The Circadian Protein BMAL1 Regulates Translation in Response to S6K1-Mediated Phosphorylation

Abstract

The circadian timing system synchronizes cellular function by coordinating rhythmic transcription via a transcription-translational feedback loop. How the circadian system regulates gene expression at the translational level remains a mystery. Here, we show that the key circadian transcription factor BMAL1 associates with the translational machinery in the cytosol and promotes protein synthesis. The mTOR-effector kinase, ribosomal S6 protein kinase 1 (S6K1), an important regulator of translation, rhythmically phosphorylates BMAL1 at an evolutionarily conserved site. S6K1-mediated phosphorylation is critical for BMAL1 to both associate with the translational machinery and stimulate protein synthesis. Protein synthesis rates demonstrate circadian oscillations dependent on BMAL1. Thus, in addition to its critical role in circadian transcription, BMAL1 is a translation factor that links circadian timing and the mTOR signaling pathway. More broadly, these results expand the role of the circadian clock to the regulation of protein synthesis.

Figure 1. A Screen for BMAL1 Cytosolic Interactions Nominates Translation

(A) Schematic of experimental plan for cellular fractionation, immunoprecipitation, and tandem mass spectrometry (LC-MS/MS) to isolate proteins that putatively interact with BMAL1 in the cytosol. (B) Representative immunoblots of WT MEF lysates separated into nuclear and cytosolic fractions probed with indicated antibodies. (C) BMAL1 Cytosolic Interaction Network demonstrates a single major cluster. (D) The largest and most connected cluster in the network is the “Translation” Cluster, comprised of 89 proteins including over 50 ribosomal proteins and 15 translation factors. See also Table S1 and Figure S1. (E) List of a subset of translation factors identified by LC-MS/MS found in the “Translation” Cluster. (F) The list of annotated putative co-precipitating proteins was loaded into the DAVID algorithm. The only significantly nominated pathway from KEGG analysis was “Ribosome.”

Figure 2. BMAL1 Associates with Translation Initiation Factors and Stimulates Translation in Cells

(A) Representative western blot of unsynchronized WT MEFs. Cells were fractionated and immunoprecipitations performed with anti-BMAL1 monoclonal antibody or mouse IgG. (B) BMAL1 associates with translation factors in vivo. Cytoplasmic extract from mouse liver or brain were immunoprecipitated with anti-BMAL1 monoclonal antibody or mouse IgG. Western blots were performed with indicated antibodies. (C) Cytoplasmic MEF lysates were prepared as in (A) and incubated with m⁷-GTP-agarose beads. Co-eluting proteins were resolved by SDS-PAGE, and probed as indicated. m⁷-GTP (1 mM) was used as a cap competitor. (D) BMAL1 m⁷-GTP pull-down is specific. Bmal1^+/+ or Bmal1^−/− MEF lysates were prepared as in (A). (E) BMAL1 associates with CBCs in vivo. Cytoplasmic extracts of mouse liver (as in B) were normalized to total protein and incubated with m⁷-GTP-agarose beads. Western blots were performed with indicated antibodies. (F) An extract of Bmal1^+/+:GFP MEFs was subjected to sucrose density gradient centrifugation. The absorbance profile at 260 nm is shown. The gradient was fractionated; fractions were analyzed by western blotting. Note that BMAL1 and eIF4E were found in the same ribosomal fractions. Ribosomal protein (rpS6) identifies the small ribosomal subunit. (G) Primary MEFs (passage 2–4) of indicated genotype were pulsed with 1 μM puromycin for 30 min, and de novo protein synthesis was quantified by detection of mean ratio of puromycin to total protein (Coomassie) ± SEM normalized to WT (n = 7, **p = 0.002, Student's t test). Independent cell lines from at least two mice per genotype were used. The same cell lines were pulsed with ³⁵S-Met/Cys for 30 min followed by lysis and SDS-PAGE. De novo protein synthesis was determined as the mean ratio of ³⁵S incorporation relative to total protein (Coomassie) ± SEM (n = 3, *p = 0.04, Student's t test). (H) BMAL1 regulates protein synthesis in the liver. Puromycin was injected 30 min before indicated time points into WT or Bmal1^−/− mice kept on a 12 hr/12 hr light/dark cycle. Livers were harvested and whole cell lysates prepared. Protein synthesis was calculated as the ratio of puromycin/total protein, and the mean ± SEM is shown for the indicated number of mice (p = 0.01 for interaction, ***p < 0.0001 for genotype, *p < 0.02 for time, two-way ANOVA, Tukey post test).

Figure 3. BMAL1 Can Stimulate Translation Independent of Its Role in Transcription

(A) Recombinant full-length human GST-tagged BMAL1 stimulates translation in vitro in rabbit reticulocyte lysates. Synthetic m⁷-GTP capped Xef1 mRNA was added to each reaction with increasing amounts of GST-BMAL1 incubated with ³⁵S-Met/Cys. Lysates were resolved by SDS-PAGE, autoradiography performed, and western blots were probed as indicated. Quantification of de novo protein synthesis was performed by calculating the mean volume of ³⁵S compared to control ± SEM (**p = 0.005, *p < 0.05, one-way ANOVA with Tukey post test, n = 3 replicates). (B) Cartoon of CMV-FLAG-BMAL1 plasmids. CMV-FLAG-BMAL1(ΔbHLH) yields a predicted protein product ~15 kDa smaller than WT. (C) Cartoon of the pYIC bicistronic translational reporter plasmid (Nie and Htun, 2006). His, 6xHis tag. (D) Representative western blot of HEK293T cells transfected with pYIC and indicated amounts of FLAG-BMAL1 or FLAG-BMAL1(ΔbHLH) and/or empty vector (pcDNA3.1). Total DNA, 1.8 μg/condition. (E) Histograms of data from (D). Cap-dependent YFP signal was normalized to actin by densitometry. Mean intensity ± SEM are shown normalized to no BMAL1 plasmid (*p < 0.05; **p < 0.01, amount of BMAL1 plasmid; WT versus ΔbHLH, not significant; two-way ANOVA, n = 4–6 per condition). (F) BMAL1(ΔbHLH) did not significantly activate the BMAL1 transcriptional target Rev-Erbα. HEK293T cells were transfected as in (D) with the addition of the BMAL1 co-activator myc-CLOCK. Data are represented as the mean ± SEM of Rev-Erbα RNA normalized to control (***p < 0.0001, one-way ANOVA, Tukey post test; BMAL1(ΔbHLH), not significant). (G) BMAL1 overexpression has no effect on pYIC transcription in HEK293T cells. Data are represented as the mean of YFP(pYIC) RNA ± SEM (one-way ANOVA not significant, Tukey post test, n = 3 replicates).

Figure 4. BMAL1 Is Phosphorylated by S6K1 In Vitro and In Vivo

(A) Recombinant full-length GST-BMAL1 was incubated with His-S6K1 and ³²P-γATP. Reactions were resolved by SDS-PAGE, and autoradiography was performed followed by western blotting. Arrows indicate GST monomers (solid) or GST dimers (dashed). (B) Outline of workflow for LC-MS/MS experiments to identify sites of S6K1-mediated BMAL1 phosphorylation. Dark blue rectangles approximate the relative representation of BMAL1 peptides in the LC-MS/MS data. Below, alignment of the putative S6K1 phosphorylation site at S42 (in human BMAL1) demonstrates evolutionary conservation across phyla. (C) S42 is an S6K1 phosphorylation site in vitro. HEK293T cells were transfected with either FLAG-BMAL1 or FLAG-BMAL1(S42G). FLAG immunoprecipitations were performed and washed. Recombinant S6K1 was added in the presence of ³²P-γATP. Autoradiography was performed followed by western blotting. (D) BMAL1 is phosphorylated at S42 in vivo. Western blot of lysates from immortalized Bmal1 cell lines as indicated. Overexpressed BMAL1 is detected at higher level and is larger in size because of FLAG-HA tag. (E) S6Ks phosphorylate BMAL1 in cells. Whole cell lysates from WT or S6K1/2 double knockout (DKO) cells were probed with indicated antibodies. (F) BMAL1 is phosphorylated at S42 downstream of mTORC1. WT MEFs were infected with lentivirus expressing a scrambled shRNA (control) or sh-raptor. (G and H) Hippocampal lysates from SynapsinI-Cre;Tsc1-flox/flox (KO) or SynapsinI-Cre;Tsc1^+/+ (control) mice were probed with indicated antibodies. Histogram of quantified western blots from (G). Data are represented as the mean ± SEM relative to WT (n = 3 mice per genotype, t test, *p < 0.05, **p < 0.005).

Figure 5. BMAL1 Phosphorylation at S42 Mediates Its Interaction with Translational Machinery and Promotes Cap-Dependent Translation

(A) HEK293T cells were transfected with FLAG-BMAL1 or FLAG-BMAL1(S42G). 24 hr later, cells were lysed and immunoprecipitated with FLAG antibody. Western blots were probed with indicated antibodies. (B) HEK293T cells were transfected with 10 μg or 20 μg of FLAG-BMAL1 or FLAG-BMAL(S42G) for 24 hr. Post-nuclear cell fractions were treated with RNase A followed by m⁷-GTP pull-down assays. Western blots were probed as indicated. (C) HEK293T cells were transfected with pYIC and FLAG-BMAL1 or FLAG-BMAL1(S42G) as in Figure 3D. Cell lysates were probed with indicated antibodies. See also Figures S2, S3, and S4. (D) Quantification of data from (C). His-YFP (cap-dependent) was normalized to actin and the mean ± SEM is represented normalized to control. **p < 0.01, BMAL1 plasmid for WT and non-significant (ns) for S42G. **p < 0.01, WT versus S42G (two-way ANOVA) n = 4– replicates per condition. (E) Western blots of hippocampal lysates of SynapsinI-Cre;Tsc1^+/+ (Tsc1-Control) or SynapsinI-Cre;Tsc1-flox/flox (Tsc1-KO) treated with vehicle or rapamycin. Data are represented as the mean ± SEM relative to WT (n = 2 mice per genotype per condition, one-way ANOVA, Tukey post test, *p < 0.05). (F) m⁷-GTP pull-down assays were performed on pooled cytoplasmic lysates from (E) and immunoblots of eluted proteins were performed with indicated antibodies.

Figure 6. BMAL1 Rhythmically Associates with S6K1 and the Translational Machinery

(A) WT MEFs were synchronized with dexamethasone followed by subcellular fractionation over a sucrose cushion at indicated time points. Representative western blot is shown (n = 4 biological replicates; see Figure S5 for nuclear lysates). (B and C) BMAL1 immunoprecipitation of normalized cytosolic lysates were treated with RNase and subjected to immunoprecipitation with BMAL1 antibody (B) or pull-down assays with m⁷-GTP beads (C). Western blots were performed with indicated antibodies. Asterisk corresponds to the IgG-heavy chain and small arrow to IgG-light chain, respectively (see Figure S5). (D) BMAL1 phosphorylation rises during the active period in liver. Mice were maintained on a 12 hr/12 hr light/dark schedule with ad libitum availability to food prior to release into constant darkness. Tissues were collected at indicated time points. Liver cytoplasmic extracts were collected, normalized to total protein, and immunoblots were performed. Representative western blot from n = 3 mice per time point is shown. (E) Extracts prepared from (D) were normalized to total protein content, and subjected to m⁷-GTP pull-down assays followed by immunoblotting as indicated.

Figure 7. Circadian Rhythms of Protein Synthesis Rates Are Partially BMAL1 and Phospho-BMAL1(S42) Dependent

(A–D) Bmal1^+/+:GFP MEFs were synchronized with dexamethasone. At indicated time points, cells were pulsed with ³⁵S-Met/Cys every 4 hr from hours 16–60 after synchronization followed by cell lysis, SDS-PAGE, autoradiography, and western blotting (below). Representative autoradiograms of the same exposure were pseudocolored for visual clarity (see Figures S7A–S7D for originals), n = 3 replicates per cell line. See also Figures S6 and S7. (E–H) Quantification of (A)–(D). ³⁵S autoradiograms were quantified by densitometry, normalized to Coomassie-stained gels for total protein and normalized to Bmal1^+/+:GFP cells at ZT16. (I) Schematic model of BMAL1 phosphorylation and association with the translational machinery in the cytosol during the circadian cycle.