Elucidation of the Role of Clp Protease Components in Circadian Rhythm by Genetic Deletion and Overexpression in Cyanobacteria

Abstract

In the cyanobacterium Synechococcus elongatus PCC7942, KaiA, KaiB, and KaiC are essential elements of the circadian clock, and Kai-based oscillation is thought to be the basic circadian timing mechanism. The Kai-based oscillator coupled with transcription/translation feedback and other intercellular factors maintains the stability of the 24-hour period in vivo. In this study, we showed that disruption of the Clp protease family genes clpP1, clpP2, and clpX and the overexpression of clpP3 cause long-period phenotypes. There were no significant changes in the levels of the clock proteins in these mutants. The overexpression of clpX led to a decrease in kaiBC promoter activity, the disruption of the circadian rhythm, and eventually cell death. However, after the transient overexpression of clpX, the kaiBC gene expression rhythm recovered after a few days. The rhythm phase after recovery was almost the same as the phase before clpX overexpression. These results suggest that the core Kai-based oscillation was not affected by clpX overexpression. Moreover, we showed that the overexpression of clpX sequentially upregulated ribosomal protein subunit mRNA levels, followed by upregulation of other genes, including the clock genes. Additionally, we found that the disruption of clpX decreased the expression of the ribosomal protein subunits. Finally, we showed that the circadian period was prolonged following the addition of a translation inhibitor at a low concentration. These results suggest that translational efficiency affects the circadian period and that clpX participates in the control of translation efficiency by regulating the transcription of ribosomal protein genes.

INTRODUCTION

Circadian rhythms, biological oscillations with 24-hour periodicity, are observed in organisms ranging in complexity from bacteria to mammals and allow organisms to adapt to environmental changes (1). Cyanobacteria are the simplest organisms known to have a circadian clock. In the cyanobacterium Synechococcus elongatus PCC7942 (here referred to as Synechococcus), KaiA, KaiB, and KaiC have been identified as essential components involved in circadian oscillation (2). KaiC, an autokinase, autophosphatase, and ATPase, is the central component of the cyanobacterial circadian clock and interacts with KaiA and KaiB (2, 3, 4, 5, 6). KaiA stimulates KaiC autophosphorylation, and KaiB opposes the stimulatory activity of KaiA (7, 8, 9). The circadian rhythm of KaiC phosphorylation persists even in the absence of transcription and translation (10). Moreover, the circadian oscillation of KaiC phosphorylation can be reconstituted in vitro by incubating recombinant KaiA, KaiB, and KaiC with ATP (11). Thus, the posttranslational activities of the KaiABC oscillator are sufficient to generate the Synechococcus circadian rhythm.

Transcription/translation feedback pathways appear to be required for the maintenance of a stable circadian clock in vivo. Most genes in Synechococcus, including kaiBC, are clock controlled (12). Therefore, KaiC is thought to be a promoter-nonspecific, genome-wide transcriptional modifier, possibly acting via its effect on the basic transcriptional machinery (13). Additionally, the circadian gene expression rhythm persists in the absence of the KaiC phosphorylation rhythm, and rhythmic kaiBC transcription and translation are thought to play a part in circadian rhythm generation in cyanobacteria (14). Temporal information from the KaiABC oscillator is transmitted to downstream genes via the histidine kinase SasA. SasA and its cognate response regulator, RpaA, which has a putative DNA binding domain, are positive transcriptional regulators of clock-controlled genes, including kaiBC (15, 16). It was recently shown that RpaB, a paralog of RpaA, inhibits clock-dependent promoters, suggesting that specific sigma factors target the SasA-RpaAB output system, thereby propagating genome-wide transcriptional oscillation (17). It has been proposed that clock-controlled dynamic contraction of the nucleoid structure is the mechanism of feedback regulation (18). Therefore, multiple factors, including the KaiABC oscillator and transcription/translation feedback, are important for maintaining the periodicity of the intracellular clock.

Transient increases in KaiC shift the phase of the Synechococcus clock (2, 19), and the quantitative proportions of KaiA, KaiB, and KaiC modulate the period and amplitude of KaiC phosphorylation in vitro (20). Thus, the balance of the Kai protein levels is also important for maintaining normal oscillation of the circadian clock. The stability of KaiC fluctuates in a circadian manner, with maximum stability occurring at subjective midnight (21). A previous study using complete genome sequencing following transposon mutagenesis revealed that the disruption of two protease-related genes, clpP2 and clpX, results in a long-period phenotype (22). The mechanism of period lengthening due to the mutation of these proteases is not clear.

The Clp protease is one of the most prominent ATP-dependent Ser-type proteases, and it is conserved among eubacteria and most eukaryotes (23). Synechococcus has three distinct ClpP paralogs (ClpP1, ClpP2, and ClpP3) as well as a ClpP-like protein, ClpR, which lacks the catalytic triad typical of Ser-type proteases (24). Each protease consists of a unique proteolytic core composed of separate Clp subunits, one with ClpP1 and ClpP2 and the other with ClpP3 and ClpR. Each core also associates with a particular HSP100 chaperone partner: the ClpP1/ClpP2 protease interacts with ClpX, and the ClpP3/ClpR protease interacts with ClpC (25). The levels of Clp family proteins increase in response to stress, for example, high light or low temperature (24, 26, 27). However, little is known about the targets of the different Clp complexes in cyanobacteria. It is possible that Clp proteases degrade Kai proteins, and the resulting quantitative Kai protein balance contributes to determining the period length.

To understand the roles of Clp protease family members in the regulation of the circadian clock, we comprehensively examined the effects of clp mutations on the circadian rhythm. In the clpP1, clpP2, and clpX deletion mutants and the clpP3-overexpressing (OX-clpP3) strain, we observed the long-period phenotype. Moreover, we found that the overexpression of clpX decreased activity at the kaiA, kaiBC, and psbAII promoters following an initial increase in activity. Microarray and quantitative reverse transcription-PCR (QRT-PCR) analyses showed that ribosomal protein (r-protein) subunit genes were upregulated before other genes were upregulated in the OX-clpX strain. Conversely, the expression of r-protein subunit genes was decreased by the deletion of clpX. Finally, we showed that the circadian period was prolonged by the addition of a translation inhibitor. These results suggest that translational efficiency affects the circadian period and that clpX participates in the control of translation efficiency by regulating the transcription of ribosomal protein genes.

MATERIALS AND METHODS

Bacterial strains, culture, and media.

The Synechococcus elongatus PCC7942-based strains used in this study were wild-type (WT) strains containing a PkaiBC::lux luciferase reporter selected with chloramphenicol (NUC42) (28) or kanamycin (NUC301) (21) in neutral site I (NSI), a PpsbAII::luxAB luciferase reporter (AMC520) (29), or a PkaiA::luxAB luciferase reporter (NUC35) (15). Cells were cultured at 30°C under constant white light (40.5 μE · m⁻² · s⁻¹) in BG-11 medium (30) unless otherwise noted.

Construction of protease gene disruption mutants.

The deletion of protease genes was performed by substitution with (clpP1 and clpP2) or the insertion of (clpX) the spectinomycin resistance gene. For the substitution mutants, two PCR steps were used. In the first step, approximately 500 bp of the up- and downstream regions of the target protease gene were amplified using primer sets 1 and 2 or 3 and 4 (see Table S1 in the supplemental material), respectively. The Ω cassette sequence was added to the fragments at this step by including the Ω cassette sequence in the primers (upstream region, 5′-CTGCGGGTCAAGGATCTGGATTTCG-3′; downstream region, 5′-CAATTCGTTCAAGCCGACGCCGCTTC-3′). The fragment of the Ω cassette was amplified by PCR using two primers: 5′-CTGCGGGTCAAGGATCTGGATTTCG-3′ and 5′-CAATTCGTTCAAGCCGACGCCGCTTC-3′. The second PCR step used the three PCR products from the first step as templates and the primers that were used in the first PCR (primers 1 and 3 in Table S1 in the supplemental material). PCR fragments in which the Ω cassette was inserted between the up- and downstream sequences of the protease gene were obtained. For insertional deletion of clpX, fragments containing clpX were PCR amplified using two primers, 5′-GTGATGTCGAGATACGACTC-3′ and 5′-GATATCGAGATCGGCCACTTG-3′. The PCR products were subcloned into pGEM-T (Promega). Then, following cleavage from pBS322ΩE by SmaI, the Ω cassette fragment was inserted into the Eco47III restriction site of pGEM-T/clpX. The constructs were transformed into NUC301 cells, which were then selected with 40 μg ml⁻¹ spectinomycin. Correct insertion and complete segregation of the disrupted construct in selected transformants were confirmed by Southern blotting and PCR (data not shown).

Construction of clp-overexpressing mutants.

p322P_trc (31) was digested with HincII to remove a short fragment and religated. This product was then digested with BglII, and the shorter fragment carrying lacI^q-P_trc was subcloned into the BamHI site of the pTS2KC targeting vector (31) to obtain pTS2KC::P_trcΔHincII. To construct the IPTG (isopropyl-β-d-thiogalactopyranoside)-inducible, C-terminal-FLAG-tagged clp protease vector, we performed PCR using primers that included an appended restriction site and a FLAG tag sequence (see Table S2 in the supplemental material). The PCR products were inserted into pTS2KC::P_trcΔHincII after digestion with the appropriate restriction enzymes (see Table S2). The construct was transformed into NUC42, AMC520, and NUC35 cells and selected with 25 μg ml⁻¹ kanamycin. We confirmed the overexpression of the clp gene following the addition of IPTG by Western blot analysis to detect the C-terminal FLAG tag (data not shown).

Bioluminescence assay.

Bioluminescence assays were performed as described previously (3). Briefly, cells were grown on agar plates under constant light (LL), and the bioluminescence was monitored under LL (40.5 μE · m⁻² · s⁻¹) at 30°C after the cells had been exposed to darkness for 12 h. The conditions described here are referred to as the standard conditions in this study.

To evaluate the stability of the period in response to light intensity, we calculated the rate of change in the period when the light intensity was doubled. Based on the frequencies at 1.7, 40.5, and 74.3 μE · m⁻² · s⁻¹, the slope and intercept were calculated, and the rate of change in the period from 22.8 to 45.6 μE · m⁻² · s⁻¹ was calculated from each slope and intercept.

Western blot analysis.

Synechococcus whole-cell extracts were prepared as described previously (9). The protein concentration was determined by the bicinchoninic acid (BCA) method using bovine serum albumin (BSA) as a standard. Whole-cell extracts (5 μg) were subjected to SDS-PAGE (15% total acrylamide [T] with 2.7% cross-linking factor [C] for KaiB, KaiA, ClpP1, ClpP2, ClpP3, and ClpR, 10% T with 2.7% C for KaiC and ClpX, and 11% T with 0.67% C for KaiC) and transferred to nitrocellulose membranes. Immunoblots were incubated in the presence of anti-KaiA, anti-KaiB, anti-KaiC, and anti-Flag (Wako; diluted 1:10,000) antibodies, as described previously (3), and proteins were detected with enhanced chemiluminescence. Serial dilutions of the standard proteins (recombinant Kai proteins) and Synechococcus cell extracts were loaded onto the same gel and subjected to immunoblotting assays. The intensity of each relevant signal was measured with a densitometer. In each experiment, we confirmed that the intensities of the relevant immunoreactive bands for the whole-cell extracts and the standard proteins yielded a linear dose response (data not shown). Equal loading of extracts was confirmed by Coomassie blue staining in the gel and/or by measuring the density of nonspecific bands on the immunoblots (data not shown). The immunoblot signals for the Kai proteins were analyzed with NIH ImageJ software.

Determination of the growth rate.

Initial cultures were grown in liquid BG-11 medium at 30°C under constant illumination (30 μE · m⁻² · s⁻¹) with air bubbling. Cell growth was monitored by measuring the optical density at 750 nm (OD₇₅₀). When the cell densities reached an OD₇₅₀ of 0.8, the cultures were diluted to an OD₇₅₀ of 0.005 and grown under constant illumination (30 μE · m⁻² · s⁻¹) at 30°C with air bubbling. The cell densities were determined as OD₇₅₀s at the times indicated in Fig. 1C. The doubling time was calculated from the growth rate of the exponential growth phase.

Fig 1.

Long-period phenotypes of clp deletion mutants. (A) The effect of protease deletion on the rhythm of kaiBC promoter activity. kaiBC promoter activity was monitored with a PkaiBC luciferase reporter gene set in wild-type (black) and protease deletion strains (ΔclpP1, ΔclpP2, and ΔclpX; red). The peak bioluminescence intensity for the kaiBC promoter activity of each strain was normalized to 1. (B) Accumulation of KaiA, KaiB, and KaiC and the phosphorylation profile of KaiC in clp protease family deletion mutants. Cells were collected at LL16 and LL28, and cell extracts were analyzed by immunoblotting. The upper and lower bands in the KaiC panel correspond to phosphorylated KaiC (P-KaiC) and nonphosphorylated KaiC (NP-KaiC), respectively. The lower panel depicts the relative KaiA, KaiB, and KaiC protein abundances. The signals were normalized to the value for the WT at the same time point, which was set to 1. The results are shown as the means ± standard deviations (SD) (n = 6). Wild type, white bar; ΔclpP1 mutant, gray bar; ΔclpP2 mutant, black bar; ΔclpX mutant, blue bar. (C) Comparison of the growth rates of the ΔclpP1, ΔclpP2, and ΔclpX deletion mutants. Black and red circles indicate the wild type and deletion mutants, respectively. The horizontal axes indicate the incubation time. Three or four independent experiments were performed for each strain, and the growth curves were plotted as the average OD₇₅₀ values ± SD. (D) ΔclpX cells exhibited filamentation. Shown are microscopy images of wild-type and ΔclpX Synechococcus cells. Bars = 15 μm.

For comparisons of the growth at various temperatures and light intensities, cells were cultured in BG-11 medium under LL conditions with shaking for 7 days. The cultures were diluted with BG-11 medium, and 2 μl of each culture was spotted onto a BG-11 agar plate. Wild-type and mutant colonies were grown on solid media for the indicated number of days (see Fig. S1 in the supplemental material) under five conditions (standard conditions, 30°C and 40.5 μE · m⁻² · s⁻¹; 35°C, 35°C and 40.5 μE · m⁻² · s⁻¹; 25°C, 25°C and 40.5 μE · m⁻² · s⁻¹; high light, 30°C and 74.3 μE · m⁻² · s⁻¹; low light, 30°C and 6.7 μE · m⁻² · s⁻¹). After growth, the colonies were photographed.

Preparation of cDNA.

Cells were collected and stored at −80°C. Total RNA was purified with the modified acid-phenol method (15), and cDNA was prepared with SuperScript III reverse transcriptase (Invitrogen) using random hexamers as primers (Invitrogen).

Microarray analysis.

Microarray analysis was performed using an Affymetrix GeneChip design based on the Synechococcus genome (32). The signal from each array was corrected using DNA and global normalization such that the average expression levels for all open reading frames (ORFs) within a microarray were equal across replicate arrays (32). For better estimation of the relative expression levels of different genes, we standardized each cDNA-derived signal based on the corresponding genomic DNA signals as described previously (32). Additionally, global normalization was applied to the RNA signal profiles under LL conditions such that the averages of the expression levels for all ORFs within a microarray were equal across replicate arrays as described previously (32).

QRT-PCR.

To synthesize cDNA, RNA (1 μg) was converted into cDNA with SuperScript III reverse transcriptase (Invitrogen) using random hexamers as primers (Invitrogen). The synthesized cDNAs were amplified with FastStart Essential DNA Green Master hot start reaction mix (Roche Applied Science) and the primer set for each target gene (see Table S3 in the supplemental material) and then analyzed using a LightCycler Nano system (Roche Applied Science). A linear standard curve and a plot of the threshold cycle number versus the log of the designated transcript level were prepared using purified total DNA from NUC42 cells. The following standard thermal cycling program was used for all PCRs: 95°C for 600 s and 45 cycles of 95°C for 10 s, 60°C for 10 s, and 72°C for 15 s. The efficiency of each QRT-PCR and the melting curves (from 60 to 95°C at 0.1°C s⁻¹) of the products were also analyzed to ensure the existence of a single amplification peak corresponding to a unique target gene. The data were analyzed using LightCycler Nano software 1.0 (Roche Applied Science).

Microarray data accession number.

The microarray data reported in this paper have been deposited in the National Center for Biotechnology Information Gene Expression Omnibus (GEO) under accession no. GSE42542.

RESULTS

The effect of clp protease family disruption on circadian gene expression.

To determine the contributions of members of the clp protease family to the circadian phenotype, we attempted to introduce gene-specific disruptions of the clpP1, clpP2, clpP3, clpR, clpX, and clpC genes by substitution with or insertion of the spectinomycin resistance gene (ΔclpP1, ΔclpP2, ΔclpP3, ΔclpR, ΔclpX, and ΔclpC). Viable transformants were obtained for ΔclpP1, ΔclpP2, and ΔclpX, and we confirmed the complete segregation of the protease inactivation mutants and did not detect wild-type (WT) alleles (data not shown). Viable transformants were not obtained for the clpC, clpR, and clpP3 genes, which is consistent with a previous report (24). We then measured the kaiBC gene expression patterns over time in the disruption mutants using a PkaiBC bioluminescence reporter (PkaiBC::luxAB). As shown in Fig. 1A, the circadian period was extended in ΔclpP2 and ΔclpX strains, as previously reported (22). The disruption of ClpP1, which forms a complex with ClpP2 and ClpX (25), also led to an increase in period length (Fig. 1A and Table 1).

Table 1.

Characterization of circadian rhythm of clp mutants

Strain	Period (h)		Q10a	Rate of period changeb
	Without IPTG	With IPTG
NUC42 (WT)	25.41	—c	0.98	1.02
OX-clpP1 strain	25.18	24.93	0.97	1.01
OX-clpP2 strain	25.02	25.06	0.98	1.01
OX-clpP3 strain	24.97	27.35	1.01	1.01
OX-clpR strain	25.19	25.49	0.98	1.01
OX-clpR-clpP3 strain	25.02	25.45	0.98	—
OX-clpX strain	25.10	—	—	—
NUC301 (WT)	24.65	—	0.99	1.02
ΔclpX strain	25.74	—	1.00	1.01
ΔclpP1 strain	25.06	—	1.00	1.01
ΔclpP2 strain	25.01	—	0.98	1.01

^aThe Q₁₀ temperature coefficient is the rate of change of a biological system as a consequence of increasing the temperature by 10°C. It was calculated based on the period length measured for kaiBC promoter activity at 25, 30, and 35°C.

^bThe rate of the period change when light intensity doubles was calculated from period length measurements at 1.7, 40.5, and 74.3 μE · m⁻² · s⁻¹.

^c—, not measured.

Because the quantitative proportions of KaiA, KaiB, and KaiC modulate the period (20), it is possible that the long-period phenotype of these protease deletion mutants resulted from an imbalance in the clock protein levels. Thus, we examined the amounts of the KaiA, KaiB, and KaiC proteins in WT, ΔclpP1, ΔclpP2, and ΔclpX cells collected at the 16th hour of constant light (LL16), when the levels of KaiB and KaiC and the phosphorylation of KaiC are maximal in WT cells, and at LL28, when the levels and phosphorylation are minimal (7, 19). In ΔclpX cells, the KaiB level was decreased by approximately 0.5-fold, and the level of KaiA tended to increase in ΔclpP1 and ΔclpX cells; however, these differences were not significant. The total KaiC levels and the phosphorylation profiles in all mutants were similar to those in WT cells. Thus, we concluded that changes in the levels and balance of the clock proteins did not occur in the deletion mutants (Fig. 1B).

To confirm the growth phenotype of the protease mutants, we determined the growth rate in liquid medium under LL at 30°C. The growth of ΔclpX mutants was a little slow, but the growth rates of the ΔclpP1 and ΔclpP2 mutants were approximately normal under our experimental conditions (Fig. 1C). The doubling times of the wild-type, ΔclpP1, ΔclpP2, and ΔclpX cells were 7.8 ± 0.48 h, 7.8 ± 0.16 h, 8.2 ± 0.38 h, and 8.6 ± 0.20 h, respectively. The ΔclpP1 mutant has previously been reported to grow slowly, and the ΔclpP2 mutant has been reported to have normal growth (24, 26). This discrepancy may be due to differences in the growth conditions because the growth of the ΔclpP1 mutant is affected by the environmental conditions (see Fig. S1 in the supplemental material) (26). Additionally, we showed that ΔclpX caused cell filamentation in liquid culture (Fig. 1D).

Effect of the overexpression of clp protease subunits on circadian gene expression.

To analyze the effects of the clp protease multigene family on the circadian system in detail, we next examined the effect of subunit overexpression on circadian gene expression in reporter strains carrying IPTG-inducible Ptrc::clpP1, Ptrc::clpP2, Ptrc::clpX, Ptrc::clpP3, Ptrc::clpR, or Ptrc::clpC transgenes (OX-clpP1, OX-clpP2, OX-clpX, OX-clpP3, OX-clpR, and OX-clpC strains, respectively). In addition, because the genomic distance between clpR and clpP3 is very short (42 bases), with the predominant expression of these genes being monocistronic (24), we examined the effect of the cooverexpression of clpR-clpP3 in cells carrying Ptrc::clpR-clpP3 (OX-clpR-clpP3 strain). We confirmed the cooverproduction of ClpP3 by Western blotting to detect its C-terminal Flag tag (data not shown). No viable transformants were obtained for the OX-clpC strain. OX-clpX and OX-clpP3 strains did not grow when exposed to 10 and 500 μM IPTG, respectively (Fig. 2A). The effects of the overexpression of clp genes on the bioluminescence rhythms following induction by the addition of 100 μM IPTG were assayed in PkaiBC reporter strains (Fig. 2B and Table 1). The period of OX-clpP3 cells was lengthened by approximately 2.4 h. In OX-clpP1, OX-clpP2, OX-clpR, and OX-clpR-P3 cells, there were no apparent effects on the kaiBC promoter activity rhythm.

Fig 2.

Phenotypes of cells overexpressing clp genes. (A) Growth characteristics of OX-clp cells. Wild-type and clp-overexpressing colonies were grown on solid media containing IPTG at the indicated concentrations for 12 days under standard conditions. (B) kaiBC promoter activity in clp-overexpressing cells. The bioluminescence levels of wild-type and clp-overexpressing strains were monitored. Cells were grown on agar plates for 4 days under LL with (red) or without (black) 100 μM IPTG, and the bioluminescence rhythm was monitored. The bioluminescence intensity per colony was normalized such that the peak value for the kaiBC promoter activity of each strain with IPTG was 1.

Phenotype of Clp protease mutants under various environmental conditions.

Next, we examined the growth rates of clp protease family deletion mutants and overexpressing cells under various environmental conditions on agar plates (see Fig. S1 in the supplemental material). The growth of ΔclpX mutants was slow under all experimental conditions, and the ΔclpP1 mutant grew slowly under the standard, 35°C, and high-light conditions. The growth of the ΔclpP2 mutant did not differ from that of the wild-type strain. The slow growth of the ΔclpP1 mutant under the high-light and 35°C conditions was consistent with a previous report (26). OX-clpP3, OX-clpR, and OX-clpR/clpP3 strains grew slowly under the standard and 35°C conditions. OX-clpR and OX-clpR-clpP3 strains also showed slow growth under the high-light condition, and the OX-clpR strain showed slow growth at 25°C. It is notable that the OX-clpP3 strain was not viable under the low-light-intensity conditions. The growth patterns of OX-clpP1 and OX-clpP2 strains did not differ from that of the wild-type strain. The slow growth of the OX-clpR-clpP3 strain without the addition of IPTG might due to leaky transcription of clpR-clpP3 from the trc promoter. Thus, the growth rates of the clp protease family mutants changed with the temperature and light conditions, indicating that each Clp protein has a specific function in the responses to different temperature and light conditions.

Given the data implicating Clp protease family members in the regulation of the circadian period, we hypothesized that Clp proteases may participate in temperature compensation and contribute to the robustness of the period under varying light intensities. To investigate this possibility, we examined the stability of the period of clp mutants under varying environmental conditions. Following growth under standard conditions for 3 days, we measured the bioluminescence rhythms of clp overexpression and deletion mutants under various temperature and light intensity conditions. As shown in Table 1, the Q₁₀ values (where the Q₁₀ temperature coefficient is the rate of change of a biological system as a consequence of increasing the temperature by 10°C) for these mutants were between 0.97 and 1.01, and the rate of change in the period in response to changes in light intensity was 1.01 (see Materials and Methods). Thus, in the clp mutants, although the growth rate changed with changing environmental conditions, the period of the circadian rhythm was stable. These results suggest that the modulation of the circadian period by Clp proteases and the stabilization of the period during exposure to environmental changes are controlled by different mechanisms.

Additionally, these results indicate that slow growth does not necessarily cause period alterations and that abnormal periods do not necessarily cause slow growth. Thus, we did not observe a clear correlation between rhythm abnormalities and the growth rate in clp mutants.

Effect of clpX overexpression on kaiBC promoter activity.

The OX-clpX strain did not grow under any experimental conditions (Fig. 2A; see Fig. S1 in the supplemental material). To examine the effect on kaiBC expression immediately after the induction of overexpression, we added 100 μM IPTG every 4 h from LL16 to LL36 and monitored the bioluminescence from the kaiBC promoter reporter. As shown in Fig. 3A, the kaiBC promoter activity began to increase approximately 6 to 8 h after the addition of IPTG regardless of the phase of the circadian rhythm. The bioluminescence eventually decreased, the circadian rhythm was disrupted, and the cells did not survive. To examine the effect of the transient overexpression of clpX, the OX-clpX strain was activated with 100 μM IPTG for 4 h at LL16 and LL28. OX-clpX cells that had been returned to IPTG-free plates after 4 h of transient activation were able to survive (data not shown). As shown in Fig. 3B and C, the kaiBC promoter activity decreased after IPTG addition, concomitant with the disappearance of the circadian rhythm, although the rhythm recovered within several days. Surprisingly, the phase of the recovered rhythm was almost the same as the phase of uninduced cells. These results suggest that the Kai-based core oscillator was not affected by the overexpression of clpX and imply that the balance of Kai proteins may not change.

Fig 3.

The effect of clpX overexpression on circadian oscillation. (A) Constant clpX overexpression produced a transient increase and then a decrease in kaiBC promoter activity independent of the circadian phase. The NUC42 OX-clpX strain was treated with 100 μM IPTG (red circles) at the indicated time. Arrows indicate the addition of IPTG. (B and C) Transient ClpX induction does not affect the phase of kaiBC promoter activity. Wild-type (NUC42; black circles) and OX-clpX (red circles) cells were grown on IPTG-free agar medium and treated with 100 μM IPTG at LL16 (B) or LL28 (C) for 4 h. The orange bars indicate the addition of IPTG. The lower panels show enlarged bioluminescence intensity scales for the wild-type (×10) and OX-clpX (×40) cells. (D) Effect of constant clpX overexpression on the bioluminescence rhythm of kaiA and psbAII reporter activity. NUC35 OX-clpX (PkaiA::luxAB) and AMC520 OX-clpX (PpsbAII::luxAB) cells were grown and treated (red circles) or not treated (black circles) with 100 μM IPTG at LL16 or LL28. Arrows indicate the addition of IPTG. The bioluminescence intensity per colony was normalized such that the peak value for the kaiBC or kaiA promoter activity of each strain without IPTG was 1 and the peak value for psbAII with IPTG was 1.

We also evaluated the effect of constant clpX overexpression on the kaiA and psbAII promoter activities at LL16 and LL28. The bioluminescence from PkaiA and PpsbAII cells increased several hours after induction and then decreased as the cells exhibited a disrupted circadian rhythm similar to that observed for PkaiBC cells (Fig. 3D). The overexpression of ClpX affected the promoter activity of not only clock genes but also general photosynthesis genes.

ClpX is involved in the regulation of ribosomal protein subunit expression.

We performed microarray analysis to examine the effect of clpX overexpression on gene expression genome wide. We compared the gene expression levels between OX-clpX and wild-type cells after the addition of 100 μM IPTG for 4 h at LL16. The expression levels did not differ significantly between the wild-type and OX-clpX cells in the absence of IPTG (Fig. 4A). In contrast, several genes were up- and downregulated following 4 h of clpX overexpression (Fig. 4B) (GEO accession no. GSE42542). Notably, the expression levels of several r-protein subunit operons were upregulated significantly. The level of kaiC mRNA did not change after 4 h of clpX overexpression. We also confirmed by reverse transcription-PCR (RT-PCR) that the mRNA levels of the r-protein subunits increased and the level of kaiBC mRNA did not change within 4 h after the addition of IPTG (see Fig. S2 in the supplemental material). This result is consistent with the microarray results.

Fig 4.

Upregulation of r-protein subunit genes following the overexpression of clpX. (A and B) Comparison of the gene expression levels between wild-type and clpX-overexpressing cells by microarray analysis. Wild-type and NUC42 OX-clpX cells were grown under constant light conditions. After two treatments of darkness, the cells were returned to LL, and IPTG was added at LL16. The cells were collected before the addition of IPTG (A) and 4 h after the addition of 100 μM IPTG (B), and total RNA was isolated. The gene expression profiles of the wild-type and OX-clpX cells are plotted on the horizontal and vertical axes, respectively. Black circles, all detectable genes; red open circles, clpX; red solid circles, r-protein subunit genes; blue circles, kaiC; green circles, rpoA. (C) Comparison of the mRNA levels in wild-type and clpX-overexpressing cells determined by QRT-PCR. Wild-type and NUC42 OX-clpX cells were grown under standard conditions in liquid medium. The cells were exposed to darkness for 12 h and then returned to LL. IPTG (100 μM) was added at LL16. Cells were collected at 0 h (LL16), 2 h (LL18), and 8 h (LL24) after treatment, and total RNA was isolated. The mRNA levels at each time point were normalized to the average value at 0 h, which was taken as 1. Two independent samples were analyzed for each experiment. Each sample was run in three or more PCRs. The vertical bar on each column indicates the standard error. Wild-type cells were also treated with 100 μM IPTG to serve as a negative control for IPTG addition because they have no trc promoter and because the kaiBC promoter activity in WT cells does not change with the addition of IPTG (2). Statistical significance between wild-type and OX-clpX strains, identified by Student's t test, is indicated with asterisks (∗∗, P < 0.005; ∗∗∗, P < 0.001).

We next examined the temporal pattern of expression of kaiBC, rpoA, and several r-protein subunit genes (rpl2, rpl3, rps1, and rps6) after the overexpression of clpX by QRT-PCR (Fig. 4C). rpoA encodes the alpha subunit of RNA polymerase, and arrhythmic high gene expression was observed for the LL and constant darkness (DD) conditions in a previous microarray analysis (32). The expression of clpX increased within 2 h after IPTG addition, and the expression of kaiBC, rpoA, and r-protein subunit genes increased 8 h after IPTG addition (Fig. 4C). kaiBC expression deceased in wild-type cells 2 h after IPTG addition. The accumulation of kaiBC mRNA displayed circadian cycling under LL in wild-type cells, and the peaks of the mRNA levels occurred between LL9 and LL12 (2, 15, 21). These expression levels then decreased from LL16 to LL24 (2, 15, 21). Thus, the decrease in kaiBC mRNA observed 2 h after IPTG addition, which corresponds to LL18, is thought to be caused by the rhythmic accumulation pattern of kaiBC mRNA. According to the microarray and RT-PCR results, r-protein subunit genes were upregulated within 4 h after IPTG addition. In contrast, the expression levels of kaiBC and rpoA did not change after 4 h (Fig. 4B; see Fig. S2 in the supplemental material), although the expression levels of these genes increased conspicuously 8 h after induction (Fig. 4C). The results of the bioluminescence measurements showed that kaiBC promoter activity also began to increase approximately 6 to 8 h after the addition of IPTG. A previous study demonstrated that bioluminescence from the PkaiBC reporter cells reflects the transcriptional activity of the kaiBC promoter (33). These results suggest that the overexpression of clpX upregulated the transcription of r-protein genes, which then caused changes in the expression of other genes, including kaiBC and rpoA. Additionally, we analyzed the effect on the expression of r-protein subunit genes (rpl2, rpl3, rps1, and rps13) in ΔclpX cells using QRT-PCR. The r-protein subunit genes that we examined were downregulated, especially at LL16 (Fig. 5). These results suggest that ClpX is involved in the regulation of the expression of r-protein subunits.

Fig 5.

Expression of the r-protein subunit genes was decreased in ΔclpX cells. QRT-PCR analysis of r-protein subunit genes in ΔclpX cells. Wild-type and ΔclpX cells were grown under LL. After two 12-h light, 12-h dark cycles, the cells were returned to LL. The cells were collected at LL4 and LL16, and total RNA was isolated. The mRNA levels of each sample were normalized to the average value for the WT cells at LL4, which was taken as 1. Three independent samples were analyzed for each experiment. Each sample was run in three or more PCRs. The vertical bar on each column indicates the standard error. Levels of statistical significance between the wild type and ΔclpX mutant, identified by Student's t test, are indicated with asterisks: ∗, P < 0.05; ∗∗∗, P < 0.001.

Because the deletion of clpX prolonged the period of gene expression, it is assumed that translation activity, which is closely related to r-proteins, is linked to the regulation of the circadian period. To examine the effect of perturbed translation on the circadian rhythm, we measured the kaiBC cell bioluminescence rhythm in the presence of a low concentration of the translation inhibitor chloramphenicol. Wild-type cells were grown on chloramphenicol-free agar plates under LL until colonies formed and were then treated with 25 ng ml⁻¹ chloramphenicol. The period of the rhythm became long after the addition of chloramphenicol (Fig. 6). This result is consistent with the period lengthening in ΔclpX cells, in which the expression levels of r-protein subunits were downregulated.

Fig 6.

The effect of the addition of a translational inhibitor at a low concentration on the period of kaiBC promoter activity. Wild-type cells were grown on chloramphenicol-free agar medium and treated with 25 ng ml⁻¹ (red circles) chloramphenicol at LL52 or not treated (black circles). The lower panels show enlarged bioluminescence intensity scales (×10). The arrow indicates chloramphenicol addition. The peak bioluminescence intensity for kaiBC promoter activity was normalized to 1. The average increase in the period length was 1.3 h (n = 4; P < 0.005).

DISCUSSION

In this study, we showed that the deletion of clpP1, clpP2, or clpX, which together function as a complex, results in an extension of the circadian period in Synechococcus (Fig. 1A). We also evaluated the effect of clp gene overexpression. In OX-clpX cells, the promoter activity of kaiBC, kaiA, and psbAII initially increased before decreasing and displaying a disrupted rhythm (Fig. 3A and D). The microarray and QRT-PCR analyses showed that clpX overexpression upregulated the transcription of r-protein subunit genes and subsequently upregulated other genes, including kaiBC and rpoA (Fig. 4). Moreover, the deletion of clpX caused the downregulation of the expression of r-protein subunit genes (Fig. 5). The effect of clpX deletion was not as strong as that of the overexpression of clpX. The constitutive inactivation of clpX might have been complemented by ClpC, which is another ATPase in the Clp family. Finally, we showed that the addition of a low concentration of a translation inhibitor increased the period length (Fig. 6). We added chloramphenicol at a low concentration at which the cells would be viable, but growth was slower than that of wild-type cells, and it is possible that the slow growth affected the period length. However, we also showed that the period of the rhythm was stable even if the growth rate changed (Table 1; see Fig. S1 in the supplemental material). Based on the long-period phenotype of ΔclpX, in which the expression of r-protein subunits was downregulated, we hypothesized that the inhibition of protein synthesis caused the period to lengthen. Our results suggest that translational efficiency affects the circadian period and that the clp protease family participates in the control of translation efficiency by regulating the transcription of r-protein genes.

Our Western blotting results did not show remarkable changes in the accumulation or balance of KaiA, KaiB, and KaiC in ΔclpP1, ΔclpP2, and ΔclpX cells (Fig. 1B); however, protein accumulation is controlled by the balance between synthesis and degradation. Moreover, the expression of kaiBC is inhibited by KaiC as part of an autoregulatory negative-feedback loop (2). Thus, the result of our Western blotting may not reflect a change in protein stability; instead, it may reflect a change in the balance of transcription, translation, and degradation. It is possible that ClpXP might degrade KaiC directly or indirectly, and it is necessary to confirm the participation of clock protein turnover in clp mutants in the future.

In this study, we demonstrated the importance of wild-type levels of Clp protease family members for the maintenance of normal circadian periodicity. We found that clpP3 overexpression extended the period of kaiBC promoter activity but that cells cooverexpressing clpP3 and clpR, which are coexpressed from the same operon, had a wild-type rhythm (Fig. 2B). Note that OX-clpP2-clpX cells exhibited a prolonged PpsbAI rhythm (22), although OX-clpX cells did not survive and OX-clpP2 cells showed no change in the rhythm or growth phenotypes (Fig. 2 and 3A; see Fig. S1 in the supplemental material). We also found that the deletion of clpX resulted in stronger growth and period phenotypes than the deletion of clpP1 or clpP2, whose corresponding proteins are partners of ClpX (Fig. 1). This result suggests that there is a compensation effect with ClpP1 and ClpP2. In addition, the period of the ΔclpP1 mutant was almost same as that of the ΔclpP2 mutant (Fig. 1A), and the ClpP2 protein was undetectable and the ClpR and ClpP3 protein levels increased in ΔclpP1 mutants (24). It is likely that the increases in ClpR and/or ClpP3 compensate for the absence of ClpP2. These compensatory effects between ClpP1, ClpP2, ClpP3, and/or ClpR may be important for the determination of the circadian period, as they are for Synechococcus viability (24). Interestingly, the expression of clpX peaks at subjective dusk, whereas the expression of clpP1, clpP2, clpR, and clpP3 peaks at subjective dawn under LL (32). Taken together, these results suggest that the balance of the Clp protease subunits and their chaperone partners is important for maintaining normal biological activity, including circadian periodicity. Further research is necessary to fully understand how the Clp subunit levels alter periodicity.

We found that the r-protein subunit genes from some operons were upregulated following the overexpression of clpX, which suggests a novel function for ClpX. The mechanism of upregulation of the r-protein genes is not clear, but one conceivable possibility is that the ClpX-associated protease directly or indirectly degrades a factor or factors that suppress r-protein gene transcription. Because the amount of r-protein is controlled by autofeedback regulation (34) and because some r-proteins have been found to interact with ClpXP (35), the r-protein itself may be degraded by the ClpXP protease.

The possible involvement of ClpX in controlling cell division and septum formation is also worth considering. The deletion of clpX and clpP1 in Synechococcus results in filamentation (Fig. 1D) (26). The production of ribosomes is one of the major programmed cell events in Synechococcus, and the synthesis of r-proteins occurs early in the cell cycle, followed by rRNA synthesis, genome replication, genome segregation, and cell septum formation (36, 37). Our results suggest that ClpX regulates the early stages of cell division by adjusting the transcription levels of r-protein genes. In addition, ClpX participates in the regulation of FtsZ stability, which is regulated by the circadian clock. Cell septum assembly is regulated by the circadian cycle (38). Ring localization and the assembly of FtsZ are subject to regulation by circadian clock machinery, thereby pacing cell septum formation (39). In Escherichia coli, ClpX disassembles FtsZ polymers, presumably by blocking the reassembly of FtsZ, and the overexpression of the ClpXP complex causes filamentation by increasing FtsZ degradation (40). These observations imply that ClpX participates in the degradation of FtsZ, thereby suppressing cell septum formation in Synechococcus.

In this study, we showed that Clp proteases affect the period of the circadian clock in Synechococcus. ClpX is important for the transcription of r-protein genes. Moreover, we suggest that the rhythm period might be affected by translational efficiency. Other clp genes whose deletion or overexpression leads to increased period length appear to participate in the regulation of the clock period, although the mechanism responsible for this behavior remains to be elucidated. Further analysis is needed to determine how other Clp protease family members regulate the cyanobacterial circadian clock system.