8%, and 27.5%, respectively. KaiC2 and KaiC3 exhibit autophosphorylation while LbKaiC2 does not (Fig. 3)²⁴. Therefore, LbKaiC2 may represent a unique group different from the C2 or C3 group.

In L. boryana, the bioluminescence rhythm of cells has been demonstrated²⁹, and this study revealed that LbKaiC1 functions as the clock oscillator. In addition, the other clock-related genes are well conserved in the L. boryana genome³⁰. These findings indicate that a circadian clock similar to that in S. elongatus operates and plays important physiological roles in L. boryana. Approximately half of cyanobacterial species can fix nitrogen⁴³. Nitrogen fixation is a process biochemically incompatible with oxygenic photosynthesis because of the oxygen vulnerability of nitrogenase⁴⁴. Some cyanobacteria have developed specialized nitrogen-fixing cells known as heterocysts, which exhibit protection of nitrogenase against oxygen from photosynthesis⁴⁵. The survival strategies of nonheterocystous nitrogen-fixing cyanobacteria such as L. boryana and Cyanothece to accommodate these two incompatible metabolisms in the same cell are very interesting, but they remain largely unknown. Circadian oscillations with half-day shifted phases in nitrogen fixation and photosynthesis have been reported in nonheterocystous nitrogen-fixing cyanobacteria^46,47. Circadian rhythms generated by Kai proteins would enable the temporal separation of these incompatible processes. The CnfR protein, as the master regulator of cyanobacterial nitrogen-fixing genes, activates the transcription of nif genes in response to hypoxia and nitrogen deficiency⁴⁸. Nitrogen fixation in nonheterocystous cyanobacteria may be finely regulated by the cooperative activities of CnfR and Kai proteins. Based on our analysis of these two KaiC homologs in L. boryana, this cyanobacterium contains a standard clock system that is generated by LbKaiC1, and LbKaiC2 is a KaiC-like ATPase with an unknown biological function. Apart from elucidating the function of LbKaiC2, the possible involvement of the Kai oscillatory system in the temporal separation of nitrogen fixation and oxygenic photosynthesis in L. boryana should also be clarified.

Materials and methods

Bacterial strains and plasmid construction for overexpression of Kai proteins

Escherichia coli DH5α cells were used as hosts for plasmid construction, and C41(DE3) cells were used to express Strep-tagged fusion proteins. LbkaiC1 and LbkaiC2 in L. boryana were cloned into the BsaI site of pASK-IBA-5plus vectors (IBA). The following oligonucleotide primers were also used in this study: LBWT_14830 (LbKaiC1), 5′-GTTAATGGTCTCAGCGCCACTTCGATTAATCAAAC-3′ and 5′-CAAGTCGGTCTCGTATCATTATTTGTCTTCTGAGC-3′; LBWT_17950 (LbKaiC2), 5′-GTTAATGGTCTCAGCGCCAGGGGAGATGGCTTACA-3′ and 5′-CAAGTCGGTCTCGTATCATTACTCACCCTCTAGCA-3′.

The nucleotide sequences in the constructed plasmids were confirmed via Sanger sequencing. Sequence data of LbkaiC1 and LbkaiC2 used in this study are available in BioSample with ID SAMD00019988, accession number AP014638³⁰.

Protein expression and purification

KaiA, KaiB, and KaiC from S. elongatus PCC 7942 were expressed and purified as previously described⁴⁹. LbKaiC1 and LBKaiC2 from L. boryana were expressed as follows: E. coli C41 (DE3) cells were transformed with the expression plasmids, after which the cells were cultured in an LB medium containing 100 µg/mL ampicillin with vigorous agitation. At the mid-exponential growth phase (with an OD₆₀₀ value of approximately 0.4), the overexpression of recombinant proteins was induced by adding 0.2 µg/mL anhydrotetracycline. Furthermore, cells expressing LbKaiC1 were cultured at 37°C and harvested at 6 h after induction. Meanwhile, cells expressing LbKaiC2 were cultured at 20°C and harvested at 33 h after induction. LbKaiC1 and LBKaiC2 were purified similarly to S. elongatus KaiC.

Gel-filtration chromatography

The molecular masses of KaiC, LbKaiC1, and LBKaiC2 were evaluated using gel-filtration chromatography. Each protein was loaded onto a Sephacryl S-300 HR column using the ÄKTA prime system (GE Healthcare, Piscataway, NJ, USA). The column was equilibrated and developed using buffer A (20mM Tris (pH 8.0), 150 mM NaCl, 1 mM ATP, 5 mM MgCl₂, and 2 mM DTT). The flow rate was set to 0.4 mL/min. The experiments were performed at 4°C. The column was calibrated with gel-filtration standards, including carbonic anhydrase (29 kDa), serum albumin (66 kDa), alcohol dehydrogenase (150 kDa), β-amylase (200 kDa), and apoferritin (443 kDa). Void volume was determined by using blue dextran.

KaiC autophosphorylation assay using radioisotopes

The autophosphorylation assay was performed as previously described⁹ with some modifications. The purified KaiC proteins (approximately 0.3 mg/mL) were incubated in 20 mM Tris–HCl (pH 8.0), 150 mM NaCl, 5 mM MgCl₂, and 1 mM ATP containing 3.7 MBq/mL [γ-³²P] ATP (NEG002Z, Perkin Elmer, Shelton, CT, USA) at 30°C for 20 h. The reaction was terminated by adding 4× SDS sample buffer. In addition, 1 µg of the proteins was applied to 12.5% precast gel (E-12.5 L, ATTO, Tokyo, Japan) for SDS–PAGE. The gel was dried in an air dryer system (BIO-RAD, CA, USA). ³²P-derived phosphorylated proteins were detected by autoradiography using Typhoon 9400 (Amersham Biosciences, GE Healthcare, Piscataway, NJ, USA).

Phosphatase treatment

KaiC hexamers were dissociated into monomers according to the previously reported³⁵. ATP contained in buffer A of the gel-filtered hexameric KaiC fraction was replaced with ADP using Spin-X UF centrifugal concentrators (30,000 MWCO) (Corning, NY, USA). KaiC in buffer B (20mM Tris (pH 8.0), 150 mM NaCl, 5 mM MgCl₂, 2 mM DTT and 0.1 mM ADP) were then incubated on ice for 24 h. KaiC monomer fraction was obtained by gel filtration chromatography using a Sephacryl S-300 HR column equilibrated with buffer B. Moreover, 2 µg of the proteins purified in NEBuffer for phosphatase treatment with 1 mM MnCl₂ was added to 400 U of lambda protein phosphatase (NEB, MA, USA) and incubated at 30°C for 1 h. The reaction was stopped by adding SDS sample buffer, which was then boiled for 5 min. An aliquot of the sample was subjected to SDS–PAGE, followed by Coomassie Brilliant blue (CBB) staining.

Western blotting of KaiC

Western blotting was performed as previously described with some modifications³⁴. In generating antisera against KaiC, rabbits were immunized using the purified Strep-KaiC. The purified proteins were incubated at 4–30°C in the presence of ATP for 24 h and separated by SDS–PAGE using a gel containing 10% T with 0.67% C. The purified proteins were transferred to PVDF membranes, followed by incubation with anti-Strep-tag (Novagen, WI, USA) at a 1:1,000 dilution or the anti-Strep-KaiC antiserum at a 1:10,000 dilution. Signals were detected using chemiluminescence assay reagents (Chemi-Lumi One Super, Nacalai Tesque, Kyoto, Japan).

In vitro reconstitution of the KaiC phosphorylation cycle

The KaiC phosphorylation cycle was reconstituted in vitro as previously described³. In brief, KaiC was incubated with KaiA and KaiB at 30°C in 20 mM Tris–HCl (pH 8.0) buffer containing 150 mM NaCl, 3 mM ATP, and 5 mM MgCl₂. The final concentrations of KaiA, KaiB, and KaiC were 1.2, 3.5, and 3.5 µM, respectively. At each time point of incubation, an aliquot of the reaction mixture was collected, and the reaction was stopped with an equal volume of 2× SDS sample buffer. The phosphorylated forms of KaiC were separated using SDS–PAGE on a gel containing 10% T with 0.67% C, followed by CBB staining. The relative amount of phosphorylated KaiC in each sample was determined through densitometric analysis using ImageJ⁵⁰.

ATPase activity measurement

The ATPase activities of SyKaiC, LbKaiC1, and LbKaiC2 were measured by determining the amounts of ADP using an HPLC system (Shimadzu, Kyoto, Japan) equipped with a Shim-Pack VP-ODS column (Shimadzu, Kyoto, Japan) as previously described⁴⁹.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Author contributions

YM, KO, CA, YO, and KT performed the experiments; all authors analyzed the data; KT conceived and designed the experiments; KT, KO, YO, and YF wrote the paper; all authors read and approved the manuscript.

Funding

This work was supported in part by the Japan Society for the Promotion of Science (grant-in-aid for 17K19247, 19K05833, 24H02301, and 24K01686) to KT.

Data availability

The data supporting the findings of this study are available from the corresponding author upon reasonable request.

Declarations

Competing interests

The authors declare no competing interests.