Pharmacological targeting of BMAL1 modulates circadian and immune pathways

Abstract

The basic helix-loop-helix PER-ARNT-SIM (bHLH-PAS) proteins BMAL1 and CLOCK heterodimerize to form the master transcription factor governing rhythmic gene expression. Owing to connections between circadian regulation and numerous physiological pathways, targeting the BMAL1-CLOCK complex pharmacologically is an attractive entry point for intervening in circadian-related processes. In this study, we developed a small molecule, Core Circadian Modulator (CCM), that targets the cavity in the PASB domain of BMAL1, causing it to expand, leading to conformational changes in the PASB domain and altering the functions of BMAL1 as a transcription factor. Biochemical, structural and cellular investigations validate the high level of selectivity of CCM in engaging BMAL1, enabling direct access to BMAL1-CLOCK cellular activities. CCM induces dose-dependent alterations in PER2-Luc oscillations and orchestrates the downregulation of inflammatory and phagocytic pathways in macrophages. These findings collectively reveal that the BMAL1 protein architecture is inherently configured to enable the binding of chemical ligands for functional modulation.

Fig. 1. Characterization of CCM binding to human BMAL1(PASB).

a, Hit-to-lead optimization of BMAL1 ligand. The K_d values for SPR were determined using equilibrium analysis. b, CCM enhances the Tm of hBMAL1(PASB) by 9.7 °C in a PTS assay. c, ITC detects CCM binding to hBMAL1(PASB) protein with K_d of 1.99 ± 0.38 μM, stoichiometry (n) value of 1.201 ± 0.028 and enthalpy (ΔH) of −29.67 ± 0.826 (kJ mol⁻¹). d, SPR demonstrates the binding of CCM to hBMAL1(PASB) with affinity of 4.05 μM calculated from kinetic analysis. e, CCM stabilizes WT hBMAL1(PASB) but not Y340F/Y404F double mutant in CETSA (left). Isothermal dose–response curves of the effects of CCM on both WT and double mutant at 48 °C to obtain EC₅₀ values reveal key residues necessary for compound stabilization (right). Four technical replicates were applied. Data are presented as mean values ± s.d. f, Volcano plot displays the thermal profile of the U2OS proteome with CCM treatment (50 μM) compared to DMSO treatment. U2OS cell lysate (n = 3) was heated at a temperature gradient (37–60.9 °C) for 3 min, and then denatured proteins were removed by centrifugation. The combined stabilization of the proteins in the supernatant across the heat gradient was detected by MS. The P value was generated by two-tailed t-test. g, CCM stabilization of HiBiT–hBMAL1(PASB) in the same U2OS cells as in f. The y axis shows its remaining protein levels after heating/denaturation steps, detected via the intensity of the Nano-Glo HiBiT signal. Error bars denote the mean ± s.d., and unpaired two-tailed t-test was used for statistical analyses. M.W., molecular weight; Rn, normalized reporter value. Source data

Fig. 2. Structural basis for CCM binding to human Bmal1(PASB) and induced conformational changes.

a, Crystal structure of BMAL1(PASB) in complex with CCM. b, Electron density maps in the central cavity of BMAL1(PASB), showing three water molecules in the apo structure (left) and the CCM molecule bound in the complex (right). The R versus S isomers of CCM are shown together. c, CCM binding repositions residues in BMAL1(PASB). d, CCM expands the central pocket volume three-fold. The accessible cavity size (when CCM is removed in the complex and waters are removed in the apo structure) was calculated in PyMol with PvVOL. e, Location of PASB domains of BMAL1 and HIF2α in their respective complexes. Structures used here are from PDB IDs 4F3L, 4H10, 4ZPK and 6E3S. The position of CCM in the complex was added manually based on the BMAL1(PAS) complex structure determined in this study.

Fig. 3. CCM modulation of BMAL1–CLOCK target gene expressions.

a, Effect of CCM (100 μM) on target genes compared to BMAL1 knockdown. The RNA levels of BMAL1 target genes were quantified by RT–qPCR in unsynchronized U2OS cells. Six and 12 biological replicates were used for knockdown and CCM treatment, respectively. b, Dose-dependent effects of CCM on BMAL1 target genes in unsynchronized U2OS cells (n = 3, biological replicates). c, The effects of CCM (100 μM) are impaired when BMAL1, CLOCK or NPAS2 is knocked down in unsynchronized U2OS cells (n = 5, biological replicates). d, Validation of BMAL1 and CLOCK siRNAs by confirming reduction in their protein levels in U2OS cells. This experiment was performed independently two times with similar results. e, CCM dose-dependent modulation of real-time circadian rhythm observed in peritoneal macrophages directly isolated from Per2–Luc (PER2::Luc) mice (synchronized, n = 3–6). f, CCM demonstrates no significant cytotoxicity on U2OS cells (n = 3, biological replicates) at concentrations of 100 µM and 200 µM for 48 h. The lysis buffer provided in the CyQUANT LDH Cytotoxicity Assay Kit was used as a positive control. g, Monitoring of the circadian rhythm in synchronized U2OS cells (n = 3, biological replicates) using RT–qPCR, showing differences between DMSO vehicle (black) and 100 μM CCM treatment (red). All error bars (a–c and e–g) denote the mean ± s.d. Statistical tests were performed using unpaired two-tailed t-test. Amp., amplitude; Conc., concentration; NS, not significant; Veh, vehicle; RLU, relative luminescence. Source data

Fig. 4. Effect of CCM treatment on macrophage gene expression signature.

a, Volcano plots display differentially expressed genes resulting from CCM treatment in WT BMDMs in both basal and stimulated states (from treatment with LPS). Female mice were used for all conditions (n = 3, WT; n = 6, KO). Differential expression analysis used DESeq2, and P values were adjusted for multiple comparisons using the Benjamini–Hochberg procedure to control for FDRs. b, Gene Ontology (GO) for the 1,391 genes modulated by CCM in WT-stimulated macrophages. The inserted Venn diagram shows the numbers of genes regulated in the WT-stimulated macrophages, intersecting with the genes regulated in the BMAL1-KO-stimulated macrophages. c, Genes involved in phagocytosis are predominantly downregulated by CCM in macrophages (n = 3, biological replicates). Data are presented as mean values ± s.d. FC, fold change. Source data

Fig. 5. Mechanism of action studies.

a, Co-IP studies between BMAL1 and CLOCK in U2OS cells (n = 3, biological replicates), demonstrating minimal effect of CCM on the stability of their complex. The ratio of BMAL1/CLOCK intensity in each IP sample was calculated and then normalized to DMSO control for each independent experiment. Three independent experiments were integrated for statistical analysis. Error bars denote the mean values ± s.d., and P value was generated by unpaired two-tailed t-test. b, BMAL1 localization in the nucleus of U2OS cells remains unchanged, with no observed alterations in the distribution of cytosolic versus nuclear levels after 3 h of CCM treatment at 100 μM. This experiment was performed independently two times with similar results. c, Overall BMAL1 protein levels in U2OS cells (n = 6, biological replicates), as detected by western blot, also remain unaltered after CCM treatment at 100 μM (left). BMAL1/β-actin intensity ratio was quantified (right). d, The PASB α-helix F and amino acids positioned on this helix, His389 and Gln379, shift positions with CCM binding. e, CCM weakens the association of histone H2A with BMAL1. Protein interactions of BMAL1 were analyzed by RIME. U2OS cells (n = 3) were treated with 100 μM CCM or 0.1% DMSO for 3 h. f, CCM does not change the binding of BMAL1 to the promotor of corresponding genes in U2OS cells (n = 3, biological replicates) detected by ChIP–qPCR. Error bars in a, c and f are presented as mean values ± s.d., and statistical tests were performed using unpaired two-tailed t-test. FC, fold change; RMSD, root mean square deviation. Source data

Extended Data Fig. 1. Developing small molecules for BMAL1.

a, Structure of hit01 (Left). Melting curve of BMAL1(PASB) with DMSO or hit01 (Middle). Shift of BMAL1(PASB) Tm value with DMSO, hit01 or CCM at various concentration (Right). b, Affinity of compounds to BMAL1(PASB) measured by surface plasmon resonance (SPR). c, Affinity fitting of SPR data at equilibrium points for binding of CCM to BMAL1(PASB) protein. Source data

Extended Data Fig. 2. Binding specificity of CCM to BMAL1.

a and b, CETSA melt curves and ITDR of the effect of CCM on mutated hBMAL1(PASB) protein Y340F (a) or Y404F (b). Four technical replicates were applied, and all error bars are presented as mean values +/− SD in a-e. c, Tm shift values of WT or mutated hBMAL1(PASB) in CETSA at 200 μM CCM in (a), (b) and Fig. 1e. d, CCM does not stabilize proteins in cell lysate that are closely related to BMAL1. CETSA melt curves are shown with DMSO or 200 μM CCM. Tm shift values are shown in (e). f, Thermal profile of U2OS proteome under 200 µM CCM treatment. SHROOM1 shows up as the only significantly stabilized protein. Source data

Extended Data Fig. 3. CCM-induced BMAL1 conformation changes.

a, Three water molecules bind in hBMAL1(PASB)-apo structure through hydrogen-bond interaction with surrounding residues. b, Different binding mode for S-CCM and R-CCM in hBMAL1(PASB). c, Electron density maps. Left is the omit map. Center are maps with only S-CCM included in the model. Right are maps with only R-CCM included in the model. In each case, the 2F_o-F_c (blue) were set at 1.0 rmsd, and F_o-F_c maps (red for negative and green for positive) at 3.0 rmsd. d, CCM relocates strands βA and βB, in addition to helix αF. A hydrogen bond forms between His389 and the main chain of Ile346 after CCM bound. e, Conformation changes in the BMAL1(PASB) main chain upon CCM binding.

Extended Data Fig. 4. Selectivity of CCM for BMAL1.

a, Sequence alignments of ARNT, ARNT2, BMAL1 and BMAL2. The sequence identity of PASBs to BMAL1(PASB) 338–441 are 38.5% (ARNT), 38.5% (ARNT2) and 70.2% (BMAL2). b, Structural basis for lack of CCM binding to related PASBs. Protein structures used here are extracted from PDB 2KDK (BMAL2), 4ZP4 (ARNT) and 7XI3 (ARNT2).

Extended Data Fig. 5. Rhythmic modulation of other BMAL1-target genes by CCM.

a, Representative raw traces of the Per2-Luc reporter rhythms. b, The effects of CCM (red) versus vehicle control (black) are shown for RT-qPCR studies conducted using synchronized U2OS cells (n = 3, biological replicates). Data are presented as mean values +/− SD. Source data

Extended Data Fig. 6. Phenotypic evaluation of macrophage after CCM treatment.

Principal component analysis of the wild-type samples suggest that CCM treatment can partially revert the gene expression profile of LPS-treated cells towards the baseline profile of vehicle-treated controls. CCM alone does not significantly alter the gene expression profile compared to the control. Source data

Extended Data Fig. 7. Association of BMAL1 on promotor regions of target genes.

Association of BMAL1 on promotor regions of target genes as detected by ChIP qPCR in synchronized U2OS cells (n = 3, biological replicates) with CCM treatment at 100 μM for 24 h. Data are presented as mean values +/− SD. Statistical tests were performed using unpaired two-tailed t-test. Source data