PBAF/cBAF reorganization on H3.3 chromatin regulates BMAL1 activity in the absence of circadian negative feedback

We therefore investigated whether the assembly of PBAF-BMAL1 is dynamic on H3.3 variant nucleosomes. To address this, we performed native HA-tag IP on soluble chromatin from FH-H3.3A livers collected at circadian times (every 4 hr over 24 hr in constant darkness, CT0-CT20), using wild-type non-tagged liver chromatin as a negative control. Since the FH-H3.3A animals used in the study are heterozygous, H3.3 nucleosomes could, in principle, retain either a canonical H3 variant or the untagged native H3.3 copy. The absence of canonical histone H3.1/2 in the H3.3 complex across all circadian times indicated that the IP enriches for homotypic H3.3 nucleosomes (Fig. 1b). We found that PBAF remodeling complexes (PBRM1, ARID2 and the shared BRG1 subunit) assemble on H3.3 chromatin independent of circadian time. However, the presence of BMAL1 on H3.3 chromatin was enriched during the daytime, peaking at CT8.

To address whether endogenous PBAF complexes associate with BMAL1, we performed PBRM1 IP from native soluble chromatin from wild-type livers, collected at circadian times CT0-CT20. PBRM1 co-immunoprecipitated with other specific PBAF components, including ARID2, BRD7, and PHF10, as well as the shared ATPase subunit BRG1, across all CTs. However, PBAF-BMAL1 complexes were specifically detected during the active phase at CT4 and CT8 (Fig. 1c), consistent with observations on H3.3A complexes. A control IP experiment using an isotype-matched IgG control showed no specific enrichment of either BMAL1 or PBAF complex proteins at any time point, confirming the specificity of the observed interactions (Fig. 1c).

To determine if H3.3 presence on liver chromatin follows a 24-hour circadian rhythm, we performed native HA-tag Chromatin-Immunoprecipitation Sequencing (ChIP-Seq) on mononucleosome-enriched chromatin from FH-H3.3A livers isolated over circadian time (Supplementary Data 2 and 3). We found that H3.3A deposition exhibits rhythmic presence at CLOCK-BMAL1 target genes and binding sites, with levels steadily increasing from CT20 and peaking at CT4 (Fig. 1d, and Supplementary Fig. 2a). At clock-output genes, well-positioned H3.3A nucleosomes were observed downstream of the transcription start sites (TSS) extending up to nearly 1 kb at all times (Fig. 1d, left panel). In contrast, TSS-upstream nucleosomes were primarily detected during the active phase. Circadian E-box sequences were flanked by cycling H3.3A nucleosomes (Fig. 1d, middle panel), while no such rhythms were detected at short interspersed nuclear element (SINE) repeat regions (Fig. 1d, right panel) or at intergenic regions (Supplementary Fig. 3a). The time of enrichment of H3.3 at clock genes preceded the peak recruitment of CLOCK-BMAL1 and active RNA Polymerase II^2,4,9. As a control, we performed H3.1/2 ChIP-Seq at CT8 and CT20. Unlike canonical H3 nucleosomes, H3.3A nucleosomes were specifically enriched at TSSs (Fig. 1e) including clock-controlled genes (Supplementary Fig. 3b).

H3.3 is encoded primarily by 2 genes, H3f3a and H3f3b, in both human and mouse genomes. In mice, additional subsidiary H3.3 variants exhibiting tissue-specific expression are also present²⁹. As both H3f3a and H3f3b encode identical H3.3 proteins, we generated H3.3A and H3.3B knockout mouse embryonic fibroblasts (MEFs) by crossing homozygous floxed strains. In order to determine if H3.3 is important for clock function, double-floxed MEFs were stably transduced with a Bmal1:Luciferase reporter and a Tamoxifen-inducible Cre recombinase. The addition of Tamoxifen to growing fibroblast cultures resulted in a significant reduction of H3.3 protein levels within 4 days (Fig. 1f, Source Data file). We then recorded bioluminescence from the confluent cultures of fibroblasts starting on the fourth day of tamoxifen treatment, for 7 consecutive days. H3.3 depletion caused a significant shortening of the circadian period by nearly 2 hours compared to H3.3 wild-type (floxed) controls (Fig. 1f), indicating a critical role for the H3.3 variant in establishing core oscillator function.

H3.3A was highly enriched at CLOCK-BMAL1 target genes (Fig. 2a, and right panel, Fig. 2d). However, it is known that not all CLOCK-BMAL1 target genes are expressed in the same phase³⁰. To further examine this, we extracted H3.3A occupancy at transcription start sites generating rhythmic transcripts that were either in-phase or out-of-phase with BMAL1 binding, as well as arrhythmic transcripts³⁰. While overall enrichment of H3.3A-labeled nucleosomes was noted in all three transcript subclasses in PerKO chromatin, differences in nucleosome positioning were apparent between the classes (Fig. 2b). Well-positioned H3.3A nucleosomes were enriched downstream of the TSS loci that generate transcripts in- phase with BMAL1 binding, whereas the two other categories displayed strong TSS-flanking nucleosomes (Fig. 2b, left and middle panels). The H3.3 variant is also known to mark MNase sensitive CTCF sites, particularly in the context of dual-variant nucleosomes containing H2A.Z at insulators and around TSSs^31–33. In PerKO livers, H3.3A was highly enriched at CTCF sites associated with target genes that contained either CLOCK-BMAL1 binding sites or were co-enriched for PER2 and BMAL1 relative to the genome-wide background (Fig. 2c).

Despite the reduction in ARID2 and BRG1, total PBRM1 protein levels remained mostly unaffected in PerKO tissues (Fig. 3a, b). To confirm the depletion of PBAF complexes in PerKO livers, we immunoprecipitated endogenous PBRM1-containing complexes from FH-H3.3A and FH-H3.3A;PerKO livers collected at CT8 and CT20, using an isotype-matched IgG control. In wild-type livers, PBRM1-associated PBAF complexes were enriched for H3.3A at CT8 (Fig. 3a, and Supplementary Fig. 5a), while H2A.Z was detected at both CT8 and CT20 (Supplementary Fig. 5a). Indeed, BMAL1 was found only in the daytime PBAF complexes (Fig. 3b). In PerKO livers, where PBAF complexes were absent, BMAL1 was no longer detectable in PBRM1 complexes (Fig. 3b). Moreover, PBRM1 no longer interacted with H3.3A or H2A.Z in PerKO livers (Supplementary Fig. 5a). These findings suggest that PBAF and BMAL1 complexes predominantly associate with daytime chromatin marked by H3.3 to initiate circadian transcription.

The observed reduction in several PBAF and cBAF subunits in PerKO livers appears to be independent of transcriptional regulation. No significant differences in mRNA levels were observed between wild-type and PerKO livers for the key components Pbrm1, Arid2, Brd7, Phf10, and the Brg1 ATPase subunit (Fig. 3c). As a control, Per1 mRNA levels were depleted as expected, consistent with the disruption of circadian feedback in PerKO tissues. These findings suggest that the coordinated reduction of PBAF and cBAF subunits in PerKO livers likely arises from post-transcriptional mechanisms. This is consistent with previous reports implicating similar pathways in chromatin remodeler subunit dynamics^34,35.

In melanoma cell lines, ARID2 depletion leads to PBAF complex disassembly, redistribution of subunits, and a remodeler switch from PBAF to cBAF complexes. This switch is associated with increased chromatin accessibility and enhanced transcription factor binding³⁶. Given the reduction of ARID2 in PerKO soluble chromatin, we hypothesized that a similar 'switch' to BRM-associated remodeler complexes might occur to regulate BMAL1 function in the absence of circadian negative feedback.

Interestingly, despite the clear reduction of ARID1B subunit in total protein levels in PerKO input fraction, cBAF/BRM complex assembly remained unaffected in these animals (Fig. 4b). BMAL1 co-immunoprecipitated with BRM at CT8 and CT20, and surprisingly, even in PerKO chromatin (Fig. 4b). BRG1 ChIP-qPCR experiments showed differential enrichment between wildtype and PerKO livers, while no significant difference was seen in BRM ChIP-qPCR experiments at E-boxes of clock genes such as Per1, Per2, Cry2 as well as immediate target genes Dbp, Rev-erbα and Rorc (Fig. 4c, d). These findings suggest that ARID2 and BRG1 depletion in PerKO livers disrupts PBAF and cBAF/BRG1 complex assembly, favoring the reorganization of remodelers to cBAF/BRM complexes. This transition could likely target BMAL1 at E-boxes in the absence of negative feedback.

Among the mSWI/SNF remodeler family, active PBAF complexes exhibit a strong preference for histone H3 post-translational modifications and dual-variant nucleosomes containing H2A.Z and H3.3³⁷. Additionally, acetylation marks such as H3K115ac and H3K122ac, located near the lateral surface of the nucleosome, are associated with active chromatin states^21,22. CLOCK-BMAL1 is known to function as a pioneer transcription factor by binding directly to nucleosomes, interacting with the acidic patch, and remodeling chromatin^2,3. We hypothesized that CLOCK-BMAL1 recognizes fragile nucleosomes marked by H3K115ac and H3K122ac and sought to determine whether this interaction is altered in PerKO livers.

We noted that chromatin from PerKO livers was enriched for fragile nucleosome marks, including H3K4me3, H3K115ac, and H3K122ac, reflecting a persistent daytime-like 'active' chromatin state (Fig. 5a). Despite this enrichment, CLOCK and BMAL1 only co-immunoprecipitated with fragile H3K115ac/H3K122ac-marked homotypic H3.3 nucleosomes during the active clock phase in wild-type livers. Moreover, CLOCK and BMAL1 failed to bind fragile H3.3 nucleosomes in PerKO (Fig. 5a), despite the 'daytime-like' dual-variant chromatin state.

FH-H3.3 nucleosomes do not co-immunoprecipitate canonical H3.1/2 histones, suggesting that CLOCK-BMAL1 binding preferably favors the H3.3 variant nucleosomes and their subsequent remodeling by PBAF/cBAF complexes in wildtype livers. Interestingly, the observed increase in H3.3 deposition in PerKO chromatin (Fig. 2a, b) was not reflected in higher levels of H3.3A mRNA expression (Supplementary Fig. 6a). Total H3.3A protein expression remained unchanged across most replicates (N > 6, Supplementary Fig. 6a). We therefore wondered if changes in H3.3-specific chaperones could explain the higher loading of H3.3 in PerKO chromatin. However, nuclear protein levels of the primary H3.3 chaperones, HIRA and DAXX, were not consistently altered across multiple samples (Supplementary Fig. 6b).

Therefore, we wondered if H3K115ac marked nucleosomes retained CLOCK and BMAL1 in PerKO chromatin. To address this question, we purified endogenous H3K115ac nucleosome complexes from FH-H3.3A and FH-H3.3A;PerKO livers (Fig. 5b) using an antibody that fails to distinguish between canonical H3.1/2 and the H3.3 variant. The H3K115ac pull-down revealed fragile nucleosomes that harbor either the H3.3 variant and/or H3.1/2 canonical histones (Fig. 5b). CLOCK and BMAL1 proteins consistently associate with H3K115ac-fragile nucleosomes at subjective day and night time-points in the wild-type, as well in PerKO chromatin. In addition, H3K115ac co-immunoprecipitated with H3.3A, H2A.Z, and H3K122ac marks in all conditions. These findings indicate that circadian transcription factor complexes interact with canonical and variant H3 nucleosomes that are inherently fragile and mobile.

Finally, we analyzed H3K115ac complexes for components of the SWI/SNF remodeler subunits. Fragile H3K115ac nucleosomes retained PBAF (PBRM1, ARID2) and shared BRG1/BRM ATPase subunits at both CT8 and CT20 (Fig. 5b). However, ARID2-mediated PBAF assembly and cBAF/BRG1 complexes were disrupted on H3K115ac nucleosomes in PerKO chromatin, consistent with the general reduction in ARID2 and BRG1 proteins observed earlier (Fig. 5b, lane 9 and Fig. 3a, b). Remarkably, cBAF/BRM complex assembly on H3K115ac nucleosomes remained unaffected in PerKO chromatin, suggesting a preferential interaction of BRM remodelers with fragile canonical or variant H3 nucleosomes.

Does an increased H3.3 deposition in PerKO chromatin and the interaction with BRM remodelers impact BMAL1 binding in the absence of negative feedback? To test this, we performed native BMAL1 ChIP-seq on soluble mononucleosome-enriched chromatin from FH-H3.3A (CT8 and CT20) and FH-H3.3A;PerKO livers. Unlike crosslinked BMAL1 ChIP-seq⁴, we observed only a weak enrichment of BMAL1 on chromatin. Nevertheless, BMAL1 was enriched at H3.3 nucleosome positions that flank CLOCK-BMAL1 sites (Fig. 5c), with a greater signal at CT8 centered on binding site as well as on flanking nucleosomes. Strikingly, PerKO chromatin showed significantly enhanced BMAL1 enrichment centered at CLOCK-BMAL1 sites (Fig. 5c). These findings suggest that the fragile H3.3 chromatin landscape and altered remodeling dynamics in PerKO livers render E-box motifs more accessible. While H3.3 nucleosomes at CT8 could differ from those in PerKO livers in their PTM landscapes or the presence of other histone variants, the reduced association of BMAL1 with H3.3 complexes in PerKO chromatin (Fig. 3a) likely reflects BMAL1 directly binding E-boxes or an increased turnover at these sites.

Genetically modified heterozygous FLAG-FLAG-HA-tag (FH) - H3.3A^+/−, FH-H3.3A^+/−Per1^−/−, Per2^−/− knock-out mouse lines and wild-type (WT) C57Bl/6J male mice were used in this study. The FH-H3.3A^+/− mouse line has been generated as described previously²⁸. FH-H3.3A^+/− littermates were obtained by crossing FH-H3.3A^+/− with WT mice. Male FH-H3.3A^+/− heterozygous mice were kept for the study. The FH-H3.3A^+/−Per1^−/−, Per2^−/− knock-out line was generated by crossing FH-H3.3A^+/− with Per1^−/−, Per2^−/− knock-out mice. Animals were housed under a temperature-controlled, 12 h:12 h light-dark cycles, and were fed ad libitum. All mice used in this study were from 10- to 17-weeks old male littermates. Aged-matched littermates were randomly assigned to experimental groups. Prior tissue collection, mice were entrained to a 12 h:12 h light-dark cycle in temperature and humidity-controlled ventilated cabinets ( ~ 22 °C, 30% RH) for 2 weeks and then kept in constant darkness. Mice were euthanized at CT0, 4, 8, 12, 16 and 20 under the red light, and tissues dissected under the room light. Animal procedures were carried out according to the French guidelines for care and use of experimental animals and approved by the ethical committee of the ENS Lyon and the French ministry of science under APAFIS no. 5212-2016032914413553.

Liver tissues from FH-H3.3A^+/− and WT animals were collected at 1 day in constant darkness and from FH-H3.3A^+/−Per1^−/−, Per2^−/− KO animals at 1 or 4 days in constant darkness. Livers were mainly processed fresh, followed by nuclei purification immediately after tissue isolation, or fresh-frozen and stored at −80 °C until use. Nuclei were purified as described in ref. ⁵⁰ and stored in Nuclear Lysis Buffer at 1X final with 10% glycerol (NLB 10X stock: 10 mM HEPES-KOH pH 7.6, 100 mM KCl, 0.1 mM EDTA-NaOH pH 8.0) at −80 °C until use.

Nuclei isolated from FH-H3.3A^+/− livers collected at CT8 and CT20 and FH-H3.3A^+/−Per1^−/−, Per2^−/− KO livers collected at CT20 were used for the native purification of H3.3A protein complexes. Nuclei from WT livers collected at random time points were used as negative control for the experiment. Total nuclear protein quantity was estimated by lysing 10ul of purified nuclei with RIPA lysis buffer (Invitrogen) for 20 min on ice and sonicated for 6 cycles on Diagenode bioruptor (30 s on/ 30 s off) at 4 °C, followed by centrifugation at 14,000 rpm (18,407 g) for 10 min. Released proteins in the supernatant were quantified using DC protein assay (Bio-Rad, 5000116).

MNase-digested nuclear extracts were prepared from one full liver ( ~ 1.4 mg of total protein) for each condition according to Khan et al.⁵¹ with modifications, as follows. Nuclei were washed twice in Buffer A (10 mM HEPES-KOH pH 8.0, 10 mM KCl, 1.5 mM MgCl₂, 340 mM Sucrose, 10% Glycerol, 1 mM DTT) containing protease inhibitors and centrifuged at 4500 rpm (1902g) for 4min30. Nuclear pellets were resuspended in Cutting Buffer (15 mM NaCl, 60 mM KCl, 10 mM Tris-HCl pH 7.5, 2 mM CaCl₂) and freshly diluted MNase (Nuclease S7, Roche 10107921001, batch number 53964400; resuspended and stored in 10 mM Tris-HCl pH 7.5, 50 mM NaCl, 1 mM EDTA-NaOH pH 8.0, 50% glycerol) in Cutting buffer was added at 10U per sample. To digest chromatin to mononucleosomes, samples were vortexed at low speed and incubated in water-bath at 37 °C for 7 min, vortexed every 1min30. After digestion, samples were kept on ice and MNase activity was stopped by adding EGTA-NaOH pH 8.0 at 20 mM final. Samples were vortexed and centrifuged at 1300 g for 5 min. The totality of the supernatant was kept in separate tubes on ice as S1 MNase-digested fraction. Remaining pellets were resuspended in TE buffer (10 mM Tris-HCl pH 8.0, 1 mM EDTA-NaOH pH 8.0) containing protease inhibitors and incubated on end-to-end rotator at 4 °C for 30 min. After incubation, samples were centrifuged at 13,000 g for 5 min and the totality of the supernatant was kept in separate tubes on ice as S2 TE-soluble fraction. Salt concentration was adjusted for the totality of both, S1 and S2 fractions as follows. Equal sample volume of 2XE buffer (30 mM HEPES-KOH pH 7.5, 225 mM NaCl, 3 mM MgCl₂, 0.4% Triton X-100, 20% Glycerol) for the S1 fraction, and half of the sample volume of 3XD buffer (60 mM HEPES-KOH pH 7.5, 450 mM NaCl, 4.5 mM MgCl₂, 0.6 mM EGTA-NaOH pH 8, 0.6% Triton X-100, 30% Glycerol) for the S2 fraction was added dropwise on vortex. Samples were centrifuged at 13,000 g for 5 min and supernatants from both fractions were pooled together respectively for each condition in separate tubes, and kept on ice as nuclear extracts. All centrifugation steps were performed at 4 °C and all vortex steps at the lowest speed.

Prior purification, 5% of each nuclear extract was saved as input for further analysis. Nuclear extracts were incubated with prewashed Anti-FLAG M2 Affinity Gel agarose beads (Sigma-Aldrich), over-night (18 h) on an end-to-end rotator at 4 °C. After incubation, samples were centrifuged at 1500 rpm (211 g), 2 min and the unbound fraction was isolated. Sequential washing steps were applied to remove non-specific binding, as follows. Beads were washed 4 times with 1XD buffer made from 3XD, followed by 4 washes with 1XD buffer made from 3XD + 0.5% Triton X-100, followed by 2 washes with 1XD buffer without Triton X-100. Each washing step was done on rotation at 4 °C for 5 min, followed by centrifugation at 1500 rpm, 2 min. After the final wash, beads were dried and specifically-bound proteins eluted by a tandem elution with the 3xFLAG peptide (Sigma-Aldrich). Both elution steps were performed with 3xFLAG peptide at 500 ug/ml final concentration, resuspended and diluted in TEGN buffer (20 mM Tris-HCl pH 7.65, 0.1 mM EDTA-NaOH pH 8.0, 10% Glycerol, 150 mM NaCl, 0.01% NP-40), for 3 h at room temperature with tilting mode. Both eluates were pooled respectively for each condition and 40% of eluates kept to further assess the specificity of the pull-down by silver staining (SilverQuest Silver Staining kit, Invitrogen) and western-blotting. Remaining eluates were submitted for Mass Spectrometry analysis.

Label-free quantitative Mass Spectrometry analysis was performed at Protein Sciences Facility - SFR BioSciences (UAR3444/US8) in Lyon, using Q Exactive HF (Thermo Scientific) Mass Spectrometer coupled with nanoUHPLC U3000 (Thermo Scientific). Briefly, samples were reduced and alkylated with tris-(2-carboxyethyl) phosphine/ iodoacetamide (TCEP/IAA) and fixed on SP3 beads prior digestion. Beads were washed with 80% ethanol and protein digested in ammonium bicarbonate (100 mM), using Lys-C/Trypsin protease (enzyme/protein ratio: 1/100) for 18 h at 37 °C. Supernatants with peptides were recovered after centrifugation at 20,000 g for 1 min and peptide concentration was determined for each sample by quantitative fluorometric peptide assay (Thermo Scientific) to inject 400 ng of peptides in the nanoLC-MS/MS system.

The two biological replicates for each condition were injected in triplicates and loaded on a C18 Acclaim PepMap100 trap-column 300 µm ID x 5 mm, 5 µm, 100 Å, (Thermo Scientific) for 3.0 minutes at 20 µL/min with 2% ACN, 0.05% TFA in H2O and then separated on a C18 Acclaim Pepmap100 nano-column, 50 cm×75 µm i.d, 2 µm, 100 Å (Thermo Scientific) with a 60 minutes linear gradient from 3.2% to 40% buffer B (A: 0.1% FA in H2O, B: 0.1% FA in ACN), from 40% to 90% of B in 2 min, hold for 10 min and returned to the initial conditions in 1 min for 14 min. The total duration was set to 90 minutes with a flow rate of 300 nL/min and the oven temperature was kept constant at 40 °C.

Samples were analyzed with a TOP20 DDA HCD method: MS data were acquired in a data dependent strategy selecting the fragmentation events based on the 20 most abundant precursor ions in the survey scan (350-1600 Th). The resolution of the MS scan was 60,000 at m/z 200 Th and for MS/MS scan the resolution was set to 15,000 at m/z 200 Th. Parameters for acquiring HCD MS/MS spectra were as follows: normalized collision energy = 27 and an isolation window width of 2 m/z. The precursors with unknown charge state, charge state of 1 and 8 or greater than 8 were excluded. Peptides selected for MS/MS acquisition were then placed on an exclusion list for 20 s using the dynamic exclusion mode to limit duplicate spectra.

Sample file was treated with Proteome Discoverer 2.5 software (Thermo Scientific), using SEQUEST HT research engine and Swiss-Prot Mus musculus database (august 2021 release, 17500 entries), supplemented with protein contaminant database. Precursor mass tolerance was set at 10 ppm and fragment mass tolerance was set at 0.02 Da, and up to 2 missed cleavages were allowed. Oxidation (M), phosphorylation (S, T, Y), acetylation (Protein N-terminus) were set as variable modifications and carbamidomethylation (C) as fixed modification. False discovery rate (FDR) of peptide identifications was calculated by the Percolator algorithm method, and a cut-off FDR value of 1 % was used. Protein quantification was done by Label Free Quantification (LFQ) approach, and LFQ abundance values were obtained and normalized to the total peptide amount. Protein quantitation was performed with precursor ions quantifier node in Proteome Discoverer 2.5 software with protein quantitation based on pairwise ratios and hypothesis t-test. Proteins were considered as differentially expressed between the two conditions when FC > 2 and p-value < 0.05. Adjusted p-values were obtained by applying the Benjamini-Hochberg procedure. Generated data files were further analyzed taking into account following parameters in order to determine specific candidates of FH-H3.3A pull-down: 'not found in WT', number of peptides, number of unique peptides, Score Sequest HT: Sequest HT, abundance ratio and adjusted p-value abundance ratio for each peptide of given protein.

Native immunoprecipitation experiments were performed on MNase-digested nuclear extracts from wild-type, FH-H3.3A^+/− and FH-H3.3A^+/−Per1^−/−, Per2^−/− KO nuclei purified from livers collected either every 4 h over 24 h, or at CT8 and CT20. Nuclei from WT livers collected at random timepoints, or isotype matched Anti-Mouse IgG antibody (Jackson ImmunoResearch, 315-005-003) were used as negative controls for the IP.

Nuclear extracts were prepared as above, with few following modifications. ~530ug of total protein were used for the IP, with all buffer volumes reduced by half and diluted MNase was used at 5U per sample. Prior IP, 5% of each nuclear extract was saved as input, mixed with 1.5X SDS-blue loading buffer containing DTT at 100 mM, denatured at 80 °C for 7 min and stored at -20 °C for further analysis. Nuclear extracts were incubated overnight (18 h) on an end-to-end rotator at 4 °C with 2 - 4.5ug of following antibodies: anti-HA-Tag (C29F4) (Cell Signaling, #3724); anti-PBRM1/BAF180 (E9X2Z) (Cell Signaling, #89123); anti-BRG1/SMARCA4 (Bethyl Laboratories, A300-813A); anti-SMARCA2 antibody [HL1115] (GENETEX, GTX636330) and anti-Acetyl-Histone H3 (Lys115) (PTM Bio, PTM-170). Following overnight incubation with antibodies, equal volume of Protein A and Protein G Dynabeads (Invitrogen, 10002D, 10004D) mix was added per sample and nuclear extracts were incubated on an end-to-end rotator for 3-4 h at 4 °C. Unbound fraction was isolated and beads were washed 2-3 times with 1XD buffer made from 3XD, on rotation at 4 °C for 5 min. After final wash, beads were resuspended in 2.5X SDS-blue loading buffer containing DTT and proteins eluted for 10 min at room temperature, with shaking mode. Protein eluates were transferred to separate tubes, denatured at 80 °C for 7 min and stored at -20 °C.

SDS-PAGE electrophoresis was performed using precast Novex NuPAGE 4–12% Bis-Tris gradient gels and NuPAGE SDS MOPS Running buffer (20X) (Invitrogen) at 1X final. Prestained Protein Ladder Plus (Euromedex, 06P-0211) was used as a reference marker for protein size. For each IP experiment, 0.6–1% of input and one-third of IP eluates were loaded per lane. Transfer was performed on 0.2μm nitrocellulose membranes (Amersham, GE10600004) at constant current (0.4 A) for 2–3 h at 4 °C, using NuPAGE SDS MOPS Running buffer at 1X with 5% ethanol. After transfer, membranes were rinsed with Millipore water and incubated in 5% non-fat dry milk- PBS 1X blocking solution for 30 min at room temperature. After blocking, membranes were first rinsed with Millipore water, then with PBS 1X- 0.1% Tween 20 wash buffer prior overnight incubation at 4 °C with respective primary antibodies. After incubation with primary antibodies, membranes were washed 2-3 times with PBS 1X- 0.1% Tween 20 wash buffer for 5-10 min and incubated with respective secondary antibodies for 1 h–1 h.30 at room temperature. Peroxidase Goat Anti-Rabbit IgG (H + L) (Jackson ImmunoResearch, 111-035-144) or Goat Anti-Mouse IgG (H + L)- HRP Conjugate (Bio-Rad, #1706516) secondary antibodies were used at 1:10,000 in 5% milk- PBS 1X. Membranes were washed 2-3 times with PBS 1X- 0.1% Tween 20 wash buffer for 5–10 min, and incubated with ECL Prime western blotting reagent (CYTIVA, RPN2232) prior developing. Image acquisitions were performed using ChemiDoc Imaging system (Bio-Rad) and ImageLab software (version 6.0.1).

Following primary antibodies were used for western-blotting in this study, dilutions were prepared in 3% BSA- PBS 1X as indicated: HA-Tag (C29F4) Rabbit mAb (Cell Signaling, #3724), 1:1000; PBRM1/BAF180 (E9X2Z) Rabbit mAb (Cell Signaling, #89123), 1:1000; ARID2 (GT7311) Mouse mAb (Sigma-Aldrich, SAB2702340), 1:500; ARID2 (GT7311) Mouse mAb (GENETEX, GTX632011), 1:500; BRG1/SMARCA4 Rabbit pAb (Bethyl Laboratories, A300-813A), 1:500; BRD7 Rabbit pAb (Proteintech, 51009-2-AP), 1:500; PHF10 Mouse mAb (Proteintech, 66341-1-Ig), 1:500; PHF10 Rabbit pAb (GENETEX, GTX116314), 1:500; BMAL1 Rabbit pAb (Bethyl Laboratories, A302-616A), 1:1000; CLOCK Mouse mAb (MBL, D334-3), 1:500; Histone H3.1/3.2 (1D4F2) Mouse mAb (Active Motif, 61629), 1:500; Tri-Methyl-Histone H3 (Lys4) Rabbit pAb (Cell Signaling, #9727), 1:500; Acetyl-Histone H3 (Lys115) Rabbit pAb (PTM Bio, PTM-170), 1:1000; Histone H3 (acetyl K122) Rabbit pAb (Abcam, ab33309), 1:500; Histone H2A.Z Rabbit pAb (Active Motif, 39113), 1:1000; Homemade H2A.Z Rabbit pAb (kind gift from Stefan Dimitrov), 1:500; ARID1B Rabbit pAb (GENETEX, GTX130708), 1:500; SMARCA2 (HL1115) Rabbit mAb (GENETEX, GTX636330), 1:500; Histone H3 Rabbit pAb (Proteintech, 17168-1-AP), 1:2000 in 1% milk- PBS 1X; Histone H4 Rabbit pAb (Proteintech, 16047-1-AP), 1:2000 in 1% milk- PBS 1X; U2AF65 Rabbit mAb (Abcam, ab197031), 1:500; HIRA (clone WC119) Mouse mAb (Sigma-Aldrich, 04-1488), 1:500; DAXX anti-Mouse (Precision antibody), 1:500.

If the proteins of interest were migrating at the same size, mainly different gels and membranes were used. In some cases, primary anti-ARID2 mouse antibody was used first and after extensive stringent washing for few hours, membranes were probed in primary anti-BRG1 rabbit antibody (before probing in primary rabbit antibody, extensively washed membranes were incubated with the anti-mouse secondary antibody and developed to verify the removal of the signal). Membranes were probed in the following order: PBRM1 and CLOCK (on the same membranes); ARID2, BRG1 on different membranes (and both on the same membranes in Fig. 1a replicate 1 and 2, Fig. 1b replicate 2, Fig. 3a replicate 1 and 3, Fig. 3b replicate 1 and 2; see Source Data file for the replicates); BMAL1, BRD7 on different membranes (and on the same membrane in Fig. 3a replicate 1); PHF10 on different membranes; ARID1B, SMARCA2 (BRM) on different membranes; the histones and histone marks were probed mainly on different membranes, except in Fig. 3b replicates 1 and 3 for H3K4me3 and HA-tag; Fig. 5a replicate 1 for H2A.Z and H3K4me3 (indicated respectively to the probing order). Fig. 5b replicate 1 for H3K122ac and HA-tag, replicate 2 for H3K115ac and HA-tag; Fig. 5b all replicates for H2A.Z and H3.1/3.2.

Native chromatin-immunoprecipitation (ChIP) experiments were performed on MNase-digested nuclear extracts from FH-H3.3A^+/− and FH-H3.3A^+/−Per1^−/−, Per2^−/− KO nuclei purified from livers collected either every 4 h over 24 h, or at CT8 and CT20. Nuclear extracts were prepared with ~350–530 ug of total protein, according to Khan et al. 2020 and as described above for the IP experiments. DNA quantity in nuclear extracts was measured on Nanodrop 2000c spectrophotometer (ThermoFisher Scientific). ~15–30 ug of chromatin was used per sample for ChIP experiments and 1% of each nuclear extract was saved as input. Inputs from CT0 - CT20 circadian nuclear extracts (related to FH-H3.3A ChIP) and inputs from CT8 and CT20 (related to H3.1/3.2 and BMAL1 ChIP) nuclear extracts were pooled together and stored at −80 °C for further analysis. Nuclear extracts were incubated overnight (18 h) on an end-to-end rotator at 4 °C with 2–4 ug of following antibodies: anti-HA-Tag (C29F4) (Cell Signaling, #3724); anti-Histone H3.1/3.2 (1D4F2) (Active Motif, 61629) and anti-BMAL1 (Bethyl Laboratories, A302-616A). Following overnight incubation with antibodies, equal volume of Protein A and Protein G Dynabeads (Invitrogen) mix was added per sample and nuclear extracts were incubated on an end-to-end rotator for 3–4 h at 4 °C. Unbound fraction was isolated and beads were washed 3 times with 1XD buffer made from 3XD, on rotation at 4 °C for 5 min. After final wash, beads were resuspended in ChIP Elution buffer (1% SDS, 100 mM NaHCO₃) and co-immunoprecipitated DNA eluted at 65 °C, with shaking mode at 600 rpm for 15 min. Recovered eluates and inputs were treated with Proteinase K at 10 mg/ml (NEB Molecular Biology Grade, P8107S) at 37 °C for 2h30, followed by DNA purification using QIAquick PCR Purification Kit (QIAGEN). Purified ChIP DNA was stored at -80 °C until further analysis. Concentration of purified DNA samples was assessed on Qubit 4 Fluorometer (Invitrogen) using Qubit dsDNA HS Assay Kit (Invitrogen). Quality of purified DNA samples was assessed on Agilent 4150 TapeStation System using High Sensitivity D5000 ScreenTape assay (Agilent) and analyzed with Agilent TapeStation Software 5.1 prior sequencing.

Sequencing libraries were prepared with NEBNext UltraII DNA Library Prep Kit for Illumina sequencing, enriching for fragments from 100-150 bp with the 0.9X beads. 150b paired-end (PE) sequencing was performed on Illumina NovaSeq6000 system using NovaSeqS4 flow cell. Approximately 30,000,000 + PE reads were obtained for each replicate (Supplementary Data 3). These were filtered to remove poor quality reads (Q < 20), short reads ( < 75b) and Illumina adapters using cutadapt 4.5, the results were visualized using FastQC 0.12.1. The trimmed reads from each replicate were then mapped separately on the mouse genome GRCm39 with Bowtie2 2.5.4⁵² using – very-sensitive—no-mixed—no-discordant and—dovetail options. Samtools 1.18⁵³, command view -F 1804 -q 20 -f 2, was used to remove poor quality mappings ( < 20), unmapped reads and their mate, non-primary alignments and reads not correctly matched. The resulting alignment was then filtered to remove duplicates and blacklisted regions, following ENCODE recommendations⁵⁴, using Picard MarkDuplicates 3.1.1 (Broad Institute 2019) and Bedtools intersect 2.30.0⁵⁵ respectively. The alignment of each replicate was then down-sampled to ensure that the same number of sequences were present for each replicate of the same ChIP Seq sample, and thus avoid over-representation of any one of them (Supplementary Data 3). This sub-sampling was carried out using samtools view, and the alignments were sorted using samtools sort. The input samples (ChIP seq control samples) were processed in the same way as the others samples.

In order to compare the alignments obtained for each sample and represent them, deepTools 3.5.5⁵⁶ was selected to create a matrix of intensity scores (computeMatrix—beforeRegionStartLength 1000—afterRegionStartLength 1000, combined with plotProfile) from the normalized alignments (bamcoverage—normalizeUsing RPGC—Mnase—maxFragmentLength 200—minFragmentLength 80). These were then plotted either as a heatmap (plotHeatmap) or as an average profile plot (plotProfile). The regions of interest studied in these graphs were the regions of interest were the CLOCK-BMAL1 binding sites recovered from the work of Trott and Menet (2018)³⁰, the CTCF binding sites from ENCODE project (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSM918715↗) and the SINEs class annotation from UCSC Genome Browser (https://genome-euro.ucsc.edu/↗). If the sites were not annotated for the mm39 mouse genome, then the coordinates were converted using UCSC genome (https://genome.ucsc.edu/cgi-bin/hgLiftOver↗).

Cross-linked material was directly prepared during the nuclei purification from isolated livers as following. Tissues were grinded in PBS 1X without protease inhibitors. Filtered tissue homogenates were then cross-linked with Formaldehyde solution (Sigma) at 1% final on rotation for 5 min at room temperature, followed by quenching with Glycine at 150 mM final on rotation for 3 min at room temperature. Subsequent nuclei purification steps remained unchanged and done as described in ref. ⁵⁰, except that Glycine at 150 mM final was added to the sucrose solutions prior ultracentrifugation step. Purified nuclei were stored in Nuclear Lysis Buffer at 1X final with 10% glycerol (NLB 10X stock: 10 mM HEPES-KOH pH 7.6, 100 mM KCl, 0.1 mM EDTA-NaOH pH 8.0) at -80 °C until further use.

To proceed for chromatin preparation, cross-linked nuclei were washed once with 1X Ripperger buffer (Ripperger 10X stock: 200 mM Tris-HCl pH 7.5, 1.5 M NaCl, 20 mM EDTA-NaOH pH 8.0), pelleted at 4000 rpm (1503 g) for 2 min and resuspended again in 1X Ripperger buffer. 1X Ripperger buffer containing 2% SDS was added to the samples and after immediate homogenization, samples were incubated on ice for 30 min, vortexed at low speed every 10 min during the incubation time. Samples were then sonicated 4 times for 3 cycles on Diagenode bioruptor (30 s on/ 30 s off; total 12 cycles) to obtain a smear between 100–300 bp with enrichment for mononucleosomes at 150 bp. In between each 4 cycles, samples were quickly warmed up at 37 °C to dissolve precipitated SDS. Following sonication, samples were centrifuged at 10,000 rpm (9391 g) for 10 min and supernatants kept as cross-linked chromatin for the ChIP experiments, stored at −80 °C until further use. All centrifugation steps were done at 4 °C.

In order to determine DNA concentration prior ChIP experiments, 5% of cross-linked chromatin were decrosslinked in ChIP Elution buffer (1% SDS, 100 mM NaHCO₃) for 2 h at 65 °C, treated with Proteinase K at 10 mg/ml (NEB Molecular Biology Grade, P8107S) for 1h30 at 37 °C, and DNA purified using QIAquick PCR Purification Kit (QIAGEN). DNA quantity was measured on Nanodrop 2000c spectrophotometer and sonication efficiency checked on 1.2% agarose gel. For each ChIP, 30 ug of cross-linked chromatin were mixed with 1X Ripperger buffer containing 1% Triton X-100 to get the final concentration of SDS at 0.1% within samples to avoid its precipitation at 4 °C. 1% of each sample was kept as input. ChIP was performed respectively with 2.5 ug and 4 ug of anti-BRG1/SMARCA4 (Bethyl Laboratories, A300-813A) and anti-SMARCA2 antibody [HL1115] (GENETEX, GTX636330), overnight (18 h) on an end-to-end rotator at 4 °C. Following overnight incubation with antibodies, equal volume of Protein A and Protein G Dynabeads (Invitrogen) mix was added per sample and chromatin was incubated on an end-to-end rotator for 3–4 h at 4 °C. Unbound fraction was isolated and beads were first washed 2 times with Wash buffer 1 (0.1% SDS, 1% Triton X-100, 2 mM EDTA-NaOH pH 8.0, 20 mM Tris-HCl pH 8.1, 150 mM NaCl) for 5 min, followed by one wash with Wash buffer 2 (0.1% SDS, 1% Triton X-100, 2 mM EDTA-NaOH pH 8.0, 20 mM Tris-HCl pH 8.1, 500 mM NaCl) for 5 min. All washes were done on rotation at 4 °C. After final wash, beads were resuspended in ChIP Elution buffer (1% SDS, 100 mM NaHCO₃) and co-immunoprecipitated DNA eluted at 65 °C, with shaking mode at 600 rpm for 20 min. Recovered eluates and inputs were decross-linked at 65 °C for 2 h and treated with Proteinase K at 10 mg/ml for 1h30 at 37 °C, followed by DNA purification using QIAquick PCR Purification Kit. Purified ChIP DNA was stored at -80 °C until further analysis.

Real-time qPCR reactions were performed as follows. Each sample was deposited in three wells as technical replicates on Hard-Shell PCR 96-well thin wall plates (Bio-Rad). Reaction master mixes were prepared using ITaq Universal SYBR Green Supermix (Bio-Rad) and specific forward/ reverse primer sets. RT-qPCR were performed with CFX96 Real-Time PCR Detection System (Bio-Rad) and analyzed by CFX Maestro Software (version 2.3). Enrichment at clock-gene E-boxes (Per1, Dbp, Per2, Reverba, Cry2, Rorc) was assessed, calculated as percentage of the immunoprecipitated input. All the results are represented as the average ± s.e.m of three independent biological replicates. Primers for these regions⁵⁷ were synthesized by Eurogentec; primer sequences are available in the Supplementary Data 4.

Whole liver extracts from FH-H3.3A^+/− and FH-H3.3A^+/−Per1^−/−, Per2^−/− KO livers collected at CT8 and CT20 were obtained during the nuclei purification as follows. Tissues were grinded in PBS 1X containing protease inhibitors and aliquots of tissue homogenates were stored at −80 °C until further use. Total RNA was extracted from whole liver extracts with TRIzol LS Reagent (Invitrogen) according to manufacturer's instructions and quantified on Nanodrop 2000c spectrophotometer. cDNA was prepared from 2 ug of purified RNA using IScript cDNA Synthesis kit (Bio-Rad) according to manufacturer's instructions. Real-time qPCR reactions were performed as follows. Each sample was deposited in two to three wells as technical replicates on Hard-Shell PCR 96-well thin wall plates (Bio-Rad). Reaction master mixes were prepared using ITaq Universal SYBR Green Supermix (Bio-Rad) and specific forward/ reverse primer sets. RT-qPCR were performed with CFX96 Real-Time PCR Detection System (Bio-Rad) and analyzed by CFX Maestro Software (version 2.3). All the results are represented as the average ± s.e.m. of three independent biological replicates of relative expression calculated with the 2^{- ΔΔCt} method and normalized to Rps9 cDNA levels. Primer sequences were mainly obtained from the Harvard PrimerBank database and synthesized by Eurogentec; they are available in the Supplementary Data 4.

FLAG-HA-H3.3B f/f animals were crossed to H3.3A f/f strain⁵⁸ and embryos were sacrificed at E14 to generate mouse embryonic fibroblasts. Cell lines were cultured at 37 °C with a 5% CO₂ atmosphere in DMEM-high glucose medium (Sigma-Aldrich) supplemented with 10% FBS (Gibco) and 1% penicillin-streptomycin at 25 units/ml (ThermoFisher).

Circadian bioluminescence rhythms were monitored on confluent wild-type and tamoxifen-treated double-floxed FH-H3.3A^−/−, FH-H3.3B^−/− ERT2-Cre Bmal1: luciferase reporter mouse embryonic fibroblasts. To obtain H3.3 knock-out, cells were treated with 4-OHT, cis/trans-4-hydroxytamoxifen (Sigma-Aldrich) at 5 μM for 3.5 days and knock-out efficiency was verified by western-blotting. Both WT and KO cells were seeded at high confluency in a black 24 well plate with flat and clear bottom for high-throughput microscopy (Ibidi) and cells were synchronized with Dexamethasone (Sigma-Aldrich) at 100 nM for 1h30. After synchronization, cell culture medium was replaced, supplemented with HEPES (Gibco) at 10 mM, sodium bicarbonate solution (Gibco) at 1% and 200 μM of D-Luciferin Sodium Salt (PJK GmbH) and bioluminescence was monitored with LumiCycle 96 luminometer for circadian biology (ACTIMETRICS). LumiCycle Data Analysis software was used to visualize the data, to perform detrending analysis and calculation of oscillation period and amplitude.

Nucleoplasm and chromatin fractions were extracted from FH-H3.3A^+/− and FH-H3.3A^+/−Per1^−/−, Per2^−/− KO liver pieces collected at CT8 and CT20, using Subcellular Protein Fractionation Kit for Tissues (Thermo Scientific) according to the manufacturer's instructions. Whole liver lysates were obtained by incubating small liver pieces from the same tissues as for the fractionation in RIPA lysis buffer for 30 min on ice, then sonicated for 6 cycles on Diagenode bioruptor (30 s on/ 30 s off) at 4 °C, followed by centrifugation at 14,000 rpm for 10 min. Proteins from all fractions were quantified using DC protein assay.

Unsynchronized confluent murine wild-type and Per1^−/−,Per2^−/−,Per3^−/− triple-knockout (PerTKO) lung fibroblasts were treated with MG132 proteasome inhibitor (Sigma-Aldrich) for 4 h prior chromatin extraction. Subcellular protein fractionation was performed using Subcellular Protein Fractionation Kit for Cultured Cells (Thermo Scientific) according to the manufacturer's instructions. Whole cell lysates were obtained by incubating a part of the same cell pellets that were used for the fractionation, in RIPA lysis buffer for 20 min on ice, then sonicated for 6 cycles on Diagenode bioruptor (30 s on/ 30 s off) at 4 °C, followed by centrifugation at 14,000 rpm for 10 min. Proteins from all fractions were quantified using DC protein assay.

To verify whether specific PBAF and cBAF complex components have impact on circadian clock regulation, we searched and reanalyzed data from the previously published genome-wide siRNA screen for modifiers of the circadian clock⁴⁵. Data from all available replicates for the siRNAs of target genes, Cry2 positive and negative controls were taken into account from all plates. Technical duplicates for given control siRNAs were averaged to obtain a single value and thus in total four traces for each siRNA, corresponding to two averaged duplicates from each plate. For the siRNAs of target genes, since one technical duplicate was available on each plate, obtained values were kept without calculating the average and the four traces correspond to individual technical replicates. Statistical analysis for period length was performed with available data from the published study (see NIHMS144235-supplement-01 file⁴⁵ for period distribution for negative and positive control siRNA, respectively 25.20 h ± 0.55 for and 29.43 h ± 0.65; BioGPS centralized gene-annotation portal⁵⁹ with Circadian Layout plugin—Circadian Genomics Screen Database - BioGPS - your Gene Portal System for respective target siRNA made available by the authors of the genome wide RNAi screen⁴⁵). For the negative and positive control siRNA and Phf10 screen, data from only two replicates were available in the database. In order to perform statistical analysis for all replicates of all target genes, the duplicates of negative control siRNA were taken twice into account for the t-test analysis.

To assess the impact of H3.3 loss on core-clock and clock output gene expression, we collected non-synchronized H3.3 wild-type (untreated) and H3.3DKO (tamoxifen-treated) MEFs 6.5 d post-treatment. Total RNA was extracted from cell pellets with TRIzol LS Reagent (Invitrogen) according to manufacturer's instructions and quantified on Nanodrop 2000c spectrophotometer. cDNA was prepared from 2 ug of purified RNA using IScript cDNA Synthesis kit (Bio-Rad) according to manufacturer's instructions. Real-time qPCR reactions were performed as follows. Each sample was deposited in three wells as technical replicates on Hard-Shell PCR 96-well thin wall plates (Bio-Rad). Reaction master mixes were prepared using ITaq Universal SYBR Green Supermix (Bio-Rad) and specific forward/ reverse primer sets. RT-qPCR were performed with CFX96 Real-Time PCR Detection System (Bio-Rad) and analyzed by CFX Maestro Software (version 2.3). All the results are represented as the average ± s.e.m. of three independent biological replicates of relative expression calculated with the 2^{- ΔΔCt} method and normalized to Rps9 cDNA levels. Primer sequences were mainly obtained from the Harvard PrimerBank database and synthesized by Eurogentec; they are available in the Supplementary Data 4.

In order to assess the impact of PBAF-cBAF specific components on core-clock and clock output gene expression, we performed siRNA-mediated knockdown of Pbrm1, Arid2, Brd7, Phf10, Smarca4 (Brg1), Smarca2 (Brm) and Arid1b in U2OS Bmal1:Luciferase cell line. 30-50% confluent non-synchronized cells were transfected with either siRNA Negative Control, or a specific siRNA targeting a given PBAF/cBAF complex component, using INTERFERin® (Polyplus) transfection reagent according to the manufacturer's instructions. siRNA duplex sequences were provided and synthetized by Eurogentec on request. Total RNA was extracted from cell pellets with TRIzol LS Reagent (Invitrogen) according to manufacturer's instructions and quantified on Nanodrop 2000c spectrophotometer. cDNA was prepared from 2 ug of purified RNA using IScript cDNA Synthesis kit (Bio-Rad) according to manufacturer's instructions. Real-time qPCR reactions were performed as follows. Each sample was deposited in three wells as technical replicates on Hard-Shell PCR 96-well thin wall plates (Bio-Rad). Reaction master mixes were prepared using ITaq Universal SYBR Green Supermix (Bio-Rad) and specific forward/ reverse primer sets. RT-qPCR were performed with CFX96 Real-Time PCR Detection System (Bio-Rad) and analyzed by CFX Maestro Software (version 2.3). All the results are represented as the average ± s.e.m. of three independent biological replicates of relative expression calculated with the 2^{- ΔΔCt} method and normalized to Rps9 cDNA levels. Primer sequences were mainly obtained from the Harvard PrimerBank database and synthesized by Eurogentec; they are available in the Supplementary Data 4. Part of the same cell pellets was kept to assess the knockdown efficiency by western-blotting; whole cell extracts in RIPA buffer were obtained as described above.

Further information on research design is available in the linked to this article. Nature Portfolio Reporting Summary