Screening for mutants of ROC15 light response

Mutants were generated by random insertional mutagenesis of the ROC15-LUC reporter strain using the hygromycin resistance gene, aph7” [29]. Approximately 4700 transformants were subjected to two cycles of a 6 h dark/18 h light schedule (one red and one blue light cycle). Under these conditions, the wild type (WT) ROC15 bioluminescence levels increased in the dark period, decreased acutely at the start of light-on conditions, and remained low until the end of the light period (Fig 1A, WT). These patterns reflected the expression levels of ROC15 [27]. Two mutants (tentatively named M1 and M2) were isolated. The first mutant (M1) showed a gradual decrease instead of an acute one in ROC15 bioluminescence levels after the transition to the light phase (Fig 1A). In addition, M1 bioluminescence levels failed to remain low until the start of the dark phase (Fig 1A). These altered patterns of ROC15-LUC bioluminescence were observed in both the red and blue cycles of screening (Fig 1A). The ROC15 light response in the mutant was further examined by exposing M1 to 5-min pulses of red and blue light. Unlike the WT, M1 failed to show an acute decrease in ROC15 bioluminescence levels in response to both light pulses (Fig 1B and 1C). Taken together, these results suggested that the light-dependent degradation of ROC15 was affected in the mutant. The second mutant (M2) also showed an altered ROC15 bioluminescence response to both red and blue lights, similar to the M1 phenotype (S1 Fig). In conclusion, we were able to isolate two light response mutants that showed impaired ROC15 bioluminescence responses to both red and blue light.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Representative bioluminescence traces of the M1 mutant.

A. ROC15-LUC bioluminescence pattern of M1 under diurnal conditions. The red, blue, and black bars above the graph represent red light (8 μmol m^-2 s^-1), blue light (20 μmol m^-2 s^-1), and darkness, respectively. Cells were prepared for the screening as described in Materials and Methods. Mean ± standard deviation (SD) of 32–36 biological replicates are plotted. B and C. Representative trace of the ROC15-LUC bioluminescence response to light pulses in M1 mutant. Unsynchronized TAP cultures were transferred into 24 well black plates, maintained in darkness for at least 6 hours to allow for the accumulation of ROC15-LUC, and exposed to a 5 min pulse of red light (B, 10 μmol m^-2 s^-1) or blue light (C, 20 μmol m^-2 s^-1). Bioluminescence has been calculated relative to that at the time point just before light pulse (Time 0). Background was not subtracted. Mean ± SD of 6 biological replicates are shown. Arrows indicate the approximate time points of light pulses. The increase in bioluminescence levels after the light pulse probably corresponds to the disturbance of the plate during the experiment.

https://doi.org/10.1371/journal.pgen.1010449.g001

Gene disrupted in the M1 mutant

To identify the disrupted genes in the mutants, we performed a genetic linkage analysis to confirm that the mutant phenotype was linked to hygromycin resistance. This was verified by backcrossing the mutants with the parental WT strain (ROC15-LUC reporter strain). In mutant M1, all hygromycin-resistant progeny failed to show an acute decrease in ROC15 bioluminescence after light-on, whereas all the hygromycin-sensitive progeny showed an acute decrease (S2A and S2B Fig). These results confirmed the genetic linkage of hygromycin resistance to the mutant phenotype. However, this pattern was not observed in the second mutant (M2) (S2C and S2D Fig). Therefore, we continued our investigations only in the M1 mutant.

In M1, the DNA flanking the hygromycin resistance marker was amplified by thermal asymmetric interlaced–polymerase chain reaction (TAIL-PCR) [30]. This sequence was compared to the C. reinhardtii reference genome (C. reinhardtii v5.6, Joint Genome Institute), and the Cre07.g357500 gene on chromosome 7 was found to be disrupted (Corresponds to the ID Cre07.g800875 in the latest version (v6.1)). This gene appeared to have eight exons, and the insertion of the hygromycin marker was found to be in the third intron (Fig 2A). The insertion was confirmed by PCR using a gene-specific primer pair (S3A and S3B Fig). We further performed a reverse transcription-PCR (RT-PCR) on the transcript (S3A Fig). No band with a WT band size was detected in the mutant. The two bands detected could possibly correspond to mutant transcripts resulting from unexpected alternative splicing due to the insertion of the marker (S3C Fig). We also performed a sequencing analysis of the RT-PCR product of the entire coding sequence (CDS) of this gene from our laboratory WT strain. The results revealed that compared to the database sequence (v5.6), our laboratory strain had an in-frame 264 nucleotide insertion in the fifth exon (Fig 2A, purple box). Mutants with similar phenotypes (i.e., impaired ROC15 light response to both red and blue wavelengths) had been isolated previously (b19 and b20 mutants) [28]. TAIL-PCR analysis of b19 revealed that the same gene (Cre07.g357500) was disrupted (Fig 2A). The insertion locus of the hygromycin marker was in the fifth intron in the b19 mutant (Figs 2A, S3A, and S3B).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Gene responsible for the mutant phenotype.

A. Schematic representation of Cre07.g357500 and its corresponding transcript. 5’UTR and 3’UTR ends of the transcript are illustrated based on information from the database (C. reinhardtii v5.6, Joint Genome Institute). This gene corresponds to Cre07.g800875 in the latest version (v6.1). The N-terminus appears to be missing in the gene prediction of this version. M1 (cetl-1) and b19 (cetl-2) indicate the insertion loci of aph7” in the mutants. The purple box indicates the location of the additional 264 nucleotides present in our laboratory strain. The location of these additional nucleotides in the transcript and the location of the corresponding amino acids in the encoded protein are also indicated by a purple box. B. Protein sequence alignment. A schematic depiction of amino acid sequences, representing the sequences from the two phyla used for alignment. The two regions of similarity (Region 1 and Region 2) are also shown. The patterned red box on the Arabidopsis ELF3 sequence represents the region that was shown to be responsible for interaction with ELF4 [33]. The patterned blue boxes beneath the Arabidopsis ELF3 sequence represents the regions identified as Block I, II, III and IV [11]. C. Alignment of the amino acid sequences at Region 1. The amino acids have been colored according to the Clustal X -scheme i.e. according to their properties and conservation. The red inverted triangle indicates conservation at the Glycine residue, which was found to be substituted in the elf3-12 mutant [32]. The colored bars above the alignment indicate the percentage of conservation. Sequences were aligned using MAFFT multiple sequence alignment program (MAFFT 7.471, SnapGene). The details of the amino acid sequences used for the alignment are listed in S2 Table.

https://doi.org/10.1371/journal.pgen.1010449.g002

To confirm that the disruption of the Cre07.g357500 gene caused the altered response to light in mutants, the M1 mutant was transformed with the WT gene (S4A Fig) to determine whether the ROC15 light response would be restored to WT levels. A total of 654 transformants (414 transformed with the WT gene and 240 transformed with only the antibiotic resistance cassette as a control group) were subjected to two cycles of a 6 h dark/18 h light schedule, with the first and second light phases consisting of red and blue light, respectively (S4B Fig). The ROC15 light response to both blue and red wavelengths was restored to WT levels in 6 of 414 transformants that had been transformed with the WT Cre07.g357500 gene (i.e., ROC15 bioluminescence levels declined steeply in response to light) (S4B and S4C Fig). The complementation rate was comparable to the rates observed in other mutants in previous studies [26]. In contrast, 0 of the 240 control transformants showed a WT ROC15 light response (S4D Fig). The ROC15 light responses of the six complements were further confirmed via 5-min pulses of red and blue light. The ROC15 bioluminescence levels decreased acutely in response to red and blue light pulses in all six complements (S4E and S4F Fig). These results confirmed that the disruption of the Cre07.g357500 caused the altered response to light in the mutant.

Protein sequence alignment

The longest open reading frame of the Cre07.g357500 gene of our laboratory strain encodes a protein with an expected length of 2392 amino acids. This protein has not been well characterized in Chlamydomonas. Therefore, we performed a BLAST search of this protein sequence against Archaeplastida proteins via the algal multi-omics portal, PhycoCosm [31]. Using the default BLAST settings (e-value threshold = 1 × 10⁻⁵), homologs were detected only in members of Chlorophyceae. Some of the organisms that showed homologs included other species of Chlamydomonas (such as C. incerta and C. schloesseri), other members of the order Chlamydomonadales (such as Volvox carteri, Gonium pectorale, and Edaphochlamys debaryana), and organisms belonging to the order Sphaeropleales (such as Scenedesmus obliquus). BLAST was performed once again with a lower E-value threshold (= 1). This resulted in hits in many more chlorophytes, including members of Trebouxiophyceae (such as Chlorella sorokiniana), Mamiellophyceae (such as Micromonas commoda), and members of another phylum—Streptophyta, such as Selaginella moellendorfii and Populus trichocarpa. Surprisingly, the streptophyte proteins included homologs of the A. thaliana ELF3. Additionally, a BLAST analysis of A. thaliana ELF3 against the C. reinhardtii proteome (v5.6_281) revealed the protein encoded by Cre07.g357500 as the best hit with an E-value of 0.016. Similarly, a BLAST analysis of the protein encoded by the gene Cre07.g357500 against the A. thaliana Araport11 protein sequences database, revealed ELF3 as the best hit with an E-value of 0.11. We then aligned 33 sequences (S2 Table) obtained from the BLAST searches and the ELF3 homologs of the major species in Streptophyta. The results revealed two regions of similarity (Region 1 and Region 2) (Figs 2B, 2C, and S5). Region 2 was found only in chlorophyte proteins (S5 Fig), whereas Region 1 was found in the proteins of both chlorophytes and streptophytes (Fig 2B and 2C). Region 1 happened to be part of a region in Arabidopsis ELF3 referred to as Block II, which is one of the four highly conserved regions (the other three being Block I, Block III and Block IV) in angiosperm ELF3 [11]. To further observe the conservation at all the Blocks, the protein encoded by Cre07. g357500 was aligned with A. thaliana ELF3 and its homologs from only other angiosperms using MAFFT and MUSCLE multiple sequence alignment algorithms. It was observed that while Block II aligned to the same sequence of the Cre07. g357500 protein in both algorithms, Blocks I, III, and IV aligned to slightly different sequences depending on the algorithm used (S6 Fig). This data only weakly suggests the possibility of Cre07.g357500 being a homolog of ELF3. However, it is important to note that the Glycine residue (Gly-326) in Region 1, which when mutated is known to affect the circadian function of ELF3 [32], was fully conserved in the proteins analyzed (Fig 2C). In addition, Region 1 was immediately adjacent to the region that has been shown to interact with ELF4 in Arabidopsis [33] (Fig 2B, patterned red box). Therefore, we hereafter refer to this gene as Chlamydomonas-ELF3-like (CETL), and to the M1 and b19 mutants as cetl-1 and cetl-2, respectively.

CETL expression analyses

We analyzed whether CETL transcript level was rhythmic under constant light conditions (LL). The results of a reverse transcription-quantitative PCR (RT-qPCR) analysis revealed that the transcript-level was rhythmic, with a peak at approximately subjective dusk or early subjective night (Fig 3A). This result indicated that CETL gene expression was evening-phased due to circadian control.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. CETL gene expression pattern in LL conditions, CETL protein light response and subcellular localization.

A. CETL mRNA rhythm in LL conditions. A free-running high-salt culture of the wild type (WT; ROC15-LUC mt⁺) was used. Cells were harvested every 4 hours between 24–72 hours in LL conditions and then subjected to RT-qPCR analysis. A representative trace of CETL mRNA levels (normalized by RCK1 mRNA levels) is shown. The white bar above the graph represents the light conditions (i.e., LL conditions). The alternating white and grey background in the graph represents the subjective day and subjective night, respectively, that is expected from the light conditions prior to LL. B. CETL light response. High-salt cultures of complemented cells expressing HA-CETL were used. Cultures were maintained in LD conditions and exposed to 5 min pulses of red (10 μmol m^-2 s^-1) and blue (20 μmol m^-2 s^-1) light at ZT18. Western blot analysis was performed on whole protein extracts of cells harvested at ZT6 during the light phase (L), in the darkness before light pulse (0 min), and, 10 and 30 minutes after exposure to light pulses. Dark control was sampled at the same time points as the light pulsed samples. α-tubulin was used as loading control. C. CETL subcellular localization in comparison to ROC15 localization. Unsynchronized TAP cultures were maintained in darkness for 6 hours before cells were harvested in darkness and subjected to immunocytochemistry. The nuclear pore complex (NPC) was stained as a marker for the nuclear membrane. The ROC15-HA strain was used to observe ROC15 localization. The BR mt⁺ strain was used as a negative control. Results for a representative strain are shown. Scale bar: 2μm.

https://doi.org/10.1371/journal.pgen.1010449.g003

To assess the CETL protein expression, the mutant (cetl-1) was transformed with a gene fragment encoding hemagglutinin (HA) tagged CETL (S7A and S7B Fig). A total of 462 transformants were screened for restoration of ROC15 response to 6 h dark/18 h red light cycles (S7C Fig). ROC15 response was restored to WT levels in six of the 462 transformants (S7D Fig). HA-CETL expression was then analyzed in these six transformants by western blot analysis. A band of size 237kDa corresponding to HA-CETL was detected in five of the six complements (S7E Fig). The ROC15 light response of these five complements was further confirmed by exposure to 5-min pulses of red and blue light. The ROC15 bioluminescence levels decreased acutely in response to red and blue light in all five complements (S7F and S7G Fig), indicating that the HA tagged CETL was functional with respect to ROC15 light induced degradation. CETL protein expression was then examined in one of these five complements by a western blot analysis. Bands corresponding to HA-CETL were detected in the light phase in a LD cycle and even 30 minutes after exposure to red and blue light pulses in the dark phase (Fig 3B). These results suggest that CETL does not show ROC15-like responses to either red or blue light. Immunocytochemical staining was also performed in one of the five complements by using anti-HA antibody. The HA-CETL signals were detected just inside the Nuclear Pore complex (NPC) signal, similar to where the ROC15-HA signals were detected (Fig 3C). These results suggest that CETL localizes in close proximity to ROC15.

ROC15-LUC light response in the cetl-1 mutant

The cetl-1 mutant showed an impaired response to red and blue light (Fig 1). To understand the wavelength specificity of this mutant, its ROC15 light response was further characterized by testing other wavelengths (i.e., violet and yellow). The mutant and WT were exposed to a light pulse of four different wavelengths at two different intensities. At an intensity of 2 μmol m^-2 s^-1, the WT showed an acute decrease in ROC15 bioluminescence levels at all wavelengths. In contrast, the mutant failed to show this acute decrease (Fig 4A). At a lower intensity of 0.2 μmol m^-2 s^-1, the WT cells showed an acute decrease in ROC15 bioluminescence levels in response to red and violet wavelengths, and a slight decrease in ROC15 bioluminescence levels in response to yellow and blue wavelengths (Fig 4B). This result was consistent with those of previous studies [27,28]. However, the mutant failed to show these decreases (Fig 4B). These results demonstrate that the mutant showed defective responses to violet and yellow wavelengths in addition to defective responses to blue and red wavelengths. Therefore, the ROC15 light response in the mutant was impaired across the visible range.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. ROC15-LUC light response at different wavelengths in the cetl-1 mutant.

Spot cultures on high-salt agar were transferred to 96 well black plates, and bioluminescence was monitored as described in Materials and Methods. Akalumine was used as substrate. Cultures were maintained in darkness for at least 3 hours to allow for the build-up of ROC15-LUC, and then exposed to a 5-minute light pulse (violet, blue, yellow, or red) at two different intensities. A and B. Representative traces of ROC15-LUC bioluminescence in response to light pulses at intensities of 2 μmol m^-2 s^-1 and 0.2 μmol m^-2 s^-1, respectively, in WT (top panel) and cetl-1 (bottom panel). Bioluminescence values have been calculated relative to those at the time point just before the light pulse. Mean ± SD of 4 biological replicates are shown. Arrows refer to the approximate time of the light pulse. Each color represents the wavelength of light to which the sample was exposed.

https://doi.org/10.1371/journal.pgen.1010449.g004

The effect of the cetl-1 mutation on light-induced degradation of ROC15 was further observed in a csl mutant genetic background. The cetl-1/csl double mutant strain was obtained via a genetic cross. The double mutant was exposed to three different wavelengths (red, blue, and violet) at two different intensities. At both 2 μmol m^-2 s^-1 and 0.2 μmol m^-2 s^-1, the cetl-1/csl double mutant failed to show the ROC15-LUC bioluminescence decreases observed in the WT in response to all the wavelengths tested (S8A and S8B Fig). These results indicate that the blue light response in csl was lost due to the cetl-1 mutation [28].

Intracellular localization and light-induced phosphorylation of ROC15 in the cetl-1 mutant

Previous studies have suggested that ROC15 localizes to the nucleus and undergoes light-induced phosphorylation [27]. To investigate these characteristics of ROC15 in the cetl-1 mutant, the ROC15-HA genetic construct was introduced into the mutant genetic background via a genetic cross. To investigate ROC15 localization, we performed immunocytochemical staining with an anti-HA antibody. The results showed that ROC15-HA signals were detected immediately inside the NPC signals in both the WT and mutant cells (Figs 5A and S9A). This suggests that the cetl-1 mutation did not affect the localization of the ROC15 protein. However, it was observed that the ROC15 expression between mutant cells was not as uniform as between WT cells (S9A Fig). This is probably due to the weak ability of this mutant to respond to synchronizing light/dark conditions (See below). This was followed by an analysis of ROC15 light-induced phosphorylation in response to red and blue lights. ROC15 phosphorylation has been reported to be visible as an electrophoretic mobility shift of ROC15-HA, as the shift was not observed when cell-extracts were subjected to phosphatase treatment [27]. This electrophoretic mobility shift was not observed in the mutant in response to red and blue light pulses (Fig 5B). The mobility shift was restored in the complement strains in both cases (S9B and S9C Fig), indicating that CETL was required for ROC15 to undergo light-induced phosphorylation.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Subcellular localization and phosphorylation of the ROC15 protein in the cetl-1 mutant.

The complement (S4 Fig) was crossed with the ROC15-HA strain (wild type [WT]) to obtain progenies that inherited ROC15-HA (but not ROC15-LUC) along with the cetl-1 mutation (cetl-1). Progenies were selected by antibiotic screening and genomic PCRs. A. ROC15 localization in the mutant. Unsynchronized TAP cultures were maintained in darkness for at least 6 hours to allow for accumulation of ROC15-HA. Cells were harvested in darkness and then subjected to immunocytochemistry. The NPC was counterstained as a marker for the nuclear membrane. The BR mt⁺ strain was used as a negative control. Scale bar: 2μm. B. Absence of light-induced ROC15 phosphorylation in the mutant. Cells in LD-entrained high-salt cultures were exposed to a 0.5-min light pulse of red (2 μmol m^-2 s^-1) and blue (20 μmol m^-2 s^-1) wavelengths at midnight (between ZT14 and ZT17). Western blot analysis was performed on whole protein extracts from cells sampled before the light pulse and at approximately 5, 10, and 15 minutes after the light pulse. The 0 min data corresponds to the sample just before the light pulse. WT in this experiment refers to a wild type sibling from the cross. α–tubulin was used as the loading control.

https://doi.org/10.1371/journal.pgen.1010449.g005

Circadian rhythm and resetting in the cetl-1 mutant

To characterize the mutant, we observed its circadian rhythm in constant darkness (DD) and LL conditions. We introduced the chloroplast luciferase reporter gene (tufA promoter-lucCP; lucCP is a firefly luciferase gene that is codon-optimized for the C. reinhardtii chloroplast) [34] of the CBR strain [26] into a mutant genetic background via a genetic cross. A stable rhythm was detected in the cetl-1 mutant under both DD and LL conditions (Fig 6A), with a slight tendency toward longer period lengths and lower amplitudes (Fig 6B). Similar results were observed in the cetl-2 mutant under DD and LL conditions (S10A and S10B Fig). This result suggested that CETL did not play a major role in the oscillator of the C. reinhardtii circadian clock. However, the peak phase of cetl-1 showed a wider distribution with a tendency to advance the phase, especially under LL conditions (Fig 6C). This is again close to what was observed in the cetl-2 mutant under LL conditions (S10C Fig). As observed in the cetl-1 mutant, the variation in phase was pronounced between independent trials of a single progeny, and also between progenies within a single trial, rather than between biological replicates of a single progeny within a single trial (S11 and S12 Figs). These unstable phase phenotypes might be indications of insufficient phase resetting in the mutants. Therefore, we investigated phase resetting using light pulses. The cetl-1 mutant and WT were maintained in DD conditions and exposed to light pulses of four different wavelengths at late subjective night. The WT showed an advance in the phase in response to all four wavelengths, consistent with previous reports [28]. Although the cetl-1 mutant also showed a slight phase advance, the advance was significantly reduced compared to that of the WT (Fig 6D). These results indicated that phase resetting was impaired in the cetl-1 mutant.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Circadian rhythm under DD and LL conditions and light-induced phase resetting.

The mating type^- (mt^-) strain of the cetl-1 mutant (S2 Fig) and the complement (S4 Fig) were crossed with the CBR strain (mt⁺) (wild type [WT]) to obtain progenies that inherited the chloroplast luciferase reporter (but not the ROC15-LUC) along with cetl-1 mutation (1–3), or along with both the cetl-1 mutation and the CETL transgene (Compl.). These progenies were selected by antibiotic screening and genomic PCRs. Spot cultures of these progenies were synchronized by a 12-hour dark/12-hour light or 12-hour darkness before release into DD or LL conditions (2 μmol m^-2 s^-1), respectively. A. Representative traces of chloroplast bioluminescence rhythm in the cetl-1 mutant under DD (top) and LL (bottom) conditions. Data of 3–5 biological replicates of the WT, a cetl-1 progeny, and a complemented progeny from a single trial are shown. The light regime used is also depicted above the graphs. B and C. The period and amplitude (B) and phase diagram indicating the circadian phases (C) of the rhythms under DD (top) and LL (bottom) conditions. Data points correspond to the results of biological replicates of the WT and three cetl-1 progenies across four trials, and of three complemented progenies from two trials (Refer to S11 Fig for the phase diagrams of individual cetl-1 progeny from four different trials). The circadian times (CT) at the beginning of the DD and LL conditions were set to 12 and 0, respectively. D. Light-induced phase resetting under DD condition. Representative bioluminescence traces (left panel) and the amount of phase shift after the light pulse (right panel) are shown. A WT sibling (WTp) from the cross was used as WT for this experiment. Spot cultures were exposed to a 5-minute pulse of violet (5 μmol m^-2 s^-1), blue (38 μmol m^-2 s^-1), yellow (20 μmol m^-2 s^-1), and red (15 μmol m^-2s ^-1) light at approximately the 34^th hour after the onset of DD conditions. The amount of phase shift was calculated with respect to the dark control. Mean ± SD of 6–8 biological replicates are shown. P-values are based on Student’s t-tests: * P<0.05, ** P< 0.001, and *** P<0.0001.

https://doi.org/10.1371/journal.pgen.1010449.g006

Phase distribution in the single colony protocol in the cetl-1 mutant

To further examine the phase resetting in the mutant, we also analyzed the chloroplast bioluminescence rhythm of cultures that had been prepared in a manner different to that of our standard protocol. The biological replicates for the current experiment originated from independent colonies (single colony protocol), as opposed to our standard protocol, where the biological replicates were obtained from a single patch on storage plates (S13 Fig). The quadruplicate spots from an independent culture of the single colony protocol were exposed to either the same light schedule as in Fig 6A (Fig 7Ai) or to additional progressive LD cycles (Fig 7Aii, 7Aiii and 7Aiv) before monitoring in LL conditions. Surprisingly, after light schedule i, the peak phases appeared to vary largely between the biological replicates of cetl-1 (Fig 7B). When visualized as a phase diagram, the wide distribution of peak phases was evident (Fig 7C). We also noted that the distribution of peak phases of cetl-1 varied with trials, and sometimes appeared more synchronized (S14A and S14B Fig). However, we again observed that the period and amplitude of the rhythm of the mutant were close to those of the WT in this protocol (Figs 7D and S14C). After each additional light/dark cycle (Fig 7Aii, 7Aiii, and 7Aiv), the peak phases of the rhythms in the mutant cultures were detected much closer to those of the WT (Fig 7E). Collectively, these results suggested that the mutant was severely affected in its ability to reset the clock, but that the ability is not completely lost.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. Phase distribution in LL in single colony protocol in cetl-1.

The CBR (mt⁺) (wild type [WT]), mutant progeny (cetl-1), and complement progeny (Compl.) from the cross mentioned in Fig 6 were used to observe the chloroplast bioluminescence rhythm. Spot cultures were prepared using the single colony protocol depicted in S13 Fig. A. Light regimes (i-iv) used for the experiment. Colonies were exposed to progressive LD (12 hours:12 hours) cycle before release into LL conditions (ii-iv). B. Chloroplast bioluminescence rhythms of the WT, cetl-1, and Compl. as observed after light schedule i. Cultures were maintained in 12 hours of darkness before release into LL conditions. Data of 28 biological replicates each, of WT and Compl., and 32 biological replicates of cetl-1 from a single trial are shown. C. Phase diagram showing the circadian phases of cetl-1, WT and Compl. rhythms shown in B. D Plot of period and amplitude of the bioluminescence rhythms shown in B. E. Peak phases of the quadruplicate spots of biological replicates after each of the four light schedules. The results after light schedules i-iv are plotted from the outermost ring to the innermost ring, respectively. Quadruplicate spots from the same independent culture are connected by a line. Data are shown only for cases where the rhythms were detected for all four spots in an independent culture, and correspond to results of a single trial.

https://doi.org/10.1371/journal.pgen.1010449.g007

Light responses of clock gene mRNAs in the cetl-1 mutant

Previous studies have demonstrated that the mRNA of some clock genes, such as those of ROC15 and ROC40, are downregulated in response to light [20,27,28,35]. Here, we investigated this phenomenon in the cetl-1 mutant. RT-qPCR was performed on samples from WT and cetl-1 cultures exposed to light for 1 h in the late night. In WT, the mRNA levels of ROC15 and ROC40 were much lower in the light-exposed samples than in the dark controls (Fig 8A and 8B). The mRNA levels of the ROC15 and ROC40 genes also appeared to be downregulated in the light-exposed sample of cetl-1 (Fig 8A and 8B). However, the decline in mRNA levels was much lesser in the cetl-1 mutant than in the WT (Fig 8C). The weak responses of both the mRNAs were restored to WT levels in the complement strain (S15 Fig). These results suggested that CETL played a role in the light responses of ROC15 and ROC40 mRNAs.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 8. Light response of clock gene mRNAs in the cetl-1 mutant.

LD-entrained high-salt cultures were exposed to 30 μmol m^-2 s^-1 of white light at late night (ZT22, 2 hours before dawn). Cells were harvested from the light-exposed culture and the corresponding dark control after 1 hour. A and B. The mRNA levels of ROC15 (A) and ROC40 (B) in the wild type (WT; ROC15-LUC strain) and cetl-1. The mRNA levels have been calculated relative to the mRNA levels of RCK1. A two-way ANOVA indicated a significant interaction between genotype and light conditions for ROC15 (genotype: F_1,8 = 0.46, P = 0.52, light conditions: F_1,8 = 54.61, P<0.0001, interaction: F_1,8 = 10.24, P< 0.05) and for ROC40 (genotype: F_1,8 = 12.46, P = 0.0077, light conditions: F_1,8 = 1199.57, P<0.0001, interaction: F_1,8 = 272.17, P< 0.0001). Tukey’s post hoc test was performed on the samples which were subdivided into four groups: WT dark control, WT light exposed, cetl-1 dark control, and cetl-1 light exposed. a (P<0.05) for comparisons with WT dark control, b (P<0.05) for comparisons with WT light exposed sample, c (P<0.05) for comparisons with cetl-1 dark control. C. Extent of downregulation in WT and cetl-1. The mRNA abundance (i.e., mRNA levels relative to RCK1 mRNA levels) of the light-exposed samples are shown relative to that of their respective dark control for easy comparison. P-values are based on Student’s t- tests. ** P< 0.01. Mean ± SD of three biological replicates is shown (A-C). Note that these strains possess two copies of ROC15: an endogenous copy and ROC15-LUC transgene. The ROC15 mRNA levels detected in this experiment are therefore a sum of the mRNA levels of the endogenous copy and the mRNA levels of the ROC15-LUC transgene.

https://doi.org/10.1371/journal.pgen.1010449.g008

Circadian rhythm of clock genes in the cetl-1 mutant

The mRNAs of certain clock genes have been shown to be rhythmic under LL conditions [26]. We examined whether the cetl-1 mutation had any effect on the expression rhythm of clock genes (ROC15, ROC40, ROC66, and ROC75) under LL conditions. The RT-qPCR analysis (S16 Fig) revealed that all four clock genes tested in the mutant had rhythmic gene expression patterns. However, some differences were observed in certain cases. The level of ROC15 mRNA appeared to be slightly lower in cetl-1 mutants, whereas the ROC40 mRNA levels appeared to be slightly higher, especially during the subjective day (S16A and S16B Fig). The phase of expression of ROC40 and ROC66 appeared to be slightly delayed in cetl-1 mutants (S16B and S16D Fig). These results suggested that CETL might not have a great influence on the transcriptional oscillator.

Light signaling pathway for ROC15 degradation

Three possible models for light induced ROC15 degradation are depicted in S17 Fig. A previous study indicated that a wide range of wavelengths were capable of causing ROC15 degradation [27]. The CSL protein, which mainly localizes in the cytoplasm, was indicated to be specific to the red/violet light signalling pathway upstream of ROC15 phosphorylation [28] (S17 Fig). In contrast, the clock protein ROC114, which localizes in the nucleus, was shown to be common to the different light signalling pathways downstream of ROC15 phosphorylation [27] (S17 Fig). In this study, we revealed that CETL, another protein which appears to localize in the nucleus (Fig 3C), is also common to the different light signalling pathways (Fig 4). In addition, impairment of the light induced phosphorylation in the cetl-1 mutant (Figs 5B, S9B and S9C) indicates that unlike ROC114, CETL plays a role upstream of ROC15 phosphorylation (S17 Fig). Furthermore, since the blue light response that remained in the csl mutant was completely lost in the csl/cetl double mutant (S8 Fig), it seems likely that these light signals are integrated downstream of CSL. However, it is still unclear whether the different light signalling pathways are integrated by CETL (S17A Fig) or another unidentified component X upstream of CETL (S17B Fig). In addition, there also exists a possibility where CETL is essential for the expression of other unknown components (X and Y) that are upstream of ROC15 phosphorylation (S17C Fig).

Mechanisms contributing to light resetting in Chlamydomonas

The light resetting ability of the circadian clock was not completely lost in the cetl-1 mutant (Figs 6 and 7) despite a defect in the light-induced degradation of ROC15 in response to a wide range of wavelengths (Fig 4). This pattern points to the possibility of other mechanisms regulating the light-induced resetting of the Chlamydomonas circadian clock. One such mechanism is the light response of the mRNAs of clock genes, such as C3, ROC15, ROC40, and ROC75 [20, 27, 28, 35]. The light-induced downregulation of ROC15 and ROC40 mRNAs was diminished in the cetl-1 mutant (Fig 8), which is in contrast to the trend observed in the csl mutant [28]. In the csl mutant, the red light-induced degradation of the ROC15 protein was impaired, whereas the light responses of the ROC15 and ROC40 mRNAs were similar to those of the WT under strong red light conditions (200 μmol m^-2 s^-1) [28]. We cannot exclude the possibility that the csl mutant may show a defect in the mRNA response at the weaker light intensities used in the current study (30 μmol m^-2 s^-1). However, the cetl-1 mutant is the first in which we have confirmed a defect in both light-induced ROC15 degradation and the light responses of clock gene mRNAs. It is intriguing that the cetl-1 mutant did not completely lose its resetting ability (Figs 6 and 7) even with both the potential resetting mechanisms affected. This suggests the existence of yet another contributing factor which is currently unknown.

Wide distribution of phases in the cetl-1 mutant

Interestingly, mutants obtained using the single colony protocol exhibited a wide distribution of phases in some cases (Fig 7). A recent study has demonstrated that a liquid culture of Chlamydomonas comprises cells in various circadian phases [36]. Another study that analyzed co-expression networks in Chlamydomonas further estimated that 21–96% of cells from a culture grown in LL conditions were synchronized [37]. It has also been shown that in mammalian cells, the circadian rhythm of mother cells is resumed in daughter cells after cell division [38]. In the single colony protocol, the colonies are therefore assumed to have a distribution of phases, with each colony possessing the phase of one of the cells in the liquid culture. This phase distribution should be reset by either the 12-hour dark period or the light/dark cycles (for the WT cells). We speculate that the wide distribution of phases observed in the mutant (Figs 7C and S14B) is a reflection of the phase distribution of the liquid culture, possibly due to the reduced ability of the mutant to be reset by the light/dark regime before release into LL conditions. This is contrary to the narrower distribution of phases observed within a single trial of the standard protocol (Fig 6). This difference may also suggest that cells in patch cultures had a more uniform phase than those in liquid culture. Furthermore, the variation in the distribution of phases between trials in the mutant (Figs 7C and S14B) may be attributed to the variation in the synchronization of cells in the liquid culture [37].

Comparison of CETL and ELF3

Conclusion

The possibility of the existence of an ELF3 homolog opens up many avenues of exploration that would enable further elucidation of the C. reinhardtii circadian clock. One such avenue is the existence of an EC in C. reinhardtii, since C. reinhardtii also has a potential homolog of the other EC component, ELF4 [39,40]. ELF3 has been shown to have thermo-sensing abilities due to a prion-like domain [33] and is shown to be involved in temperature entrainment of the A. thaliana clock [44]. Therefore, another possibility is that CETL plays a role in the integration of temperature signals to the clock. In C. reinhardtii, the C1 and C3 subunits of the clock-related protein complex, Chlamy1, can integrate temperature information into the circadian clock [49]. This may provide a gateway to study the involvement of CETL in temperature integration into the circadian clock. In conclusion, this study has revealed many pathways that could be explored to improve our understanding of temporal organization in Chlamydomonas and the evolution of circadian clocks in Viridiplantae.

Strains and media

We used the ROC15-LUC mating type ⁺ (mt⁺) and mating type ^- (mt^-) [27], CBR (mt⁺) [26], BR (mt⁺) [27], b19 (cetl-2) (mt⁺) [28], csl (mt⁺) [28], and ROC15-HA (mt⁺) [27] strains. BR is a reporter-less strain used to measure background noise during bioluminescence monitoring experiments. All strains were maintained on agar plates containing with Tris-acetate-phosphate (TAP) [50] medium and stored in constant white light conditions (10–20 μmol m^-2 s^-1) at 24°C (storage culture). The storage cultures were re-plated every 1–1.5 months. A high-salt (HS) medium [51] was also used in this study. All C. reinhardtii strains used in this study were derived from the CBR34 strain [26], which was obtained from a genetic cross between a CC2137-based reporter strain (tufA promoter-lucCP) [34] and the SAG11-32a WT strain.

Light sources

The light sources that were used in this study are detailed henceforth. LED tubes (red LED: LT20RS, 636 nm [full width at half maximum of 21 nm]; blue LED: LT20BS, 458 nm [21 nm], Beamtec, Saitama, Japan) were used for mutant screening (Figs 1, S1, S2, S4B and S7C). LED panels (red LED: ISL-150×150-RR, 660 nm [24 nm], yellow LED: ISL-150×150-YY, blue LED: ISL-150×150-BB, 470 nm [27 nm], violet LED: ISL-150×150-VV, 405 nm [14 nm], CCS, Kyoto, Japan) were used for light pulse and wavelength specificity experiments (Figs 1B, 1C, 3B, 4, 5B, 6D, S1B, S1C, S4E, S4F, S7F, S7G, S8, S9B and S9C). The wavelength of the yellow LED was measured using a multi-channel spectrometer (MC2100, Otsuka Electronics, Osaka, Japan). The peak was detected at 570nm [11nm]. White LED panels (MLP-LSK2478DA5, Musashi Electric, Saitama, Japan) and white LED tubes (LT-40KY-III, Beamtec, Saitama, Japan) were used as background light during rhythm assays (Figs 6A, 7, S10, S12 and S14). Light intensity was measured with a light meter (LI-250 equipped with LI-190, LI-COR, NE, USA; LA-105, NK systems, Osaka, Japan). Blue LEDs were used as safety lights at lower intensities that were undetectable by the light meters. Fluorescent tubes (FL20SSW/18-B, Hitachi, Tokyo, Japan) were used for culturing and mRNA analysis (Figs 3A, 8, S15 and S16).

Culture conditions

The three main types of cultures used in the study are detailed below.

1. Unsynchronized TAP cultures: Cells from the storage cultures were inoculated into TAP medium and maintained at 24°C in LL (30–40 μmol m^-2 s^-1) for 3–4 days (starter culture). Cells from starter cultures were then inoculated into fresh TAP medium to achieve a final concentration of 1 × 10⁵ cells/mL and maintained under the same conditions as the starter culture for 2 days.
2. LD-entrained and free-running HS cultures were prepared as previously described [35]. Briefly, starter cultures were inoculated into fresh HS medium at a concentration of 2 × 10⁵ cells/mL and maintained at 24°C in LL (30–40 μmol m^-2 s^-1) for 3 days. The temperature was then reduced to 17°C, and lights turned off for 12 hours to synchronize the circadian clock. After the synchronization period, the LD-entrained cultures were maintained in a 12-hour light (10 μmol m^-2 s^-1)/12-hour dark cycle, and free-running cultures were kept in LL conditions (10 μmol m^-2 s^-1).
3. Spot cultures for bioluminescence monitoring: Cells were cultured for 3 days in 96 well plates (Nunc MicroWell, Thermo Fisher Scientific, MA, USA) in 100 μL of TAP medium maintained at 24°C in LL conditions (30–40 μmol m^-2 s^-1). Following this, 5 μL of culture was spotted onto 1.5% HS agar plates (Agar BA-70, Ina Food Industry, Nagano, Japan was used for rhythm assays and Agar against dryness (LG), Kanto Chemical, Tokyo, Japan was used for light response assays). The spots were allowed to grow for 4 days at 24°C in LL conditions (30–40 μmol m^-2 s^-1).

Genetic crosses and TAIL-PCR

Genetic crosses and TAIL-PCR were performed as described previously [26].

Plasmid construction

To obtain the pLaadA/CETL plasmid, the blunt-ended 11.9 kb SacI/BamHI fragment from a bacterial artificial chromosome containing the CETL gene—obtained from the C9 strain—was sub-cloned into the EcoRV-digested pLaadA plasmid [28]. This fragment also contained the 264 nucleotide insertion. To obtain pLaadA/HA-CETL plasmid, an EcoRV restriction site was introduced just after the second predicted start codon of CETL in the pLaadA/CETL plasmid by PCR-based mutagenesis. A codon adapted HA sequence (S7B Fig) was inserted at the EcoRV site by using In-Fusion technology (TakaRa Bio, Shiga, Japan).

Bioluminescence monitoring

Cultures were transferred into 96 well white plates (Nunc F96 MicroWell, Thermo Fisher Scientific), 96 well black plates (STREIFEN-PLATTE [762076], Greiner Bio-One, Kremsmünster, Austria), or 24 well black plates (Krystal Microplates, Porvair Sciences, Norfolk, UK). In cases where spot cultures needed to be transferred, the spots on the agar were cut along with the agar and transferred into wells using a glass tube (inner diameter: 6 mm). Luciferase substrates were added to each well at a final concentration of 100 μM (liquid cultures) or approximately 200 μM (spot cultures). D-luciferin (Biosynth, Staad, Switzerland) was used unless specifically mentioned, and the luciferin analog Akalumine-HCL (FUJIFILM Wako Pure Chemical Corporation, Osaka, Japan) was used in some experiments [52]. Bioluminescence was monitored at 24°C using a custom-made automatic bioluminescence apparatus [53,54] and commercially available instruments (CL24A-LIC and CL96S-4, Churitsu Electronic Corporation, Nagoya, Japan). In each cycle, bioluminescence was measured after at least 3.5 min of dark exposure to ensure a decrease in delayed light emission of chlorophyll. The results of bioluminescence monitoring were analyzed using the Rhythm analysis program (RAP [53,54]) and Kaiseki NINJA (Churitsu Electronic Corporation, Nagoya, Japan). Bioluminescence rhythm data were de-trended by dividing by the 24-hour moving average. The Cosinor-rhythmometry method was used to analyze bioluminescence rhythm data over 3–4 days. Data for which the curve fitting was obviously inaccurate or had a high error index (>0.055) were not considered for further analysis. For the analysis of the ROC15-LUC reporter, luminescence detected in empty wells or in cultures of the reporter-less BR strain was considered background noise and subtracted unless specifically mentioned.

Transformation

Transformation of the C. reinhardtii nuclear genome was achieved by electroporation, as described previously [27,55]. For mutagenesis, ROC15-LUC mt⁺ was used as a host strain for transformation with the aph7” fragment. The aph7” fragment for mutagenesis was obtained by digesting the plasmid pHyg3 [29] with the HindIII restriction endonuclease. The 1.7 kb fragment was purified after agarose gel electrophoresis, and 30 ng of DNA (/~3 × 10⁷ cells) was used for the transformation. Transformants were selected on TAP agar plates with a final hygromycin concentration of 30 μg/mL. For complementation, the pLaadA/CETL plasmid was digested with the PacI restriction endonuclease. The resulting 14.3 kb fragment was purified after agarose gel electrophoresis. DNA fragments (300 ng /~3 × 10⁷ cells) were used for transformation, and transformants were selected on TAP agar plates with a final spectinomycin concentration of 50 μg/mL.

Mutant/Complement screening

The transformant colonies were inoculated into 100 μL of TAP medium and maintained in 96 well plates (Nunc MicroWell, Thermo Fisher Scientific) for 3 days at 24°C in LL conditions (30–40 μmol m^-2 s^-1). Following this, 5 μL of the culture was transferred into 100 μL of fresh TAP medium containing D-luciferin (final concentration, 100 μM) in 96 well white plates (Nunc F96 MicroWell, Thermo Fisher Scientific). The plates were maintained at 24°C in LL conditions (30–40 μmol m^-2 s^-1) for 1 day before bioluminescence monitoring. Bioluminescence was monitored using a custom-made automatic bioluminescence apparatus [53, 54].

Protein analysis

Western blot analysis was performed as previously described [27]. Polyacrylamide gel (12.5%) (e-PAGEL, ATTO Corporation, Tokyo, Japan) was used for the observation of light- induced phosphorylation of ROC15 (Figs 5B, S9B and S9C) and 7.5% polyacrylamide gel (e-PAGEL, ATTO Corporation) was used for the detection of HA-CETL (Figs 3B and S7E). The transfer step was performed using a Qblot kit, EZ blot kit and EZFastBlot HMW kit (ATTO Corporation)). BLOCK ACE Powder (KAC, Kyoto, Japan) was used to block the membranes. Rat monoclonal anti-HA antibody was used as the primary antibody (1:10000, clone 3F10, Roche, Basel, Switzerland). Horseradish peroxidase-conjugated goat anti-rat IgG was used as a secondary antibody (1:25000, Merck KGaA, Darmstadt, Germany).

Immunocytochemistry was performed as described previously [27]. Rat monoclonal anti-HA (1:1000) and mouse monoclonal anti-NPC (1:1000, clone MAb414, Labcorp, NC, USA) were used as primary antibodies. Alexa Fluor 488-conjugated goat anti-rat IgG (Thermo Fisher Scientific) and Alexa Fluor 647-conjugated goat anti-mouse IgG (Thermo Fisher Scientific) were used as secondary antibodies. Fluorescence was observed using a laser scanning confocal fluorescence microscope (FV10i-DOC; Olympus, Tokyo, Japan).

RT-qPCR

RT-qPCR was performed as described previously [28]. The primers used for quantification of ROC15, ROC40, ROC66, ROC75, and RCK1 transcripts are the same as previously described [35]. Primers used for quantification of CETL transcript are listed in S1 Table.