Abnormal circadian clock gene expression upon loss of 6–4phr function

It has been reported that marsupial CPDphr can functionally substitute for CRY proteins to restore a normal circadian clock in loss-of-function mouse mutants [29]. Is this functionality restricted to this artificial situation or does it reflect a broader natural photolyase function in vertebrates? To address this question, we analyzed CRISPR-Cas9 generated medaka cell lines carrying loss of function mutations in the CPDphr, 6–4phr and DASHphr genes and their corresponding wild type (WT) controls [32]. All cell lines were cultured under light-dark cycle (LD) conditions (12 hours light and 12 hours dark) for several days to directly entrain their circadian clocks. Subsequently, we assayed the levels of various clock gene mRNAs in each line at regular time points to investigate whether there were any abnormalities in rhythmic expression. Specifically, the expression levels of a set of period genes (per1b, per2a and per3) were quantified at various timepoints throughout the light-dark cycle by qRT-PCR analysis. The results revealed low basal levels and arhythmic expression for all three period genes analyzed in the 6–4phr mutant cell line compared with the corresponding WT control cells (Fig 1a-c and Tab A in S1 Data). However, in the other mutant medaka cell lines (CPDphr and DASHphr mutants), rhythmic expression of these genes closely matched that of the wild type control cell line (S1 Fig and Tab F in S1 Data). Might this abnormality be restricted to these cell culture models? To confirm these initial in vitro results, we also measured the mRNA levels of these three per genes in cultured fin clips from the 6–4phr mutant fish as well as the corresponding WT controls (iCab) which were maintained under LD conditions for 4 days to entrain their circadian clocks. Consistent with the in vitro results, we observed low basal levels and arhythmic expression of per gene mRNA in the 6–4phr mutant compared with iCab fin clips (Fig 1d-f and Tab A in S1 Data). This reveals that among the three photolyase genes, loss of 6–4phr function has a greater impact on rhythmic clock gene expression in medaka.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Circadian clock gene expression and regulation in loss of 6-4phr function medaka cells.

(a-f) Medaka 6-4 wild type (WT) and mutant (KO) cell lines (a-c) and explanted wild type (iCab) and 6-4phr mutant (6-4 KO) fish fin clips (d-f) were incubated under light-dark cycle (LD) conditions (12 hours light and 12 hours dark) for 4 days to entrain their circadian clocks and thereafter sampled at regular timepoints through a complete LD cycle. The mRNA expression levels of the clock-regulated genes per1b and per3 and the light-regulated gene per2a were analyzed by qRT-PCR. Relative mRNA expression calculated as mean ± SEM (n = 3) is plotted on the y-axes, while zeitgeber time (zt) sampling timepoints are plotted on the x-axes. (g) In vivo luciferase assay of medaka 6-4 WT (black trace) and KO (red trace) cell lines transfected with a luciferase reporter carrying the zebrafish clock-regulated per1b promoter. Bioluminescence (cps) values are plotted on the y-axis against time (hrs) on the x-axis. Each timepoint represents the mean ± SEM from independently transfected wells (n ≥ 8). White and black horizontal bars along the x-axis indicate the light and dark periods, respectively. (h) In vitro luciferase assay of medaka 6-4 WT (black) and KO (grey) cells transfected with a luciferase reporter construct containing 15 tandemly repeated copies of the D-box sequence derived from the zebrafish cry1aa gene promoter (D-box_cry1aaLuc) and exposed to visible light or maintained in constant darkness. The fold induction of relative bioluminescence values is plotted on the y-axis against exposure conditions on the x-axis. The results are presented as mean ± SEM (n ≥ 3). A ß-galactosidase assay was employed for normalization of transfection efficiency. Significant thresholds are indicated by asterisks and p > 0.05, p < 0.05, p < 0.01, p < 0.001 are represented by ns, *, ** or ***, respectively. Three biological replicates were performed for each experiment.

https://doi.org/10.1371/journal.pgen.1011971.g001

To further investigate the molecular mechanism whereby 6–4phr influences the circadian rhythmicity of per gene expression we explored the possible role of 6–4phr in transcriptional regulation. Rhythmic transcription of per1b and per3 is regulated directly by the circadian clock via E-box enhancer elements while the light-inducible expression of per2a relies on proximal E- and D-box elements [31,33,34]. Medaka 6–4phr mutant and wild type control cells were transiently transfected with a luciferase reporter driven by the zebrafish per1b promoter which includes multiple E-box enhancers [31]. Following transfection, cells were initially exposed to LD conditions in order to entrain their endogenous circadian clocks for 3 days followed by incubation under constant darkness for 2 days (free-running conditions). Consistent with our previous qRT-PCR analysis of endogenous per1b expression, we observed a substantial attenuation of clock-regulated rhythmic gene expression under both entraining and free-running conditions. These findings agree with 6–4phr playing a key role in E-box – mediated transcriptional control mechanisms in the medaka cell lines (Fig 1g and Tab A in S1 Data).

Light-driven transcription of the per2a gene has been demonstrated to be regulated by a D-box enhancer promoter element [35,36]. Therefore, we next tested the effect of loss of 6–4phr function on D-box enhancer-regulated transcription. Specifically, the 6–4phr mutant and WT cells were transfected with a heterologous reporter construct where 15 tandemly repeated copies of a D-box enhancer sequence derived from the cry1aa promoter, were cloned upstream of a minimal promoter which drives transcription of the luciferase reporter gene (D-box_cry1aaLuc) [17]. Transfected cells were exposed to visible light for 15 hours while a control set of cells was maintained in constant darkness before preparing cell extracts and performing a luciferase assay. Our results revealed significant attenuation of light-induced luciferase expression in the 6–4phr mutant cells compared with WT control cells (Fig 1h and Tab A in S1 Data). Therefore, together these results are consistent with 6–4phr playing a regulatory role in both E-box and D-box directed transcription, contributing to both light- and clock-regulated gene expression.

6-4phr differentially modulates E-box and D-box binding transcription factors

As a part of the central TTFL in the molecular circadian clock mechanism, at a critical stage of the circadian cycle, the Per and Cry proteins suppress transactivation by the Clock-Bmal heterodimer. Given the overall structural similarity between photolyases and cryptochromes, we speculated that 6–4phr may potentially function as a negative regulator of Clock-Bmal-mediated transactivation. In contrast, based on our qRT-PCR results (S1a-c Fig and Tab F in S1 Data), although CPDphr shares a similar structure with 6–4phr, this enzyme should not be expected to influence Clock-Bmal function. To test this hypothesis, wild type zebrafish cells (PAC-2 cells) were co-transfected with a luciferase reporter under the transcriptional control of a minimal promoter carrying 4 tandemly repeated copies of a canonical E-box element derived from the zebrafish per1b promoter (4xE-box (-7) [31]) together with expression vectors encoding zebrafish Clock1a and Bmal1a. These constructs were either co-transfected with expression vectors for 6–4phr, CPDphr or a Cry1aa expression vector used as an experimental positive control. While confirming that Cry1aa serves as a potent repressor of the transactivating function of Clock1a-Bmal1a on the E-box reporter, importantly our results also revealed that 6–4phr co-expression led to a dose-dependent suppression of Clock1a-Bmal1a-mediated activation. In contrast, co-expression with CPDphr failed to elicit any significant repression of Clock1a-Bmal1a transactivation (Fig 2a and Tab B in S1 Data). Comparable results were obtained by transfecting the same constructs into the murine 3T3 cell line which lacks endogenous photolyase genes (Fig 2b and Tab B in S1 Data). Therefore, these findings are consistent with 6–4phr, but not CPDphr, sharing a Clock-Bmal inhibitory function with Crys, resulting in down regulation of E-box-mediated transcription.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. 6–4phr differentially modulates E-box and D-box binding transcription factors.

(a-b) In vitro luciferase assay of zebrafish PAC-2 (a) and mammalian 3T3 (b) cells co-transfected with an E-box luciferase reporter construct containing 4 tandemly multimerized copies of the E-box sequence derived from the zebrafish per1b gene promoter (4xE-box (-7)) and expression vectors for zebrafish Clock1a and Bmal1a, 6–4phr, CPDphr and Cry1aa (from 50 ng to 150 ng). (c-d) In vitro luciferase assay of zebrafish PAC-2 cells co-transfected with a D-box reporter construct containing 15 tandemly multimerized copies of the D-box sequence derived from the zebrafish cry1aa gene promoter (D-box_cry1aaLuc) together with expression vectors for the six PAR factors, specifically Tefa, Tefb, Dbpa, Dbpb, Hlfa and Hlfb (c) and six E4bp4/nfil3 factors, specifically E4BP4–1 (nfil3-1a/e4bp4–1), E4bp4–2 (nfil3-2a/e4bp4–2), E4BP4–3 (nfil3-3b/e4bp4–3), E4BP4–4 (nfil3-1b.1/e4bp4–4), E4BP4–5 (nfil3-2b/e4bp4–5) and E4BP4–6 (nfil3-3a/e4bp4–6) (d) (25 ng) as well as zebrafish 6–4phr (30 ng). (e-f, h-j) In vitro luciferase assay of zebrafish PAC-2 (e, h-j) and mammalian 3T3 (f) cells co-transfected with the D-box_cry1aaLuc reporter construct and expression vectors for zebrafish Tefb (25 ng) together with the titration of expression vectors for 6–4phr (e-f), Cry3 (h), Cry1aa (i) or CPDphr (j). (g) In vitro luciferase assay of medaka 6–4 WT and KO cell lines co-transfected with the D-box_cry1aaLuc reporter construct and the expression vector for zebrafish Tefb. The expression plasmids included in each transfection are indicated below each graph. Fold induction of relative bioluminescence values is plotted on the y-axis. The results are indicated as mean ± SEM (n = 3). A ß-galactosidase assay was used to standardize for transfection efficiency. Three biological replicates were performed for each experiment. Significant thresholds are denoted by asterisks and p > 0.05, p < 0.05, p < 0.01, p < 0.001 are represented by ns, *, ** or ***, respectively.

https://doi.org/10.1371/journal.pgen.1011971.g002

We next wished to investigate how 6–4phr influences D-box enhancer-binding transcription factor function. Six PAR bZip transcription factors (Tefa, Tefb, Hlfa, Hlfb, Dbpa and Dbpb) and seven E4bp4/nfil3 family members (nfil3-1a/e4bp4–1, nfil3-2a/e4bp4–2, nfil3-3b/e4bp4–3, nfil3-1b.1/e4bp4–4, nfil3-2b/e4bp4–5, nfil3-3a/e4bp4–6 and nfil3-1b.2) [13,14] are recognized as D-box binding bZip transcription factors in fish. We initially investigated whether ectopic expression of 6–4phr influences transcriptional regulation by twelve of these transcription factors in a co-transfection assay in PAC-2 cells. Specifically, the D-box_cry1aaLuc reporter was co-transfected together with expression vectors for each of the twelve D-box – binding transcription factors in the presence or absence of the 6–4phr expression vector. While co-transfection with expression vectors for all six PAR bZip factors resulted in an increase in bioluminescence derived from the D-box reporter construct, co-expression of 6–4phr together with tefa, tefb and dbpb resulted in higher levels of bioluminescence than with the PAR bZip factors alone (Fig 2c and Tab B in S1 Data). However, in the case of co-expression of the E4bp4/nfil3 factors with the D-box reporter, levels of bioluminescence remained relatively low in the presence or absence of the 6–4phr expression vector (Fig 2d and Tab B in S1 Data).

To validate these initial results, we focused on the inducing effect of 6–4phr on D-box – mediated transcriptional activation by Tefb, the factor displaying the strongest activation in our assay. We co-transfected the D-box_cry1aaLuc reporter with a fixed amount of Tefb and increasing amounts of the 6–4phr expression vectors. We observed a dose-dependent increase in the Tefb, D-box – mediated transactivation (Fig 2e and Tab B in S1 Data). This result was also confirmed in mouse 3T3 cells where we observed a comparable, dose-dependent increase in Tefb-induced activation of D-box reporter transcription in the presence of increasing amounts of co-transfected 6–4phr expression vector (Fig 2f and Tab B in S1 Data).

As an alternative test for the contribution of endogenous 6–4phr to Tefb regulation, we co-transfected the D-box_cry1aaLuc reporter construct and Tefb expression vector into the medaka 6–4phr mutant and WT cell lines. Consistent with the previous results obtained using exogenously expressed 6–4phr, lower levels of Tefb-induced D-box reporter expression were observed in the 6–4phr mutant cells compared with the WT cells (Fig 2g and Tab B in S1 Data).

How specific are the effects of 6–4phr on Tefb regulation? Do Crys or other photolyases also influence activation by this PAR bZip factor? A comparable experiment was performed by co-transfecting PAC-2 cells with the D-box_cry1aaLuc reporter, a fixed amount of Tefb expression construct and increasing levels of Cry3, Cry1aa or CPDphr expression vectors. Based on phylogenetic analysis, fish Cry3 represents a close relative of 6–4phr while Cry1aa bears more similarity with mammalian cryptochromes [37]. Instead, CPDphr belongs to a distinct group of photolyases catalyzing the repair of CPD photoproducts [19]. In all cases, we failed to observe any significant effect on Tefb activation of D-box reporter transcription (Fig 2h and 2j and Tab B in S1 Data).

6-4phr physically interacts with multiple circadian clock transcription factors

Protein-protein interactions are essential for the core timing function of the circadian clock. Therefore, do the effects of 6–4phr on transcriptional regulation by the E-box and D-box result from a direct interaction between 6–4phr and the transcription factors that bind to them? As a first step to test for a direct physical interaction between 6–4phr and the Clock-Bmal heterodimer, we performed a NanoBiT (NanoLuc Binary Technology) - based assay. Expression constructs for the three proteins linked at the N- or C-terminus with the large (LgBiT) or the small (SmBiT) subunits of the Nano-luciferase enzyme were prepared. Expression of the various NanoBiT fusion constructs was initially confirmed by transfection into PAC-2 cells followed by western blotting with an anti-LgBiT or epitope-specific anti-Myc tag antibody (S2a-c Fig and Tab G in S1 Data). As a positive control, Clock1a and Bmal1a NanoBiT expression constructs were co-transfected into PAC-2 cells. We observed a robust induction of furimazine-derived bioluminescence, consistent with the well-documented heterodimerization of these two bHLH-PAS domain transcription factors (S2d Fig and Tab G in S1 Data). Subsequently, 6–4phr and Clock1a or Bmal1a NanoBiT constructs were co-transfected into PAC-2 cells in various pairwise combinations and NanoBiT luciferase-catalyzed bioluminescence was assayed. Our results reveal that co-expression of Clock1a fused with the LgBiT subunit at the N-terminus and 6–4phr linked with SmBiT at the N-terminus resulted in elevated bioluminescence signals (Fig 3a and Tab C in S1 Data). Furthermore, the Bmal1a protein fused with LgBiT at the C-terminus and 6–4phr fused with SmBiT at the N-terminus also showed significantly elevated bioluminescence values (Fig 3b and Tab C in S1 Data). These results are consistent with both Clock1a and Bmal1a proteins physically interacting with 6–4phr.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. 6–4phr physically interacts with circadian clock transcription factors.

(a-b, e) NanoLuc Binary Technology (NanoBiT) assay to test the physical interaction between zebrafish Clock1a (a), Bmal1a (b) or Tefb (e) and 6–4phr. The full-length coding sequences of clock1a, bmal1a, tefb and 6–4phr were cloned into pcDNA 3.1 or the pCS2-MTK vector incorporating a CMV promoter and fused either with LgBiT (at the 5’ end for clock1a and tefb or 3’ end for bmal1a) or with SmBiT (at the 3’ end for 6–4phr). The PRKACA:PRKAR2A pair was used as a constitutive positive control, while the Halo Tag-SmBiT was used as a negative control. Bioluminescence (cps) values are plotted on the y-axes with the mean ± SEM (n ≥ 3). (c-d, f) Yeast 2 hybrid assay (Y2H) to test the physical interaction between zebrafish Clock1a, Bmal1a (c-d) or Tefb (f) with 6–4phr. The full-length coding sequences of clock1a, bmal1a, tefb and 6–4phr were cloned into the pGBKT7 or pGADT7 vectors and the DNA-binding domain of GAL4 (BD) or activation domain of GAL4 (AD) was fused at the N- or C-terminus. The targeted constructs were transformed separately into the yeast strains AH109 or Y187 and the yeast transformants were patched onto agar plates including synthetic medium without leucine, tryptophan and histidine (TDO). A titrated dilution was performed to show the interaction between each pair of proteins. Three biological replicates were performed for each experiment.

https://doi.org/10.1371/journal.pgen.1011971.g003

This result was confirmed with a yeast two hybrid protein-protein interaction assay. Specifically, full-length 6–4phr, clock1a and bmal1a coding sequences were fused to the coding sequences for the GAL4 DNA-binding domain (BD) or the GAL4 activation domain (AD) and then the expression vectors were transformed into yeast cells. For selection of constructs, agar plates including synthetic media devoid of leucine and tryptophan (LW) were employed with various dilutions of each combination (S2e and S2f Fig) and growth on media lacking leucine, tryptophan and histidine (TDO) was tested in order to evaluate the interactions. The various dilutions were used to test the interaction strength between each combination. Consistent with our NanoBiT assay results, the colony growth confirms that Clock1a and Bmal1a display a strong physical interaction, and also that both Clock1a and Bmal1a interact robustly with 6–4phr (Fig 3c and 3d).

Does the induction of Tefb mediated transcription from the D-box enhancer by 6–4phr also rely upon direct interaction between the two factors? To address this question, initially several Tefb NanoBiT fusion constructs were generated and co-transfected in various pairwise combinations into PAC-2 cells. Consistent with the formation of Tefb-Tefb homodimers via dimerization of the bZip domains, we observed significantly elevated bioluminescence signals upon transfection with certain pairwise combinations of the Tefb fusion constructs (S2g Fig and Tab G in S1 Data). Co-transfection of PAC-2 cells with the Tefb and 6–4phr NanoBiT constructs also revealed significantly elevated bioluminescence levels, indicating that 6–4phr has the capacity to physically interact with the Tefb protein (Fig 3e and Tab C in S1 Data). We again validated these results using the yeast two hybrid assay. Expression vectors carrying the Tefb coding sequence fused to the GAL4 DNA-binding domain (BD) or the GAL4 activation domain (AD) were generated and then transformed into AH109 and Y187 yeast cells, respectively, and mated with corresponding 6–4phr expressing yeast strains. The colony growth results were used to test the expression (S2h Fig) and confirm that 6–4phr can physically interact with the Tefb protein (Fig 3f).

Photolyase domain of 6–4phr is responsible for the interaction with Clock1a, Bmal1a and Tefb

The 6–4phr protein possesses two highly conserved domains involved in chromophore binding. A FAD binding domain which binds flavin adenine dinucleotide and a photolyase domain which binds chromophores such as 5,10-Methenyltetrahydrofolate (MTHF). Via these chromophores, primarily FAD, the 6–4phr protein harvests light energy to catalyse photoreactivation repair of UV-induced 6–4 photoproducts [26,27]. Are these domains crucial for the interaction of 6–4phr with circadian clock components? To address this question, we generated several sub-deletions of 6–4phr fused with SmBiT (phr∆-1–7, Fig 4a) and then tested the interaction of these deletion constructs with full-length Clock1a, Bmal1a or Tefb fused to LgBiT. The expression of these sub-deletions was initially validated in transfected zebrafish PAC-2 cells by western blotting using an epitope-specific anti-Myc tag antibody (S3a and S3b Fig and Tab H in S1 Data). Each 6–4phr sub-deletion expression vector was then co-transfected in PAC-2 cells together with the expression constructs for LgBiT-tagged, full-length Clock1a, Bmal1a or Tefb. Cotransfection with an expression vector for SmBiT-tagged, full length 6–4phr served as a positive control. Bioluminescence assay results revealed that the phr∆-1 and phr∆-2 deletions disrupted interaction with the Clock1a and Bmal1a proteins, while only the phr∆-2 deletion interrupted the interaction with the Tefb protein. These findings are consistent with the region of the 6–4phr protein which physically interacts with these circadian clock components (Fig 4b-d and Tab D in S1 Data), overlapping with the photolyase domain.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 4. The photolyase domain of 6-4phr interacts with Clock1a, Bmal1a and Tefb.

(a) Schematic representation of the set of N-terminally, SmBiT-tagged 6-4phr sub-deletion constructs used to test the regions of the 6-4phr protein mediating interaction with full length Clock1a, Bmal1a and Tefb. (b-d) NanoBiT assays testing which domains of 6-4phr mediate the interactions with the Clock1a (b), Bmal1a (c) or Tefb (d) proteins. SmBiT-tagged 6-4phr sub-deletions as well as the LgBiT - tagged full length Clock1a, Bmal1a or Tefb expression constructs were cotransfected into zebrafish PAC-2 cells. The bioluminescence assay results (cps) are plotted on the y-axes as means ± SEM (n ≥ 3). Three biological replicates were performed for each experiment. Significant thresholds are indicated by asterisks and p > 0.05, p < 0.05, p < 0.01, p < 0.001 are represented by ns, *, ** or ***, respectively.

https://doi.org/10.1371/journal.pgen.1011971.g004

Multiple functional domains of Clock1a and Bmal1a are responsible for the interaction with 6–4phr

The arhythmic period gene expression observed in 6–4phr mutant cells and tissues points to interactions between 6–4phr, Clock1a and Bmal1a proteins playing a key role in the core clock TTFL mechanism. We therefore explored in more detail how 6–4phr interacts with Clock1a and Bmal1a. Does 6–4phr repress the transactivation of the Clock1a-Bmal1a heterodimer directly by interfering with their heterodimerization, by influencing their interaction with other transcriptional co-activators or with the RNA polymerase machinery itself [38,39]? To address these questions, we performed a NanoBiT assay involving co-transfection of PAC-2 cells with fixed amounts of the Clock1a and Bmal1a NanoBiT fusion expression constructs together with increasing amounts of an expression vector for the untagged 6–4phr. The results reveal a significant decrease in the interaction between the Clock1a and Bmal1a NanoBiT fusion proteins in the presence of increasing amounts of co-transfected 6–4phr (Fig 5a and Tab E in S1 Data). This data is consistent with 6–4phr repressing Clock1a-Bmal1a-activated transcription by interfering with the formation of Clock1a-Bmal1a heterodimers.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 5. Multiple functional domains of Clock1a and Bmal1a are responsible for the interaction with 6-4phr.

(a) NanoBiT assay to test the effect of 6-4phr on Clock1a-Bmal1a heterodimerization. The LgBiT-tagged zebrafish full length Clock1a and Bmal1a expression constructs were transfected into PAC-2 cells at optimized concentrations together with a titration of expression vector for the untagged 6-4phr. (b-c) Schematic representation of the N terminally, LgBiT-tagged Clock1a (b) and C terminally, LgBiT-tagged Bmal1a (c) deletion constructs used to test for interactions with full length 6-4phr. (d-e) NanoBiT assay to test the domains of Clock1a (d) and Bmal1a (e) responsible for the interaction with 6-4phr. The LgBiT-tagged_Clock1a or_Bmal1a_deletion expression vectors together with the SmBiT-tagged full length 6-4phr expression vector were transfected into zebrafish PAC-2 cells. Bioluminescence (cps) values are plotted on the y-axes as mean values ± SEM (n = 3). (f-g) Schematic representation of Clock1a (f) and Bmal1a (g) deletion mutants fused to the GAL4 activation domain (AD) or GAL4 DNA binding domain (BD) and used to test for interactions with full length 6-4phr fused to the GAL4 DNA binding domain (BD) or GAL4 activation domain (AD). (h-i) Yeast 2 hybrid assay to test the critical domains of Clock1a (h) and Bmal1a (i) responsible for their interaction with 6-4phr. Clock1a and Bmal1a deletions and full length 6-4phr were transformed separately into the yeast strains AH109 or Y187 and the yeast transformants were patched onto agar plates including synthetic medium without leucine, tryptophan and histidine (TDO). A titrated dilution was performed to show the interaction between each combination. Three biological replicates were performed for each experiment. Significant thresholds are denoted by asterisks and p > 0.05, p < 0.05, p < 0.01, p < 0.001 are represented by ns, *, ** or ***, respectively.

https://doi.org/10.1371/journal.pgen.1011971.g005

Which domains of Clock1a and Bmal1a are essential for the physical interaction with 6–4phr? To address this question, multiple sub-deletion constructs based on the Clock1a and Bmal1a NanoBiT constructs were generated in order to systematically eliminate each functional domain of the Clock1a and Bmal1a proteins (Fig 5b and 5c) and their expression was confirmed by western blotting (S3c-g Fig and Tab H in S1 Data). Zebrafish PAC-2 cells were then co-transfected with each of these deletion constructs together with the construct expressing full-length 6–4phr fused with SmBiT at the N-terminus. The NanoBiT results were then validated with a yeast two hybrid assay where various portions of Clock1a and Bmal1a were fused to the GAL4 DNA-binding domain (BD) or the GAL4 activation domain (AD) (Fig 5f and 5g). Then these constructs were transformed into yeast cells for mating with cells transformed with the full-length 6–4phr in order to test for protein-protein interaction (S3h and S3i Fig). Together, these results revealed that the integrity of the bHLH and PAS A domains of Clock1a as well as the bHLH and C-terminal domains of Bmal1a is essential for the physical interaction with the 6–4phr protein (Fig 5d, 5e, 5h and 5i and Tab E in S1 Data).

Fish maintenance and dissection of tissues

Wild type (iCab) and 6–4phr mutant (6–4 KO) medaka (Oryzias latipes) fish were maintained in the fish facility of the Institute of Biological and Chemical Systems, Biological Information Processing (IBCS-BIP) at the Karlsruhe Institute for Technology (KIT). The 6–4phr mutant fish line was generated by CRISPR-Cas9-based mutagenesis at Osaka University, Japan as previously described [32]. The fish were anesthetized using 0.02% Tricaine (Sigma Aldrich) diluted in embryo rearing medium (ERM) for a few minutes and thereafter fin clips were excised using a scalpel and placed in culture. The fish were then transferred to fresh ERM medium for recovery. The experiments performed with fish were conducted in compliance with the European Legislation for the Protection of Animals used for Scientific Purposes (Directive 2010/63/EU) (General license for fish maintenance and breeding: Az.: 35-9185.64/ BH KIT IBCS-BIP, Karlsruhe Institute of Technology (KIT)) and adhered to the animal protection standards of Germany.

Explanted fin clip cultures, cell culture and transfection

iCab and 6–4 KO fish fin clips were excised using a scalpel, washed three times in 1 X PBS and Leibovitz’s L15 medium (Gibco) and then cultured in Leibovitz’s L15 medium supplemented with Fetal Bovine Serum (FBS) (Gibco) and appropriate antibiotics (Gibco). A cell line derived from zebrafish embryos (PAC-2) as well as the wild type and mutant medaka cell lines generated from medaka embryos (6–4/CPD/DASH WT/KO) were maintained in Leibovitz’s L-15 medium´ supplemented with Fetal Bovine Serum (FBS) and appropriate antibiotics. All these cell lines were cultured at 26°C and transfected according to the manufacturer’s instructions using FuGene HD transfection reagent (Promega) with a ratio of FuGene (μl) to plasmid DNA (μg) of 4:1.

Gene expression analysis

Medaka wild type and photolyase mutant cells as well as fin clips were plated in petri dishes and maintained under LD cycle (12 hours light and 12 hours dark) conditions for 4 days. Afterwards, Trizol Reagent (Invitrogen) was added to the fin clips and cell layers then the lysates were transferred to 1.5 ml Eppendorf tubes and frozen at -80°C. Fin clips as well as cell samples were disrupted and homogenized by syringing repeatedly through a 25 Guage needle. Chloroform was added to separate RNA from the DNA and protein mixture and isopropanol was employed for RNA precipitation by following the manufacturer’s instructions. The integrity and quantity of the ribosomal 28S, 18S and 5S RNA bands was visualized by electrophoresis on a 1% agarose gel.

1 μg of total RNA was used for synthesis of cDNA and subsequently digested with RQ1 DNase co-incubated with a RNase Inhibitor. RQ1 DNase stop solution and a random primer were added to the mixture, and then a master mix (RT Buffer + dNTPs + RevertAid RT Enzyme) was prepared and added to the cDNA mixture and incubated according to the manufacturer’s instructions (Thermo Fisher). Electrophoresis through a 1% agarose gel was used to verify the quality of quantity of synthesized cDNA.

Quantitative Real-time PCR (qRT-PCR) was performed using the StepOnePlus Real-Time PCR System (Thermo Fisher) according to the manufacturer’s instructions. Medaka mitochondrial ribosomal protein S18B (MRPS18B) mRNA expression was used for normalization to calculate the relative mRNA expression levels of clock genes by the 2^-ΔΔCT method. A standard curve was plotted to test the primer efficiency for amplifying the per gene transcripts. The primer sequences used and test results are listed in S1 Table.

Plasmid cloning

Several plasmids (per1b-Luc [31], D-box_cry1aaLuc [17], 4xE-box(-7) [31], and pCS2-MTK expression constructs for zebrafish clock1a, bmal1a, 6–4phr, CPDphr [47], cry1aa as well as tefa, tefb, dbpa, dbpb, hlfa, hlfb, nfil3-1a/e4bp4–1, nfil3-2a/e4bp4–2, nfil3-3b/e4bp4–3, nfil3-1b.1/e4bp4–4, nfil3-2b/e4bp4–5 and nfil3-3a/e4bp4–6 [13]) were maintained using standard methods in the Foulkes lab’s plasmid collection.

The full-length NanoBiT constructs containing clock1a, bmal1a, 6–4phr and tefb coding sequences fused to coding sequences for the LgBiT or SmBiT domains at the N- or C-termini were generated by a PCR-based cloning strategy. Specifically, the inserts were PCR amplified using Q5 High-Fidelity DNA polymerase (New England Biolabs) and ligated with the pCS2-MTK or pcDNA3.1 vectors by T4 ligase (New England Biolabs). All of the NanoBiT Clock1a, Bmal1a and 6–4phr subdeletion expression plasmids were generated by using the Q5 Site-Directed Mutagenesis Kit (New England Biolabs) according to the manufacturer’s instructions. The primers used for construction are listed in S2 Table.

All the yeast two-hybrid construct cloning inserts including clock1a and its subdeletions, bmal1a and its subdeletions, 6–4phr and tefb as well as the pGBKT7, pGADT7 and pcDNA3.1 cloning vectors were PCR amplified using Q5 High-Fidelity DNA polymerase (New England Biolabs). Plasmid fragments were assembled by employing the NEBuilder HiFi DNA Assembly Cloning Kit (New England Biolabs). All the insert fragments incorporated 25 bp of overlapping sequences relative to the neighboring vector fragment. Standard transformation protocols for Escherichia coli were used. The identity of all freshly generated and re-transformed plasmids was confirmed by appropriate restriction enzyme (New England Biolabs) digestion of mini preparation (QIAGEN) samples and sequencing of midi or maxi preparation (QIAGEN) samples by a commercial supplier (Microsynth Seq Lab). A NanoDrop Spectrophotometer (PeqLab) was used to assess the concentration and purity of all plasmid preparations. All the primers used for yeast two hybrid assay plasmid construction and mutagenesis are presented in S3 Table.

In vitro luciferase assay

Cells were seeded in a 24-well plate and the next day were transfected with various plasmids of interest by using FuGene HD transfection reagent (Promega). A ß-galactosidase expression vector (pcDNA3.1/Myc-His/lacZ, Invitrogen) was included in each transfected sample to control for transfection efficiency and the final DNA quantity was adjusted by the addition of varying amounts of empty pCS2-MTK or pcDNA 3.1 vectors to ensure equal amounts of total DNA were included in each transfection. Cells were lysed by addition of firefly luciferase lysis buffer (0.1 M Tris acetate (pH = 7.5), 2 mM EDTA, 1% Triton X 100) and then collected and stored at -80°C until used for in vitro luciferase and ß-galactosidase assays.

In vitro luciferase assays were performed using the Luciferase Assay System kit (Promega) according to the manufacturer’s instructions and bioluminescence levels were assayed using a Victor Light Luminescence Plate Reader (Perkin Elmer). The ß-galactosidase assay was performed using a standard colorimetric assay and optical densities were measured using a SpectraMax iD3 Microplate Reader at a wavelength of 420 nm.

In vivo luciferase assay

Cells were seeded in a white 96-well plate (Perkin Elmer) and transfected with different plasmids of interest using FuGene HD transfection reagent (Promega) according to the manufacturer’s instructions. Thereafter, a beetle luciferin potassium salt solution (200 mM) (Promega) was added to the culture medium in each well (luciferin:medium = 1:250) and the plate was sealed with plastic sealing foil. The plates were then transferred to a Topcount NXT automatic scintillation counter (Perkin Elmer) under different lighting conditions (light/dark cycle or constant darkness). The data was analyzed by the I-and-A Excel plug-in (S. Kay, Scripps Research Institute).

Western blotting analysis

Cells were seeded in a 6-well plate and the following day, plasmids of interest were transfected by using FuGene HD transfection reagent (Promega) according to the manufacturer’s instructions. After incubation, the culture medium was aspirated and then the cells lysed by the addition of 1x Laemmli buffer (2% SDS, 10% glycerol, 0.0025% bromophenol blue, 0.1 M DTT and 63 mM Tris, pH = 6.8). A Bio-Rad mini-protean 3 chamber apparatus was employed for SDS-PAGE and then protein samples were transferred to an ImmunBlot PVDF membrane (Millipore) by electroblotting. After primary and secondary antibody incubation, the Clarity Western ECL substrate (Bio-Rad) was added to the PVDF membrane and the result was visualized by using the ChemiDoc Imaging System (Bio-Rad). Images were viewed and evaluated using ImageLab software (Bio-Rad). The primary antibody and secondary antibodies used are listed in S4 Table.

NanoBiT assay

The NanoLuc Binary Technology (NanoBiT) assay was performed by using the NanoBiT PPI MCS Starter System (Promega), which detects protein-protein interactions in living cells based on structural complementation between Large BiT (LgBiT; 18 kDa) and Small BiT (SmBiT; 11 amino acid peptide) subunits. clock1a, bmal1a and 6–4phr cDNA sequences with their sub-deletion fragments, and tefb were fused with LgBiT or SmBiT at the N- or the C-terminus. To generate the various NanoBiT fusion plasmids, the coding sequences of all the proteins of interest were PCR amplified and then cloned into CMV-based vectors (pCS2-MTK or pcDNA3.1). All pairwise combinations were transiently transfected into zebrafish PAC-2 cells using FuGene HD transfection reagent (Promega) and the cells were incubated in darkness for 2 days. After addition of Nano-Glo Live Cell Reagent, a non-lytic detection reagent including the cell-permeable furimazine substrate, the bioluminescence signals were detected by a Topcount NXT automatic scintillation counter (Perkin Elmer).

Yeast two-hybrid assay

The yeast two-hybrid assay was performed according to the user manual of the Matchmaker Gold Yeast Two-Hybrid System (Clontech Laboratories, Inc.). All the clock1a and bmal1a derived fragments, 6–4phr and tefb were amplified by PCR and ligated with either the pGBKT7 or pGADT7 vectors. For all the constructs, the DNA-binding domain of GAL4 (BD) or activation domain of GAL4 (AD) was fused at the N- or C-terminus of targeted genes or their mutants. The BD and AD-based vectors were subsequently transformed into the yeast strains AH109 and Y187 which were mated for protein-protein interaction detection.

Statistical analysis

All data were calculated and represented as mean ± SEM. Additionally, they were plotted by using GraphPad Prism 9 (GraphPad Software Inc.) and analyzed by SPSS Statistics 19.0 (IBM). To evaluate statistically significant differences, either a Student’s t-test or an analysis of variance (ANOVA) was employed, and then Sidak’s multiple comparison post-test was applied. Statistics were defined to be significant when the p value was less than 0.05 and the statistical differences of p < 0.05, p < 0.01, p < 0.001 are indicated by *, ** and ***, respectively.