Identification of the Repressive Domain of the Negative Circadian Clock Component CHRONO

Polymerase chain reaction (PCR) was used to make constructs described in this study. Optimized sequence coding for CHR-H and CHR-H3were amplified and inserted into the V28-E4 plasmid (from Dr. Ge lab, Anhui University) between NotⅠ and XhoⅠ restriction sites, resulting in a maltose binding protein (MBP) tag and 6XHis tag fused to the 5′ and 3′ ends respectively. MBP-CHRH (or H3) was expressed in the E. coil BL21 (DE3) strain. For the luciferase reporter assay, various versions of chrono as indicated in the text were cloned into pcDNA3.1(+)/Hygro (Invitrogen, Untied States). For immunoprecipitation, sequence of 5Myc-6His-hBMAL1 (5M6HA1) was cloned and inserted into pcDNA3.1 via BmaHⅠ and EcoRⅠ sites (named pcDNA3.1-5M6HA1 afterwards). On the one hand, a series of chrono with cpVenus tag followed sequences were cloned and inserted into pcDNA3.1(+)/Hygro plasmid between HindⅢ and XhoⅠ sites, on the other hand, we built series of bmal1 with 5XMyc-6XHis-tag, using two EcoRⅠ sites on the mutated pcDNA3.1.

To make L606A/L607A double mutations in the bmal1 coding region, the whole plasmid pcDNA3.1-5M6HA1 was amplified using a high fidelity Taq polymerase. A pair of complementary primers 5′GGCAGCAATGGCTGTCATCATGAGCGCAGCAGAAGCAGATGCTGGACTGGGTGGC3′ and 5′GCCACCCAGTCCAGCATCTGCTTCTGCTGCGCTCATGATGACAGCCATTGCTGCCT3′ were used for the mutagenesis. After the amplification, DpnⅠwas used to digest the template plasmids. Then the product was transformed into the DH5α competent cells. To obtain the site-directed mutagenesis plasmids, the plasmids from the transformed E.coli colonies were sent to sequencing for identification.

PredictProtein (www.predictprotein.org/↗), PSIPRED (http://bioinf.cs.ucl.ac.uk/psipred/↗), and JPred (http://www.compbio.dundee.ac.uk/jpred/↗) were used for CHR secondary structure prediction and both offered similar results.

NucPred (https://nucpred.bioinfo.se/nucpred/↗), PredictNLS (https://rostlab.org/owiki/index.php/PredictNLS↗), and PSORTII (https://psort.hgc.jp/form2.html↗), were used for CHR nuclear location signal prediction.

Codon usage bias is a general characteristic for all species. The human chronocDNA had a low yield using prokaryotic expression system. We used OPTIMIZER (http://genomes.urv.cat/OPTIMIZER/↗) and JCat (http://www.jcat.de↗) to obtain optimized sequences for chrono to express in E. coli, and the full optimized sequence was ordered from General Biosystems, Inc. Anhui Province, China.

BL21(DE3) E. coli cells that carry constructs for expression of MBP-CHRH/CHRH3 fusion protein were grown at 37 °C, 220 rpm to reach OD values up to 0.6–0.8. Then 0.2 mM IPTG was added to induce expression at 16 °C, 150 rpm, for 20–24 h. Cell pellets were lysed in 50 mM Tris-HCl (pH 8.0), and 200 mM NaCl. Ni-NTA Sefinose™ Resin (#C600033, BBI, Shanghai, China) was used for first His tag affinity purification. Then, Hiload 16/600 superdex (200pg) was used as gel filtration purification.

To verify rich α-helix occupation in the secondary structure of CHR-H, we compared secondary structures of MBP-His₆, MBP-CHRH-His₆, and MBP-CHRH3-His₆ with circular dichroism spectroscopy (Biologic/MOS-500). Protein buffer was changed to 50 mM sodium phosphate (pH 8.0), and then the protein sample was diluted to 0.1 mg/mL and clarified using by centrifugation (12,000 rpm, 4 °C, 10 min) before measurement. Each CD spectrum was collected at room temperature, with a scan rate of 1 nm/s; scanning increment, 0.5 nm; spectral bandwidth, 1.0 nm; from 190–260 nm for each measurement. The spectra represent the average of 3 scans with the bassline subtracted from analogous conditions.

HEK-293T cells were grown in DMEM (SH30022, Hyclone/GE, United States) supplemented with 10% fetal bovine serum (P30-3302, PAN-Biotech, Germany) plus 100 IU/mL penicillin and 0.1 mg/mL streptomycin (SV30010, Hyclone/GE, United States) at 37 °C, 5% CO₂. Approximately 3 × 10⁵ cells were seeded in each well of a 6-well plate 1 day before transfection. When cells reached 70%~90% confluency, 100 ng of the firefly luciferase reporter pP_PK2::Fluc, 40 ng of the renilla luciferase reporterpP_CMV::Rluc, 250 ng of a hamster Bmal1 construct, and 250 ng of a mouse Clock construct and 500 ng of mouse Per2 (or human CHR, CHR-N, CHRNH, CHR-H, CHRH2, CHRH3, CHRH4, CHRHC, and CHR-C) were co-transfected into HEK 293T by using Lipo6000™ Transfection Reagent (C0526, Beyotime, Beijing, China). The constant amount of DNA (1100 ng/well) was adjusted with pcDNA3.1(+)/Hygro vector as a carrier, and we measured single luminescence. Twenty-four hours after transfection, cells were washed three times with ice-cold PBS (SH30256.01, Hyclone/GE, United States), then processed with a TransDetect Double-Luciferase Reporter Assay Kit (FR201, Transgen Biotech, Beijing, China), and measured by SpectraMax Paradigm Multi-Mode Microplate Reader (Molecular Devices, Silicon Valley, United States). Each construct was examined at least for three independent times.

Mouse antibodies against Myc-tag (M185-3L/M047-3, MBL, Nagoya, Japan), rabbit antibody against GFP-tag (D110008, BBI, Shanghai, China) and sheep antibody against Tubulin (ATN02, Cytoskeleton, Denver, United States) were subjected to Western blot according to the manufacturer’s protocol. For IP, HEK-293T cells transfected with the desired plasmids by using Attractene Transfection Reagent (301005, QIAGEN, Hilden, Germany), according to the fast-forward protocol, were lysed in IP buffer (10 mM Tris pH 7.5, 150 mM NaCl, 1 mM EDTA, 1% Triton X-100, 0.1% sodium deoxycholate, 1 mM PMSF and Roche complete EDTA free protease inhibitor cocktail). After quantification by using BCA Protein Assay Kit (P0012, Beyotime, Beijing, China), lysates were precleared with Protein G Agarose (P2009, Beyotime, Beijing, China) and then immunoprecipitated with mouse anti-Myc antibodies. After washing five times, the precipitates were resuspended in the 5XSDS loading dye, boiled for 5 min, and run on a 12.5% SDS-PAGE gel followed by Western blot analysis. Immunoreactive bands were detected by Typhoon laser scanners (FLA-9500, GE, United States).

For IF (immunofluorescence), Hela cells were grown in DMEM supplemented with 10% FBS plus 100 IU/mL penicillin and 0.1mg/mL streptomycin at 37 °C, 5% CO₂. A sterile cover glass was placed into each well of a 6-well plate, and then 2 × 10⁵ cells per well were seeded 1 day before transfection. When cells reached 70%~90% confluency, they were transfected with Lipo6000™ Transfection Reagent with various combinations of the following constructs (all in pcDNA3.1 (+)/Hygro background): 1,250 ngCHR-cpVenus or CHRH3-cpVenus and 1,250 ng 5XMyc-6XHis-hBMAL1, 5XMyc-6XHis-hBMAL1-C, or 5XMyc-6XHis-hBMAL1-Cs. Twenty-four hours after transfection, cells were washed three times with ice-cold PBS and fixed in 4% paraformaldehyde for 30 min at room temperature. Then cells were permeabilized and blocked with 0.3% TritonX-100 and 5%FBS in PBS for 1h at room temperature follow by sequential washing with PBS, and were subjected for incubation using Myc-tag antibody (ab32, UK, abcam, diluted in PBS containing 3%BSA) for 1h at room temperature. By following three 5 min PBS washes, cells were immune-stained by Cy3-conjugated antibody (D110172, BBI, Shanghai, China, diluted in PBS containing 1%BSA/0.3%TritonX-100) for 30 min at room temperature. After last three times PBS wash, cells were embedded in Mounting Medium, Antifading (with DAPI, S2110, Solarbio). Analyses and documentation were done with immunofluorescence microscopy (TCS SP5, Leica, Germany). For localization assessment, Hela and NIH 3T3 cells were cultured in the same media, and a sterile cover glass was placed into each well of a 6-well plate, and then 2 × 10⁵ cells per well were seeded 1 day before transfection. When cells reached 70%~90% confluency, they were transfected with Lipo6000™ Transfection Reagent with series of cpVenus fused truncations of CHR (all in pcDNA3.1 (+)/Hygro background): 2500ngCHR-cpVenus/cpVenus-CHR/CHRN-cpVenus/CHRH-cpVenus/CHRC-cpVenus or EGFP. Twenty-four hours after transfection, cells were washed three times with ice-cold PBS and fixed in 4% paraformaldehyde for 30 min at room temperature. After three times PBS washes, cells were embedded in Mounting Medium, Antifading and documented by immunofluorescence microscopy.

chrono^−/−bmal1^Luccell line was generated using the CRISPR/Cas 9 system. Oligonucleotides specific for the target sites of chrono gene loci were designed using the Optimized CRISPR Design tool (www.genomeengineering.org/↗). The highest quality score sgRNA (CCGCTGCAGGCATCGATCGA) targeted to the exon 1 of chrono (on-target locus: chr1:+150255862) was chosen to insert into the expression vector pX459 (pSpCas9{BB}-2A-Puro was a gift from Feng Zhang, Addgene plasmid # 48139). 24 h after transfection, cells were screened with 2 μg/mL puromycin. Cell line clones were screened by the limiting dilution method.

2 × 10⁵ U2-OS reporter cells were seeded in 35mm dishes the day before synchronization. On the next day, cells in each dish were synchronized with 2 h treatment of 200nM dexamethasone (dissolved in DMSO) and recorded in a LumiCycle as described previously [26]. Bioluminescence data were analyzed with the LumiCycle analysis program (Actimetrics, USA) to obtain circadian parameters such as period and amplitude.

Since there were neither homologous proteins nor similar protein structures being reported, assessment of CHR amino acid sequences was used to predict its secondary structure. Predict Protein [27] and PSIPRED [28] were used to predict that CHR contains an α-helix rich region in the middle (Figure 1A) and both ends of CHR are flexible regions. Thus, CHR can be separated into three domains: a N-terminal domain (CHRN, 1–110), a helical domain (CHRH, 111–196), and a C-terminal domain (CHRC, 197–385).

CHR was identified as a circadian clock component that functions as a transcriptional repressor of B/C complexes [13,14]. However, details of the binding regions of CHR to B/C complexes, in other words the functional domains of CHR, are unknown. Then, we searched the area of CHR that functions as an inhibitor to the transcription activator B/C complexes using the luciferase reporter assay. Based on the prediction, luciferase reporter assays were used to test domain effects including CHRN, CHRH and CHRC with respect to transcription activation by the B/C complexes in HEK293 cells. In this mammalian reporter system, both mPeriod-2 (mPer2) and full-length CHR as positive controls repressed the B/C complexes using a reporter pP_PK2::Fluc (pR2.8) that contains E-box (CACGTG) elements in the promoter region and a reporter pP_CMV::Rluc (R-luc) as the internal reference [29]. However, CHRN, CHRH, or CHRC domain alone was unable to repress transcription activation by the B/C complexes (Figure 1B), which indicated the repressive domain was larger than any individual part. Therefore, we performed the same assay by combing individual parts (CHRNH and CHRHC represent the combinations of CHRN, CHRH, and CHRC) and found CHRHC can repress the transcription activity as strongly as the PER2 protein (Figure 1C). Furthermore, a series of the extended constructs, CHRH2 (111~262), CHRH3 (111~264), and CHRH4 (111~298), were made on the basis of the CHRH and CHRHC and tested by the luciferase reporter assay. As the data indicated, CHRH3 and CHRH4 prevented the activation by the B/C complexes, while CHRH2 was not able to repress the activation (Figure 1D). Thus, the functional domain of CHR protein that represses B/C complexes lies in the middle of CHR is predicted to be an α-helix rich region.

To better understand the repression domain, we went on to explore the structure of CHR protein. We first attempted to acquire its tertiary structures through protein crystallization. We tried different tags (GST, 6XHis, or MBP) and series of truncations to purify CHR proteins. Eventually, we chose MBP as a tag to purify truncated CHR proteins. However, we were not successful to get information of CHR’s tertiary structure through crystallization by now. Based on our luciferase reporter assays, we decided to validate the predicted secondary structure by CD spectroscopy. MBP tagged CHRH and CHRH3 were chosen to carry out the CD experiments. The yield of MBP-CHRH was ~ 40 mg per liter culture, while that of MBP-CHRH3 was merely 2 mg per liter which may be due to the increase of flexible region (Figure 2A,B). According to the in silico prediction, CHRH is α-helix rich, and the component of CHRH3′s elongated part is mainly flexible (Figure 2C). Based on the data from circular dichroism spectra (Figure 2D), we analyzed the secondary structure content via K2D, a neural network method [30]. According to the analysis (Figure 2E), proportion of alpha helices increased significantly when MBP fused with CHRH and CHRH3. These analyses are consistent with our secondary structure prediction that CHRH is an α-helix rich region.

In light of luciferase reporter assays, we sought to confirm the interaction between CHR and BMAL1 using co-immunoprecipitation (co-IP). HEK-293T cells were co-transfected with BMAL1 fused to a 5Myc-6His tag (5M6HA1, short for 5Myc-6His-BMAL1) on the N-terminus and CHR fused to cpVenus on the C-terminus (CHRV, short for CHR-cpVenus). The C-terminus fusion with a cpVenus tag did not interfere with its cellular localization. In consistent with previous reports that CHR interacts with BMAL1 [13,14,15], CHRV could be co-immunoprecipitated from HEK-293T cells lysates together with BMAL1 proteins using an anti-Myc antibody (Figure 3A). Subsequently, we tested the possible direct interactions between BMAL1 and a series of CHR constructs as shown in Figure 1. In line with previous luciferase reporter assays, we foundthat CHRN, CHRH, or CHRC alone fused to cpVenus (CHRNV, CHRHV, or CHRCV) was not able to interact with 5M6HA1 (Figure 3B). Again, in agreement with the reporter assays, CHRHC-cpVenus (CHRHCV) was co-immunoprecipitated with BMAL1using anti-Myc antibodies, indicating that the CHRHC region is necessary for BMAL1/CHR interaction (Figure 3C). Moreover, in the same system, shorter CHRHC truncated variants (CHRH3 and CHRH4) could also be immunoprecipitated by anti-Myc antibodies from extracts of HEK-293T cells co-transfected with 5M6HA1 (Figure 3D,E), while CHRH2, merely two amino acids shorter than CHRH3, could not (Figure 3F). These results signify that the middle region of CHR (111~264) is the minimum area for BMAL1/CHR interaction. In summary, the middle region of CHR that contains a highly helical domain is vital for CHR’sfunction as a transcriptional repressor.

Previously, a region (514~594) of BMAL1 adjacent to the CRY1 interacting terminus was reported to interact with the full-length CHR (14). This interaction disrupts the binding between B/C complexes and CRE binding protein (CBP) or p300, the transcriptional co-activator of circadian clock pathways [14]. However, more convincing evidence from immunoprecipitation or other reliable interaction assays is still lacking. To better understand how CHR interacts with BMAL1 to repress the activity of B/C complexes, we decided to confirm the interaction between CHR and BMAL1C (C-terminus of BMAL1). However, in our data, 5Myc-6His-hBMAL1-C (5M6HA1C, 514~594) failed to co-immunoprecipitate with the full-length CHR protein (Figure 4A). We also carried out immunofluorescence (IF) assays to test their co-localizations (Figure 4H). As a transcriptional repressor, exogenously expressed CHRV is located in the nuclei, as detected by the fused fluorescent protein. Our IF assay using anti-Myc antibodies indicated that BMAL1-C is located in the cytoplasm. The merged image implied that CHRV and 5M6HA1-Cwere not co-localized together (Figure 4H), suggesting that they do not interact with each other. This result support our co-IP results (Figure 4A). Under such circumstances, we believed that there is no sufficient evidence that the right domain of BMAL1 C-terminus has been identified for BMAL1/CHR interaction. Since we knew that the last 43 amino acids of BMAL1 are required for transcriptional activation [23] and CBP also binds to it [25], we hypothesized that CHR disturbed CBP’s co-activation by binding to the last 43 aa of BMAL1(583~626) as well. To test our hypothesis, firstly, we tested whether CHR interacted with the C-terminus of BMAL1. Indeed, 5Myc-6His-hBMAL1-N (5M6HA1-N, 1~490) was not capable of co-immunoprecipitating with CHRV together from extracts of HEK-293T cells (Figure 4B), while the 5Myc-6His-hBMAL1-C longer version (5M6HA1Cl, 490~626) succeeded (Figure 4C). Then, we used 5Myc-6His-hBMAL1-C shorter version (5M6HA1Cs, 579~626), containing the last 43 aa, to pull CHRV down, and it also succeeded in precipitating the CHRV protein (Figure 4D). CRY1 and CBP has been shown to compete for BMAL1 TAD domain binding [31,32], and the IxxLL motif in the TAD α-helix is the major binding site mediating the interaction with CRY1and CBP [25]. Thus, we constructed L606A and L607A double mutant 5M6HA1 (mut5M6HA1), where the IxxLL motif was destroyed [25], to examine whether CHR competes with CBP by binding to TAD domain directly. We found the interaction between BAML1 and CHR was abolished when the IxxLL motif was disrupted (Figure 4E). In other words, CHR may have a competition with CBP for BMAL1 TAD binding. Since CHRH3 retains the repression function of CHR (Figure 1D), we used 5M6HA1Cs to pull CHRH3V down. Although CHRH3V can be precipitated, the CHRH3/BMAL1Cs interaction is weaker compared with the interaction between CHR and BMAL1Cs (Figure 4D,F). It is possible that the functional domain may need extra parts to stabilize the interaction. Therefore, we tested the interaction between 5M6HA1Cs and CHRH4V, which is able to repress the transcriptional activity evoked by B/C complexes and be immunoprecipitated by BMAL1 as well (Figure 1D and Figure 3E). The CHRH4 showed a stronger binding with the C-terminus of BMAL1, compared to CHRH3 (Figure 4F,G). In summary, we suggested that CHRH3 is the minimal domain to exercise as a repressor via directly binding to BAML1-Cs, and the coils adjacent to its C-terminus might be an additional stabilizing part to maintain the binding. Similarly, the co-localization between CHR and the C-terminus of BMAL1 via IF supports the idea that they interact with each (Figure 4H). Consistently, the short construct CHRH3 showed less co-localization, compared to the full-length CHR (Figure 4H, bottom VS. middle). Thus, the repressing domain, CHRH3, is important for interactions between CHR and B/C complexes through the TAD domain at BMAL1′s C-terminus.

As one of the core components of mammalian molecular circadian clock, BMAL1′s C-terminal TAD domain determines circadian period [25]. Since CHR interacts with BMAL1′s TAD domain, we proposed that CHR also contributes to generating the 24 h period. Previous studies have demonstrated that knocking-out chrono in animals could lengthen the circadian period [13,14,15], which agrees with our proposal. Nevertheless, it is unclear whether KO of chrono in cultured cells would interfere with the circadian period. Using the CRISPR/Cas9 system [33], the highest score of a 20-nt guide sequence targeting exon 1 of chrono gene was selected as the sgRNA, which is described in the M&M. The chrono gene was knocked out in U2-OS cells which harbor a luciferase reporter under the control of bmal1 promoter, generating chrono^−/−bmal1^Luc U2-OS cells. We PCR-amplified the target region from chrono^−/−bmal1^Luc genomic DNA, followed by sequencing. The results showed that both chrono alleles were modified, with one allele losing 4 bp and the other allele having 1 bp insertion at the Cas9 cleavage site (Figure 5A). These frameshift mutations created premature translation stop codons, which eventually knocked out chrono in U2-OS cells (Figure 5B). Compared to the wild-type U2-OS cells (WT, 22.18 ± 0.068 h), chrono^−/−bmal1^Luc cells (24.47 ± 0.453 h) exhibited lengthened circadian period by ~2 h (Figure 5C,D). Our cell-based results were consistent with previous animal studies. Overall, these data suggested that CHR plays a role in maintaining the 24 h periodicity via interacting with the C-terminal TAD domain of BMAL1.

We have identified the domain of CHR that interacts with BMAL1 to repress the transcription activation of B/C complexes. This domain contains the predicted helix domain and a partial extended coiled region in its C-terminus. We further wonder what the N-terminus of CHR functions. Thus, we constructed two plasmids with cpVenus conjugated at CHR’s C-terminus (CHRV) and the N-terminus (VCHR) respectively to study the cellular localization of CHR. At first, we checked the location of those proteins in the U2-OS cell line.Since CHR is a transcriptional factor located in the nuclei [13], we predicted these two fused proteins should enter nuclei, too. However, VCHR but not CHRV tended to stay at cytoplasm (Figure 6A). On account of this phenomenon, we hypothesized that a nuclear localization signal (NLS) may lie in the N-terminus of CHR and cpVenus as a rather huge tag (26KD) which can interfere with the NLS function and block the nuclear import. However, although CHR is a nuclear protein, there is no specific sequence predicted as an NLS. Perhaps there is a novel NLS in the N-terminus, so we sought to test whether CHRNV located in the nuclei. We checked the expression of cpVenus fused truncation constructs via immunofluorescence microscopy. In Hela cell line, we found CHRNV was inclined to locate in the nuclei as much as CHRV, while CHRHV and CHRCV spread evenly in whole cell (Figure 6B). Furthermore, we used NIH 3T3 cells to check distributions of these truncations. Compared with distributions of CHRHV and CHRCV, CHRNV has s strong accumulation in the nuclei (Figure 6C). Noteworthily, the expression of CHRNV was obviously higher than CHRHV and CHRCV in both cell lines, consistent with the western blot results (Figure 3B). Perhaps, the nuclear-entrance protected CHRN from being degraded. Taken together (Figure 6B,C), we suggested that N-terminus of CHR leads itself to enter the nuclei, though we have not identified the exact NLS signal yet.