The E3 ubiquitin ligase UBE3A is an integral component of the molecular circadian clock through regulating the BMAL1 transcription factor

PER2::Luc mice were from Prof. J Takahashi (22). All animal work was carried out in accordance with the 1986 Home Office Animal Procedures Act (UK) and following local ethical review.

The following antibodies were used in this study, BMAL1 (mouse monoclonal, 23), ubiquitin (Abcam, rabbit polyclonal), REV-ERBα (kind gift from Dr J. Ripperger), UBE3A (Cell Signalling, rabbit monoclonal), total GSK3β and p-GSK3β (Cell Signalling, rabbit polyclonal), Myc and tubulin (Sigma, monoclonal), Human influenza hemagglutinin (HA) (Santa Cruz, mouse monoclonal), Histidine tag (HIS) (Santa Cruz, mouse monoclonal), V5 (Invitrogen, mouse monoclonal), p53 (Merck Millipore, mouse monoclonal) and green fluorescent protein (GFP) (Santa Cruz, rabbit polyclonal). MG132, dexamethasone, forskolin and cycloheximide were purchased from Sigma.

mCLOCK and hBMAL1 expression plasmids were kind gifts from Dr Kazuhiro Yagita (Kyoto Prefectural University of Medicine, Japan). mAVP::luc plasmid and luc assay were described previously (23,24). Full-length complementary deoxyribonucleic acid (cDNA) encoding V5-BMAL1 was a kind gift from Dr Pat Nolan (MRC Harwell, UK), pCW7-HIS-ubiquitin were described previously(25). HA-UBE3A (Addgene 8648), HA-UBE3A-C833A (Addgene 8649), Myc-UBE3A (Addgene 11468) and E6/E7 (Addgene 8641).

MEFs were isolated from E13.5 wild type (WT) PER2::Luc mice as described (26). MEFs were transfected with E6/E7 construct as described before (27) and stable transfectants confirmed by genotyping and reverse transcriptase-polymerase chain reaction (RT-PCR) of E6/E7 fragments from genomic DNA and cDNA. NIH3T3 cells were purchased from the American Type Culture Collection. Cells were grown in complete Dulbecco's modified Eagle's medium (DMEM; Invitrogen) supplemented with 10% fetal bovine serum (FBS) in a 37°C incubator maintained at 5% CO₂.

Agarose was prepared in 6-well plates, with a base layer consisting of 1% agar in 1 x DMEM supplemented with 10% FBS and a top layer consisting of 0.35% agar in 1 x DMEM with 10% FBS. Cells were seeded directly into the top agar layer at a density of 5 × 10³ cells per well and allowed to grow for approximately 3 weeks prior to imaging under phase contrast.

Immunocytochemistry was performed as described (24). Briefly, cells on sterile coverslips were fixed with 4% paraformaldehyde, then blocked in 1% donkey serum before incubation with primary antibody. Cells were then treated with an Alexa-Fluor-conjugated secondary antibody mix. Fluorescence was visualized using a Zeiss Observer D1 AX10-inverted microscope, photographed using an AxioCam ICM1 camera and AxioVision version 4.8.2 software (Carl Zeiss).

Circadian rhythms in bioluminescence from Per2::Luc in SW1353 cells or PER2::Luc in MEFs were recorded in real time using photomultiplier tube (PMT) devices (28). Confluent cells were synchronized with 50% horse serum for 2 h, Dex (100 nM), Forskolin (FSK) (10 μM) or temperature cycles (36–38.5°C) as indicated, then transferred to standard recording media as described (29). For shRNA knockdown in NIH3T3 cells with pGL3-mBmal1::Luc-Hyg (30), 35-mm dishes of cells were transiently transfected by shRNA vectors. The cells were treated with 100-nM Dex for 2 h and then transferred to standard recording media. Bioluminescence was recorded for 1 min at 9-min intervals for 5–10 days at 37ºC in air with Dish Type Luminescencer, Kronos (ATTO). Raw data of the rhythms were detrended by 24-h moving averages.

Whole cell extracts were prepared in lysis buffer containing 1% Triton X-100 as previously described (25). Plasmid transfections for HEK293T or NIH3T3 cells were performed using Polyfect (Qiagen) according to the manufacturer's instructions. To analyse protein stability, the protein synthesis inhibitor cycloheximide (20 μg/ml) was added to cells for the times indicated in the figure legends. Cells were then washed and samples analysed over a 6-h period by western blot.

Ubiquitination of BMAL1 was examined by histidine (His) purification method (25). HEK 293T cells were transfected with V5-BMAL1, His-ubiquitin, E6/E7, wild-type or mutant derivatives of UBE3A plasmids in varied combination as detailed in figure legend. After transfection, the cells were left overnight and then treated with MG132 for 6 h, followed by solubilizing the cells with lysis buffer A containing 8-M guanidinium HCl, 100-mM Na2HPO4/NaH2PO4,10-mM Tris-HCl (pH8.0) and 10-mM imidazole. Subsequently, Ni-nitrilotriacetic acid purification of the His-ubiquitin conjugates was performed under denaturing conditions.

Co-IP of over-expressed protein was performed as described before (25). Briefly, HEK 293T cells in 10-cm dishes were transfected with various combinations of tagged plasmids. Overnight after transfection, the cells were solubilized with lysis buffer (150-mM NaCl, 20-mM Tris-HCl, pH7.5, 10% glycerol, 1% Triton X-100, 1-mM phenylmethylsulfonyl fluoride, PhosSTOP and cOmplete ethylenediaminetetraacetic acid-free cocktail protease inhibitors, Roche). After centrifugation, 1-ml supernatants were used for immunoprecipitation by adding 2 μg of V5 or HA antibody as appropriate at 4°C overnight. The following day, Protein G Dynabeads (Invitrogen) were added and incubated for 2 h, followed by four washes with lysis buffer in the presence of protease inhibitors. Precipitated proteins were examined by sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) and western blot.

Ribonucleic acid (RNA) was extracted from cultured fibroblasts using the Qiagen RNeasy purification system. cDNA was prepared using the Superscript II reverse transcriptase (Invitrogen) and analysed for gene expression using quantitative real-time PCR with TaqMan (Applied Biosystems) chemistry (24). Probeset was ordered (pre-validated) from Applied Biosystems (Bmal1 Mm00500226_m1).

For knockdown of Ube3a, shRNA was designed using siDirect (http://sidirect2.rnai.jp/↗), a web-based software, and the following sequences were used: shUbe3a#1—5′-GGUUU ACUAU GCAAA UGUAG U-3′—and shUbe3a#2—5′-GGAUU AUAUU AUGAC AAUAG A-3′. The oligonucleotides to express the shRNA were inserted into the pBS-mU6 vector, as described previously (30). Small interfering RNA (siRNA) transfections were performed with RNAiMax reverse transfection reagent (Invitrogen). Cells were incubated with 10- or 20-nM SMARTpool siRNA duplexes against human Ube3a, or a scrambled duplex (Dharmacon) for 24 h before downstream analysis. For PMT studies, cells were synchronized with 50% horse serum for 2 h before PMT recording. For Western Blotting (WB) analysis of knockdown efficiency, transfected cells were harvested for protein extraction 48 h post transfection.

All Drosophila stocks were raised on standard cornmeal food at 25°C. RNAi line VDRC-45876 was from Vienna Drosophila RNA interference (RNAi) Center. The UAS-ube3a transgenic line was constructed by amplifying full-length Dube3a from a cDNA pool derived from whole adult flies and cloned into pUAST. Following sequence confirmation the UAS-Ube3a was injected into embryos for germline transformation. For specific knockdown of Dube3a, w;UAS-Gal4;pdf-Gal4 flies were crossed with w;+;UAS-Ube3a-RNAi to generate w;UAS-Gal4/+;pdf-Gal4/UAS-Ube3a-RNAi progeny. For controls each line was, instead, crossed to w;+;+ to generate w;UAS-Gal4/+;pdf-Gal4/+ and w;+;UAS-Ube3a-RNAi/+. Specific over-expression of DUBE3A was achieved by crossing w;UAS-Gal4;pdf-Gal4 and w;+;UAS-Dube3a flies to generate w;UAS-Gal4/+;pdf-Gal4/UAS-Dube3a. Controls (w;UAS-Gal4/+;pdf-Gal4/+ and w;+;UAS-Dube3a/+) were generated by crossing each line to w;+;+.

For the analysis of protein levels, larval brains were dissected and homogenized in ice-cold lysis buffer [50-mM Tris-HCl (pH 7.4), 150-mM NaCl, 1% NP-40, 0.1% SDS and 1% proteinase inhibitor (Merck)], and subjected to SDS-PAGE according to standard procedures. Western blots were probed with anti-UBE3A (8F7, 1:1000, made by Yong^’s lab). The specificity of the antibody was confirmed in larval brains from the Dube3a³⁵ null line.

Circadian locomotor activity of flies was measured as described (31). Briefly, male flies, 1–3 days old, were placed in glass tubes and entrained to a 12-h light–12-h dark cycle (LD) for 5 days and then released into constant dark conditions (DD) for 7–10 days. Locomotor activity was recorded in 30-min bins using DAM2 Drosophila Activity Monitor system (Trikinetics, Waltham). Only flies that survived for 12 days were included in analysis. Data were analysed using El Temps software (Antoni Díez Noguera, University of Barcelona). Periodogram analysis was done using chi-square analysis of LD or DD data, as standard, and categorized as arrhythmic if their rhythm strength (variance%) score was below the rhythmic threshold.

Data were evaluated using Student's t-test or one-way ANOVA with Tukey test as indicated. Results are presented as mean ± SEM from at least three independent experiments. Differences were considered significant at the values of *P < 0.05, **P < 0.01 and ***P < 0.001.

In vitro, over-expression of HPV16 oncogenes E6 and E7 is sufficient to immortalize and transform primary cells (27). In a recent attempt to immortalize MEFs isolated from PER2::Luc clock reporter mice (22), we used E6/E7 oncogenes. Transfection of MEFs with E6/E7 led to successful integration of viral DNA into the host genome (Supplementary Figure S1). While mock transfected MEFs no longer proliferated beyond passage 9, E6/E7 expression led to indefinite proliferation (beyond passage 30; Figure 1A). This was associated with a reduction in p53 protein levels, consistent with E6-mediated degradation as previously reported (Supplementary Figure S2). E6/E7-expressing cells were capable of anchorage-independent growth and colony formation in soft agar, a hallmark of tumorigenic cell lines (Supplementary Figure S3). Thus, stable integration and expression of E6/E7 led to immortalization and transformation of primary MEFs. Surprisingly, using real-time recording, the circadian rhythms in PER2::Luc bioluminescence were severely dampened in transformed MEFs (Figure 1B). This was the case in cells synchronized using various well-accepted protocols (dexamethasone, horse serum, forskolin and 12:12-h temperature cycles; Figure 1B and Supplementary Figure S4). Aberrant rhythms in transformed cells were confirmed by studying BMAL1 protein and messenger RNA (mRNA) expression following clock synchronization (Figure 1C and D).

BMAL1 availability has been proposed as ‘rate-limiting’ for the positive arm of the clock and is necessary for robust rhythmicity (32). In unsynchronized cells average BMAL1 levels were significantly lower (P < 0.01) in transformed cells than in primary cells. This was not accompanied by a reduction in steady-state Bmal1 mRNA (Figure 2A, P > 0.05). In addition, the levels of REV-ERBα, a key repressor of Bmal1, did not significantly change (Supplementary Figure S5A). These results suggest a post-transcriptional mechanism for BMAL1 protein reduction. Indeed, BMAL1 decay rate was faster in transformed cells than in primary cells (Figure 2B) and BMAL1 reduction in transformed cells could be rescued by application of the proteasomal inhibitor MG132 (Supplementary Figure S5B), suggesting the involvement of the proteasomal degradation pathway.

BMAL1 is targeted for degradation by ubiquitination (9,33,34), thus we sought to address whether E6/E7 expression affected this. A Ni-NTA His-ubiquitin pull-down assay was used to precipitate ubiquitinated proteins in HEK 293T cells. Immunoblotting revealed that E6/E7 expression enhanced polyubiquitination of V5-BMAL1 (Figure 2C). The final step of ubiquitination requires a substrate-specific E3 ubiquitin protein ligase. Due to the well-documented interaction between E6 and UBE3A (11), we predicted that UBE3A may be involved in E6/E7-mediated degradation of BMAL1. In the presence of E6/E7, over-expression of WT UBE3A, but not a dominant negative mutant (C833A) that lacks ubiquitin ligase activity, caused a reduction in the steady-state level of V5-BMAL1 in HEK 293T cells (Figure 2D). These results suggest that the E6/E7-mediated BMAL1 reduction involves the ubiquitin ligase activity of UBE3A.

To confirm that BMAL1 and UBE3A are capable of interacting with each other in cells, we used Co-IP. In the presence of E6/E7, V5-BMAL1 was pulled down specifically by HA-UBE3A and, reciprocally, myc-UBE3A was pulled down by V5-BMAL1 (Figure 3A and B). Substantially more myc-UBE3A was pulled down when the proteasome was inhibited with MG132, supporting the notion that this interaction leads to proteasomal degradation of BMAL1 (Figure 3C). Importantly, V5-BMAL1 could pull down myc-UBE3A in the presence and absence of E6/E7 (Figure 3C). This is distinct from the E6-dependent interaction between UBE3A and p53 (35), and suggests that UBE3A acts as an endogenous ubiquitin ligase of BMAL1. To prove a physiological interaction between endogenous UBE3A and BMAL1, an endogenous IP was performed in unsynchronized SW1353 cells. Endogenous BMAL1 was pulled down specifically by UBE3A using an anti-UBE3A antibody (Figure 3D). It is known that GSK3β phosphorylates BMAL1, promoting its ubiquitination and degradation (34). However, we found no difference in the level of either total or phosphorylated GSK3β (Ser9) between primary and transformed cells (Supplementary Figure S6).

Because BMAL1 is essential for rhythmic physiology and UBE3A interacted with BMAL1 under physiological conditions independently of E6/E7, we hypothesized that UBE3A-mediated BMAL1 degradation could be important for endogenous circadian timekeeping. To test this, we used shRNAs to knockdown Ube3a in NIH3T3 mouse fibroblasts containing a Bmal1::Luc reporter. Knockdown of Ube3a led to a significant reduction in the amplitude of Bmal1::Luc oscillations (P < 0.01 or 0.001; Figure 4A and C and Supplementary Figure S7A). Furthermore, knockdown of human Ube3a in a Per2::Luc SW1353 human cell line (24) led to unstable clock oscillations, with an initial period lengthening (by up to 4.5 h; P < 0.05) and amplitude reduction (∼50%; P < 0.05 for 10 nM and P < 0.01 for 20-nM siRNA), followed by a loss of (or irregular) rhythmicity (Figure 4B and C). Consistent with a role of UBE3A in regulating BMAL1, we observed increased BMAL1 levels following RNAi treatments against Ube3a (Supplementary Figure S7B). Ubiquitination of BMAL1 has been shown to be essential for its transactivational activity (33). To establish a role of endogenous UBE3A in regulating BMAL1 function, we performed transactivation assay in HEK 293 cells using a mAVP::luc promoter reporter (23). The ability of BMAL1/CLOCK to activate mAVP::luc was significantly reduced by knockdown of Ube3a (Supplementary Figure S8), supporting UBE3A as an endogenous regulator of the core clock complex. Taken together, our results demonstrate that loss-of-function of UBE3A in cells dampens circadian amplitude and lengthens clock period, indicating UBE3A as an endogenous regulator of circadian pacemaking in mammalian cells, at least partly through regulating the core clock transcription factor BMAL1.

To determine the importance of UBE3A in driving rhythmic behaviour, we used Drosophila melanogaster, which shares high homology with mammals in both the Ube3a gene and in molecular clock mechanisms (2,36). We first confirmed effective reduction/increase of average dUBE3A protein levels in all neurons by driving dUbe3a RNAi (VDRC-45876) or dUbe3a-WT (i.e. over-expression; OE) pan-neuronally (Figure 5A). Next, we used the pdf>Gal4 driver to express dUbe3a RNAi and WT exclusively in the LN_v neurons that coordinate locomotor behaviour (Figure 5B). Both the pdf>dUbe3a-RNAi and pdf>dUBE3A-OE lines exhibited normal behaviour under a 12:12-h light–dark (LD) cycle (Figure 5C and D). The pdf>dUbe3a-RNAi line gradually lost rhythmicity in DD, with 11/14 flies becoming arrhythmic in DD (Figure 5C and D and Table 1). Surprisingly, the pdf>dUBE3A-OE exhibited a similar and more pronounced arrhythmic phenotype, with 100% of flies showing an immediate loss of rhythmicity in DD (Figure 5C and D and Table 1). Taken together, our results indicate that tight control of dUBE3A levels in central clock neurons of flies is critical for the regulation of locomotion rhythms.