Cooperation between bHLH transcription factors and histones for DNA access

Human histones were expressed and purified as described previously⁵⁷. Lyophilized histones were mixed at equimolar ratios in 20 mM Tris-HCl (pH 7.5) buffer, containing 7 M guanidine hydrochloride and 20 mM 2-mercaptoethanol. Samples were dialysed against 10 mM Tris-HCl (pH 7.5) buffer, containing 2 M NaCl, 1 mM EDTA and 2 mM 2-mercaptoethanol. The resulting histone complexes were purified by size-exclusion chromatography (Superdex 200; GE Healthcare). For MYC-MAX TR-FRET experiments, H2B-biotinylated octamers were prepared using a T122C mutant introduced into H2B using site-directed mutagenesis. The purified H2A–H2B(T122C) complex (46 μM) was conjugated with biotin using 558 μM EZ-Link Maleimide-PEG2-Biotin (Thermo Fisher Scientific) in 10 mM Tris-HCl (pH 7.5) buffer, containing 2 M NaCl, 1 mM EDTA and 1 mM TCEP, at room temperature for 2 h. The reaction was stopped by adding 2-mercaptoethanol and the sample was then dialysed against 10 mM Tris-HCl (pH 7.5) buffer containing 2 M NaCl, 1 mM EDTA and 5 mM 2-mercaptoethanol. Reconstitutions of the H2A–H2B(T122C–biotin) complex, the H3.1–H4 complex and the histone octamer were performed as described previously⁵⁸.

DNA for medium- to large-scale individual nucleosome purifications was generated by Phusion (Thermo Fisher Scientific) PCR amplification. The resulting DNA fragment was purified by a Mono Q column (GE Healthcare). All purified DNA was concentrated and stored at −20 °C in 10 mM Tris-HCl pH 7.5 until use. Labelled DNA for smTIRF experiments was also generated using PCR with fluorescently labelled primers (Sigma Aldrich, see Supplementary Table). 1

A mutant version of the SpyCatcher protein (SpyCatcherS50C) was purified following previously established protocols^63,64. SpyCatcherS50C was incubated with DTT (8 mM) at 4 °C for 1 h. DTT was removed using a S200 16/60 gel filtration column (GE healthcare) in a buffer containing 50 mM Tris-HCl pH 7.3 and 150 mM NaCl. JF549-maleimide (Tocris) was dissolved in 100% DMSO and mixed with SpyCatcher to achieve a fourfold molar excess of JF549-maleimide. SpyCatcher was labelled at room temperature for 3 h in a vacuum desiccator and stored overnight at 4 °C. Labelled SpyCatcher was separated from free dye on a S200 16/60 gel filtration column in 50 mM Tris-HCl pH 7.5, 150 mM NaCl, 250 μM TCEP and 10% (v/v) glycerol, concentrated, flash-frozen in liquid nitrogen and stored at −80 °C. Purified wild-type MYC-MAX–Spy, MYC-MAX^Y73A-R76A–Spy and MYC^S405Y-A408R-MAX–Spy were mixed with JF549–SpyCatcher in a 5:1 molar ratio and incubated for 1 h at room temperature, frozen in liquid nitrogen.

Measurements were performed as described previously⁶⁵. In brief, objective-type smTIRF was performed using a Nikon Ti-E inverted fluorescence microscope, equipped with a CFI Apo TIRF 100× oil immersion objective (NA 1.49), an ANDOR iXon EM-CCD camera and a TIRF illuminator arm. Laser excitation was realized using a Coherent OBIS 640LX laser (640 nm, 40 mW) and coherent OBIS 532LS laser (532 nm, 50 mW). For all smTIRF experiments, flow channels were prepared as described before⁶⁵, washed with 500 µl degassed ultrapure water (Romil), followed by 500 µl 1× T50 (10 mM Tris pH 8, 50 mM NaCl) and background fluorescence was recorded with both 532 nm and 640 nm excitation. Fifty microlitres of 0.2 mg ml⁻¹ neutravidin was then injected and incubated for 5 min and washed using 500 µl 1×T50. Then, 50 pM of Alexa647-labelled DNA or NCPs in T50 with 2 mg ml⁻¹ bovine serum albumin (BSA, Carlroth) was flowed into the channel for immobilization. Five hundred microlitres of 1× T50 was used to wash out unbound DNA, and 1–2 nM JF549-labelled MYC-MAX was flowed in using imaging buffer (20 mM Tris-HCl pH 7.5, 150 mM NaCl, 10% (v/v) glycerol, 0.005% (v/v) Tween-20, 2 mM Trolox, 3.2% (w/v) glucose, 1× glucose oxidase/catalase oxygen scavenging system and 1 mg ml⁻¹ BSA), and movies were recorded at 2–5 Hz in TIRF illumination, alternating between far-red and green illumination (1:200 frames).

Single-molecule trace extraction and trace analysis were done as described previously⁶⁵ with some adjustments. Movies were background-corrected using a rolling ball algorithm in ImageJ. DNA positions were detected using a custom-built MATLAB (Mathworks) script using a local maxima approach. Images were aligned to compensate for stage drift. Fluorescence intensities (in the orange channel) were extracted within a 2-pixel radius of the identified DNA peaks. Individual detections were fitted with a 2D-Gaussian function to determine colocalization with immobilized DNA. Detections exceeding a PSF width of 400 nm, a 250 nm offset from the DNA position or an intensity greater than 5,000 counts were excluded from further analysis. Individual traces were analysed by a step-finding algorithm⁶⁶, followed by thresholding. Overlapping multiple binding events were excluded from the analysis. For each movie, cumulative histograms were constructed from detected bright times (t_bright) corresponding to bound MYC-MAX molecules to obtain dwell times and dark times (t_dark) to obtain on-rate constants, usually including data from around 100 individual traces. The cumulative histograms from traces corresponding to individual DNA were fitted with either di- or tri-exponential functions.

LANCE TR-FRET assays were performed with His-tagged MYC-MAX (acceptor, ULight α-6×His antibody) and donor biotinylated nucleosomes (LANCE Eu-W8044 streptavidin) following the general protocol described previously⁴². To analyse His–MYC-MAX binding to the NCP^SHL+5.1 nucleosomes, biotin was incorporated into H2B (residue T122) using maleimide chemistry (see also the Methods subsection ‘Expression, purification and reconstitution of human octamer histones’). For all other TR-FRET experiments, the biotin was incorporated into the nucleosome using a biotinylated primer proximal to the E-box motif during PCR to produce the DNA fragment (Microsynth). In the MYC-MAX forward titrations, increasing concentrations of His–MYC-MAX (mixed 1:20 with the ULight α-6×His antibody) were added to a mixture of 1 nM biotinylated nucleosome, 2 nM Lance Eu-streptavidin in a buffer containing 20 mM Tris-HCl, pH 7.5, 125 or 75 mM NaCl, 5% glycerol, 0.01% NP-40, 0.01% CHAPS, 5 mM DTT and 100 μg ml⁻¹ BSA (T75). Before TR-FRET measurements, reactions were incubated for 5 min at room temperature. For competition experiments with CLOCK-BMAL1, increasing amounts of untagged CLOCK-BMAL1 bHLH PAS-AB wild-type and mutant proteins were incubated with a preformed complex of His–MYC-MAX-nucleosome (625 nM His–MYC-MAX:31.25 nM ULight) in the T75 buffer. After excitation of europium fluorescence at 337 nm, emissions at 620 nm (europium) and 665 nm (ULight) were measured with a 75-μs delay to reduce background fluorescence and the reactions were followed by recording 30 data points of each well over 30 min using a PHERAstar FS microplate reader (BMG Labtech). The TR-FRET signal of each data point was extracted by calculating the 620:665 nm ratio. The signal was corrected for direct acceptor excitation by subtracting the signal observed in the absence of the nucleosome. The resulting raw signals were fitted to the Bmax values of 1 in Prism 7 (GraphPad), assuming equimolar binding of the TF–nucleosome substrates using a one-site specific binding curve.

For measuring nucleosomes or nucleosome complexes, microscope coverslips were treated with 10 ul of poly-l-lysine for 30 s, rinsed with Milli-Q and dried under an air stream. Before mass photometry measurements, protein dilutions were made in MP buffer (20 mM Tris-HCl pH 7.5, 100 mM KCl and 0.5 mM TCEP) and nucleosome–TF complexes were mixed in a 1:6 ratio and incubated for 30 min at room temperature. Data were acquired on a Refeyn OneMP mass photometer. First, 18 μl of MP buffer was introduced into the flow chamber and focus was determined. Then 2 μl of protein solution were added to the chamber and movies of 60 or 90 s were recorded. Nucleosomes (NCP^SHL+5.8, NCP^SHL−6.2, NCP^POR1 and NCP^{SHL+5.8-tandem}) and CLOCK-BMAL1 bHLH PAS-AB were measured individually at 20 nM (final concentration) and then in complex at 10 and 60 nM, respectively. Each sample was measured at least two times independently (n = 2). All acquired movies were processed and molecular masses were analysed using Refeyn Discover 2.3, based on a standard curve created with BSA and thyroglobulin.

Cy5-labelled nucleosomes (30 nM) were mixed with either CLOCK-BMAL1 bHLH PAS-AB wild type or mutants (0–500 nM), CLOCK-BMAL1 bHLH PAS-AB (250 nM) in the presence and absence of increasing concentrations of cGAS (18.75–150 nM) or cGAS only (75 nM).

For BAF competition assays, unlabelled nucleosomes (30 nM) containing an E-box motif at SHL+5.8 were mixed with BAF only (100 nM), BAF (100 nM) in the presence of increasing amounts of CLOCK-BMAL1 (125 nM, 250 nM and 500 nM) or CLOCK-BMAL1 only (250 nM and 500 nM).The reactions were conducted in binding buffer (BB) (20 mM Tris-HCl pH 7.5, 75 mM NaCl, 10 mM KCl, 1 mM MgCl₂, 0.1 mg ml⁻¹ BSA and 1 mM DTT) and incubated at room temperatute for around one hour. After incubation, the samples were analysed by electrophoresis on a 6% non-denaturing polyacrylamide gel (acrylamide:bis = 37.5:1) in 0.5× TGE buffer (12.5 mM Tris base, 96 mM glycine and 500 μM EDTA), and the bands were visualized with an Odyssey (LiCor) imaging analyser or with a Typhoon FLA 9500 after staining in SYBR GOLD Nucleic Acid Gel Stain (Invitrogen). Fluorescently labelled nucleosomes and DNA-binding curves were analysed using the Empiria Studio v.2.3 software.

DNA sequences were generated by replacing the Widom 601 sequence with the canonical consensus JASPAR E-box motif (GGCACGTGTC, MA0819.1, MA0059.1) at 1-bp intervals across the entire modified W601. The E-box motif present in the original Widom 601 positioning sequence at SHL+5.1 was mutated (see Supplementary Table). The W601-E-box variant DNA sequences were flanked by EcoRV sites and adapter sequences and ordered as gene fragments from TWIST Biosciences. The individual gene fragments were suspended, pooled equally and cut with EcoRV-HF (NEB), and DNA fragments (153 bp) were purified from an agarose gel using the QIAquick Gel Extraction kit (Qiagen). The W601-E-box DNA pool was spiked with an excess of W601 DNA (1:30 molar ratio; pool:601). The nucleosome pool was assembled and purified as described above. 1

SeEN-seq was performed as before²⁵ with some modifications. For SeEN-seq EMSAs, nucleosomes (100 nM) were incubated with a 62.5 nM final concentration of MYC-MAX bHLH LZ (human MYC residues 351–437, human MAX residues 22–102) or 250 nM of CLOCK-BMAL1 bHLH PAS-AB (mouse CLOCK residues 26–395, mouse BMAL1 residues 62–441) in 20-μl reactions containing 20 mM Tris-HCl pH 7.5, 75 mM NaCl, 10 mM KCl, 1 mM MgCl₂, 0.1 mg ml⁻¹ BSA and 1 mM DTT. To compensate for the loss in DNA-binding affinity in the CLOCK-BMAL1 bHLH construct⁶¹, CLOCK-BMAL bHLH SeEN-seq was performed with around fivefold higher concentrations (1,250 nM) compared to what was used for the PAS-containing construct. The reactions were incubated at room temperature for around 1 h and loaded onto a 6% non-denaturing polyacrylamide gel (acrylamide:bis = 37.5:1) in 0.5× TGE gel and run for 1 h (150 V, room temperature). Gels were then stained with a SYBR gold nucleic acid stain (around 10 min, Invitrogen). DNA bands corresponding to the size of TF-bound and unbound nucleosome complexes were imaged and excised using a C300 gel doc UV-transilluminator (Azure Biosystems). Gel slices were incubated with acrylamide gel extraction buffer (100 μl, 500 mM ammonium acetate, 10 mM magnesium acetate, 1 mM EDTA and 0.1% SDS) and heated (50 °C, 30 min). H₂O (50 μl) and the QIAquick Gel Extraction kit QG buffer (450 μl, Qiagen) were added and the samples were heated (50 °C, 30 min). Samples were briefly spun and the supernatant containing DNA fragments were transferred to QIAquick Gel Extraction spin columns. Samples were purified according to the manufacturer’s instructions and eluted in H₂O (22 μl), and the DNA was quantified by Qubit reagent (Thermo Fisher Scientific). Purified DNA (20 μl, around 2–20 ng DNA) was used for NGS library preparation (NEBNext ChIP–seq, E6240S) with dual indexing (E7600S) and no more than 10 cycles of PCR amplification. Purified sequencing libraries were quantified by Qubit reagent (Thermo Fisher Scientific) and the library size was checked on the bioanalyser platform (Agilent) before sequencing on an Illumina MiSeq or NextSeq platform (300 bp paired-end). Sequencing fragments were mapped to the W601 sequence and E-box-motif-containing variants (153 bp) using the Bioconductor package QuasR with default settings⁶⁷, which internally use Bowtie for read mapping⁶⁸. The number of sequence reads aligned to each construct was quantified by the QuasR function Qcount with every construct represented. SeEN-seq enrichments are calculated by determining the fold change between library-size normalized read counts for each 601-E-box variant in the TF-bound and unbound nucleosome fractions. These fold changes represent a relative affinity difference between all positions. In all replicates we were able to capture every motif position, suggesting that the E-box motif does not markedly affect nucleosome stability.

The TF and the nucleosomes were mixed in a 1.5:1 ratio in MS sample buffer (50 mM HEPES pH 7.5, 150 mM NaCl and 500 μM TCEP) and incubated at room temperature for around 1 h. In the meantime, an aliquot of disuccinimidyl sulfoxide (DSSO) XL reagent (Thermo Fisher Scientific, A33545↗) was warmed up to room temperature and diluted to a 100 mM stock concentration in anhydrous DMSO by shaking for 5 min, 400 rpm. After incubation, the sample was transferred to a concentrator (Amicon Ultra, Merck Millipore, 10,000 MWCO), DSSO was added and the cross-linking reaction mix was incubated for 1 h at 10 °C, while shaking at 400 rpm. The excess cross-linker was quenched by adding 1 M Tris pH 6.8 (50 mM final concentration) and incubating for an additional hour at room temperature, 400 rpm. The sample was centrifuged (5 min, 14,000g) to remove XL reagent and 400 µl of fresh 8 M urea in 50 mM HEPES, pH 8.5 for denaturing and washing were added. This step was repeated twice. Next, reduction/alkylation buffer (50 mM TCEP, 100 mM 2-chloroacetamide) was added (5 mM and 10 mM final concentration respectively) and the sample was incubated for 30 min while shaking at 400 rpm. It was centrifuged for 5 min at 14,000g and 400 µl of fresh 8 M urea was added for denaturing and washing. The sample was centrifuged again for 5 min at 14,000g. This step was repeated twice with a final centrifugation step of 15 min instead of 5 min to concentrate the sample to around 30 µl. Lys-C was added (0.2 µg µl⁻¹ stock, 1:100 enzyme to protein ratio) and the sample was digested for 1.5 h at room temperature while shaking. The sample was diluted fourfold with 50 mM HEPES, pH 8.5. Then, trypsin (0.2 mg ml⁻¹ stock, 1:100 enzyme to protein ratio) was added and the sample was incubated overnight at 37 °C, while shaking at 400 rpm. An additional aliquot of trypsin and acetonitrile to a final concentration of 5% was added the next day and the sample was incubated for another 4 h at 37 °C, while shaking at 400 rpm. The sample was transferred into an Eppendorf tube, TFA was added (1% final concentration) and the sample was briefly sonicated and spun down for 5 min at 20,000 g. The supernatant was desalted using a PreOmics iST-NHS kit and concentrated in a speedvac. Samples were reconstituted with 0.1% TFA in 2% acetonitrile.

Samples were analysed by LC–MS in two ways:The equivalent of around 1 μg peptides per sample was loaded onto a uPAC C18 trapping column, and then separated on a 50-cm uPAC C18 HPLC column (connected to an EASY-Spray source (all Thermo Fisher Scientific, columns formerly from Pharmafluidics)) connected to an Orbitrap Fusion Lumos. The following chromatography method was used: 0.1% formic acid (buffer A), 0.1% formic acid in acetonitrile (buffer B), flow rate 500 nl per min, gradient 240 min in total, (mobile phase compositions in % B): 0–5 min 3–7%, 5–195 min 7–22%, 195–225 min 22–80%, 225–240 min 80%.The equivalent of around 5 μg peptides per sample were loaded onto a Vanquish Neo chromatography system with a two-column set-up. Samples were injected with 1% TFA and 2% acetonitrile in H₂O onto a trapping column at a constant pressure of 1,000 bar. Peptides were chromatographically separated at a flow rate of 500 nl per min using a 3-h method, with a linear gradient of 2–9% B in 5 min, followed by 9–28% B in 120 min, followed by 28–100% B in 20 min, and finally washing for 15 min at 100% B (buffer A: 0.1% formic acid; buffer B: 0.1 formic acid in 80% acetonitrile) on a 15-cm EASY-Spray Neo C18 HPLC column mounted on an EASY-Spray source connected to an Orbitrap Eclipse mass spectrometer with FAIMS (all Thermo Fisher Scientific). In either case, the mass spectrometer was operated in MS2_MS3 mode, essentially according to a previous report⁶⁹. On the Orbitrap Fusion Lumos mass spectrometer, peptide MS1 precursor ions were measured in the Orbitrap at 120-k resolution. On the Orbitrap Eclipse, three experiments were defined in the MS method, with three different FAIMS compensation voltages, −50, −60 and −75 V, respectively, to increase the chances of more highly charged peptides (that is, cross-linked peptides) being identified.

For each experiment, peptide MS1 precursor ions were measured in the Orbitrap at 60-k resolution. In either case, the MS advanced peak determination (APD) feature was enabled, and those peptides with assigned charge states between 3 and 8 were subjected to CID–MS2 fragmentation (25% CID collision energy), and fragments detected in the Orbitrap at 30-k resolution. Data-dependent HCD-MS3 scans were performed if a unique mass difference (Δm) of 31.9721 Da was found in the CID–MS2 scans with detection in the ion trap (35% HCD collision energy).

MS raw data were analysed in Proteome Discoverer v.2.5 (Thermo Fisher Scientific) using a Sequest⁷⁰ database search for linear peptides, including cross-linker modifications, and an XlinkX⁶⁹ search to identify cross-linked peptides. MS2 fragment ion spectra not indicative of the DSSO cross-link delta mass were searched with the Sequest search engine against a custom protein database containing the expected protein components, as well as a database built of contaminants commonly identified during in-house analyses, from MaxQuant⁷¹, and cRAP (ftp://ftp.thegpm.org/fasta/cRAP↗), using the target-decoy search strategy⁷². The following variable cross-linker modifications were considered: DSSO hydrolysed/+176.014 Da (K); DSSO Tris/+279.078 Da (K), DSSO alkene fragment/+54.011 Da (K); DSSO sulfenic acid fragment/+103.993 Da (K), as well as oxidation/+15.995 Da (M). Carbamidomethyl/+57.021 Da (C) was set as a static modification. Trypsin was selected as the cleavage reagent, allowing a maximum of two missed cleavage sites, peptide lengths between 4 or 6 and 150, 10 ppm precursor mass tolerance and 0.02 Da fragment mass tolerance. PSM validation was performed using the Percolator node in PD and a target FDR of 1%.

XlinkX v.2.0 was used to perform a database search against a custom protein database containing the expected complex components to identify DSSO-cross-linked peptides and the following variable modification: DSSO hydrolysed/+176.014 Da (K); oxidation/+15.995 Da (M). Cross-link-to-spectrum matches (CSMs) were accepted above an XlinkX score of 40. Cross-links were grouped by sequences and link positions and exported to xiNET⁷³ format to generate cross-link network maps.

Cross-links were mapped to the structure models with an in-house script for PyMOL and the ChimeraX plug-in XMAS⁷⁴. Xwalk was used to calculate solvent accessible surface distances⁷⁵.

Data are available through ProteomeXchange⁷⁶ with the identifier PXD033181.

Nucleosomes were mixed with molar excesses of the respective TFs in a volume of around 100 μl and incubated at room temperature for 30 min (molar ratios: 1:3:3, NCP^SHL+5.1:OCT4:MYC-MAX; 1:1.5, NCP^SHL+5.8:MYC-MAX; 1:1.5, NCP^SHL+5.8:CLOCK-BMAL1; 1:3 NCP^SHL–6.2:CLOCK-BMAL1; 1:3, NCP^SHL+5.1:MAX-MAX; 1:1.5:3, NCP^LIN28-E: MYC-MAX:OCT4; 1:3, NCP^Por1:CLOCK-BMAL1) in a binding buffer containing 20 mM HEPES pH 7.4, 1 mM MgCl₂, 10 mM KCl and 0.5 mM TCEP. The molar ratio used for each considers the number of TF motifs, with an excess of TF, and the relative affinity of each TF for the nucleosome substrate. The sample was then subjected to cross-linking using the GraFix method⁷⁷. For GraFix cross-linking, the TF–NCP complexes were layered on top of a 10%–30% (w/v) sucrose gradient (20 mM HEPES pH 7.4, 50 mM NaCl, 1 mM MgCl₂, 10 mM KCl, 0.5 mM TCEP) with an increasing concentration (0–0.34% w/v) of glutaraldehyde (EMS) and subjected to ultracentrifugation (Beckman SW40Ti rotor, 30,000 rpm, 18 h, 4 °C). After centrifugation, 100-μl fractions were collected from the top of the gradient and peak fractions were analysed by native PAGE. The peak fractions were combined and sucrose was removed by dialysis into Grafix buffer (20 mM HEPES pH 7.4, 50 mM NaCl, 1 mM MgCl₂, 10 mM KCl and 0.5 mM TCEP). The resulting sample was concentrated with an Amicon Ultra 0.5-ml centrifugal filter to around 2–7 μM nucleosomes as determined by measuring the DNA concentration at an absorbance of 260 nm. After concentration, 3.5 μl of sample was applied to Quantifoil holey carbon grids (R 1.2/1.3 200-mesh, Quantifoil Micro Tools). Glow discharging was performed in a Solarus plasma cleaner (Gatan) for 15 s in a H₂/O₂ environment. Grids were blotted for 3 s at 4 °C at 100% humidity in a Vitrobot Mark IV (FEI), and then immediately plunged into liquid ethane.

Data were collected automatically with EPU 3.0 (Thermo Fisher Scientific) on a Cs-corrected (CEOS) Titan Krios (Thermo Fisher Scientific) electron microscope operated at 300 kV or on a Glacios (Thermo Fisher Scientific) electron microscope at 200 kV (NCP^SHL+5.1-MAX-MAX and NCP^Por-CLOCK-BMAL1 only). For the OCT4–MYC-MAX-bound nucleosome structure, zero-energy-loss micrographs were recorded at a nominal magnification of 130,000× using a Gatan K2 summit direct electron detector (Gatan) in counting mode located after a BioQuantum-LS energy filter (slit width of 20 eV). For the other assemblies the acquisition was performed at a nominal magnification of 75,000–96,000× with a Falcon 4 direct electron detector (Thermo Fisher Scientific). All datasets were recorded with an accumulated total dose of 50 e^–/Ų and the exposures were fractionated into 50 frames. The targeted defocus values ranged from −0.25 to −2.5 μm.

Real-time evaluation along with acquisition with EPU 3.0 (Thermo Fisher Scientific) was performed with CryoFLARE1.10 (ref. ⁷⁸). Drift correction was performed with the RELION 3 motioncorr implementation⁷⁹, in which a motion-corrected sum of all frames was generated with and without applying a dose-weighting scheme. The CTF was fitted using GCTF 1.06 (ref. ⁸⁰) or the patch CTF implementation in cryoSPARC v.3. Particles were picked using crYOLO (1.8.0)⁸¹, cisTEM (1.0.0 beta)⁸², AutoPick (implemented in RELION)⁸³ or cryoSPARC v.3 blob picker⁸⁴.

All datasets were further processed in RELION 3.0 (ref. ⁷⁹), cryoSPARC v.3 or cryoSPARC v.4 in the case of the NCP^Por structure⁸⁴ as indicated in each Extended Data figure including two-dimensional (2D) and 3D classification, 3D refinement, particle polishing and CTF refinement. The resolution values reported for all reconstructions are based on the gold-standard Fourier shell correlation curve (FSC) at 0.143 criterion^83,85 and all the related FSC curves are corrected for the effects of soft masks using high-resolution noise substitution⁸⁶. The software used for the final refinements of each map is indicated in the corresponding Extended Data figure. For the NCP^SHL–6.2-CLOCK-BMAL1 map, a composite map of two refinements was generated using combine_focus_maps implementation in PHENIX⁸⁷. LocScale implemented in CCPEM (v.1.5)^88,89 was used for sharpening and blurring the following maps: NCP^SHL+5.8-CLOCK-BMAL1, NCP^SHL+5.8-MYC-MAX and NCP^SHL+5.1-MYC-MAX-OCT4. The NCP^SHL–6.2-CLOCK-BMAL1 maps were filtered based on local resolution using cryoSPARC v.3. All local resolutions were estimated with MonoRes (XMIPP) implementation in cryoSPARC v.3 (ref. ⁹⁰).

For modelling of MYC-MAX bound to the NCP in the presence of OCT4, PDB 6T90↗ (ref. ²⁵) was used as a template for the OCT4-bound NCP, and coordinates extracted from PDB 1NKP↗ (ref. ²) were used to obtain a template for DNA-bound MYC-MAX. The two models were fitted into the cryo-EM map using ChimeraX (fit-in-map tool; ref. ⁵⁶). The gap between NCP DNA and MYC-MAX DNA was closed using ideal B-form DNA in Coot (v.0.9.6)⁹¹ and the DNA sequence was adapted accordingly. The joined DNA was refined in PHENIX⁹² using DNA restraints (base pair, stacking). MYC-MAX together with the detached DNA end as well as OCT4 together with the other DNA end were further relaxed into the density using ChimeraX/ISOLDE⁹³ in combination with adaptive distance restraints. Side chains were corrected in Coot and ChimeraX/ISOLDE (v.1.2–v.1.5) if necessary. The model coordinates and B-factors were refined using the Rosetta FastRelax and B-factor protocols (v.3.13)⁹⁴ in combination with self-restraints (torsions) and with side-chain repacking disabled. The model for MYC-MAX bound to SHL+5.8 was obtained by docking the NCP template (PDB: 6T93↗)²⁵ into the map and fitting the DNA end with ISOLDE (in combination with adaptive distance restraints). The DNA sequence was adjusted and the MYC-MAX model (PDB: 1NKP↗; ref. ⁵⁸) was docked by superposition on the E-box motif. The model was further refined with ISOLDE using adaptive distance restraints for different rigid groups (MYC-MAX in combination with released DNA, histones) as well as PHENIX (v.1.19–v.1.20.1) and Rosetta as described above. Putative side-chain density did not allow unambiguous differentiation between MYC-MAX in the quasi-homodimeric overall structure. Therefore, both orientations (MYC-MAX dimer flipped in respect to the nucleosome) were modelled with 50% occupancy, respectively, and side chains were truncated.

In the case of both NCP-bound CLOCK-BMAL1 models, PDB 6T93↗ (ref. ²⁵) was used as the NCP template, PDB 4H10↗ (ref. ⁶¹) as the template for the DNA-bound bHLH domains of CLOCK-BMAL1, and PDB 4F3L↗ (ref. ³) as the template for the CLOCK-BMAL1 PAS domains. The DNA sequence of the NCP template (6T93↗) was extended at both ends with ideal B-form DNA generated in Coot and the sequence was adjusted to the construct used in this study. The NCP model was fitted into the cryo-EM density with ChimeraX (fit-in-map tool)⁵⁶ and the detached DNA ends were semi-flexibly fitted into the density with ISOLDE⁹³ in combination with adaptive distance restraints. The DNA was refined with PHENIX⁹² and Rosetta⁹⁴ as described for the MYC-MAX structure. The PAS domains from 4F3L↗ were docked and rigid-body-refined with phenix.dock_in_map. Again, adaptive distance restraints were generated in ISOLDE for separate groups including the bHLH domains together with the detached DNA segment, the opposite DNA end and the PAS domains. This allowed the groups to be semi-flexibly relaxed into the density while maintaining the original geometry.

In the case of CLOCK-BMAL1 bound to position SHL–6.2, the DNA/bHLH model (4H10↗) and the NCP template (6T93↗) were fitted into the density and the DNAs were connected with an ideal B-form DNA generated in Coot. The DNA sequence was adapted to the position SHL–6.2 construct and refined as described for the NCP-bound MYC-MAX structure. The PAS domains from 4F3L↗ were manually docked into the density guided by the cross-link between BMAL1 K212 and H3 K57. Because accurate fitting was not possible owing to local resolution limitations and diffuse map density, the PAS domains were docked against the histones using the Rosetta local docking protocol⁹⁵ in combination with Rosetta density scoring (8°, 3 Å perturbations) and a filter for a maximum cross-link distance of 30 Å between Cα atoms of BMAL1 K212 and H3 K57. The resulting poses were ranked by interface energy and density scores and the pose with the best interface energy score was selected because it was separated from the bulk of other poses while also having a good density score. B-factors were refined as described above. Because of insufficient local resolution, side chains were removed from the CLOCK-BMAL1 models for deposition.

In the case of CLOCK-BMAL1 bound to Por, the E-box 1 protomer and the bHLH domain of the E-box 2 protomer were resolved to a resolution facilitating model building. The model from the SHL+5.8 structure was used as a template and readily fit the density of the nucleosome and the internal CLOCK-BMAL1 heterodimer. The DNA sequence was adjusted and the external-bound CLOCK-BMAL1 heterodimer was docked in ChimeraX on the basis of cross-linking data, map fit and orientations of the connecting segments of the PAS domains in respect to the bHLH domains. The model was subjected to semi-flexible fitting with ISOLDE using distance and torsion restraints and further refined with PHENIX using coordinate restraints. Observed inter-CLOCK-BMAL1 cross-links can occur either within a heterodimer or between the heterodimers. Some cross-links would be sterically implausible to occur within the heterodimer and could reflect potential inter-heterodimer cross-links. Together with a histone cross-link (external CLOCK K205 and H3 K56) these putative inter-heterodimer cross-links suggest an overall orientation in which the external CLOCK PAS domains face the internal BMAL1 PAS domains. It was not possible to find a consensus model in which all cross-link distances would be below a threshold of 30 Å. This could be due to the assignment ambiguity of the inter-CLOCK-BMAL1 cross-links or the flexibility of the PAS domains. Because of these ambiguities and the limited local map resolution, the external PAS domains are not included in the final model. B-factors were refined as described above. Because of the insufficient local resolution, side chains were removed from the CLOCK-BMAL1 and histone models for deposition.

The Rosetta cryo-EM refinement protocols were run using an in-house developed pipeline (ROSEM, https://github.com/fmi-basel/RosEM↗). Validation for all models was carried out with PHENIX⁹⁶ and MolProbity (v.4.5.2)⁹⁷.

Structural figures and cryo-EM segmented maps were produced with UCSF ChimeraX (v.1.3).

Clash scores for MYC-MAX–nucleosome and CLOCK-BMAL1–nucleosome models were calculated using a PyMOL script (scanFactor.py) as described previously^45,98 In brief, a MYC-MAX probe (1NKP↗) or a CLOCK-BMAL1 probe (4F3L↗, 4H10↗) containing an appropriately positioned DNA fragment for superimposing on a nucleosome template model was placed in all possible binding positions, and the clash score for each taken as the total number of atoms in the TF closer than an adjustable threshold distance (1 Å default) to nucleosome atoms.

NCPs reconstituted with Widom 601 DNA containing an E-box motif, at SHL −6.9 and SHL +5.1 and an OCT4 motif at SHL −6.0 were mixed with full-length human OCT4 and/or human MYC-MAX bHLH LZ (human MYC residues 351–437, human MAX residues 22–102) in a 1:2:2 molar ratio in BB buffer (20 mM HEPES pH 7.4, 1 mM MgCl₂, 10 mM KCl and 0.5 mM TCEP) and incubated on ice for around 30 min. Nucleosomes in the presence or absence of OCT4 and/or MYC-MAX were treated with a titration (0.1 U, 0.5 U) of DNaseI (NEB M0303S) in the presence of MgCl₂ (2.5 mM) and CaCl₂ (0.5 mM) for 5 min at 37 °C. The reaction was stopped by adding an equal volume of Stop Buffer (200 mM NaCl, 30 mM EDTA, 1% SDS) and incubated on ice for 10 min. Samples were treated with Proteinase K (10 μg) for 2 h and DNA was retrieved using Ampure Beads (A63881). DNA was used for sequence library preparation (NEBNext ChIP–seq, E6240S) with dual indexing, and sequenced on an Illumina MiSeq (300 bp paired-end). Sequences were mapped to the Widom 601 sequence (147 bp) containing the TF motifs using the Bioconductor package QuasR with default settings⁶⁷, which internally use Bowtie for read mapping⁶⁸. The start position of mapped reads, the DNaseI cut site, was extracted and the counts were binned into 1-bp bins across the length of the W601 sequence. Plots and comparisons were done using 100,000 reads per replicate.

One microgram of genomic DNA extracted from D. melanogaster BG-3 cells was assembled into chromatin by adding 15 µl 10× McNAP buffer (0.3 M creatine phosphate, 30 mM ATP, 3 mM MgCl₂, 1 mM DTT and 10 ng µl⁻¹ creatine phosphokinase), 35 µl EX50 buffer (10 mM HEPES/KOH pH 7.6, 50 mM KCI, 1.5 mM MgCl₂, 50 µM ZnCl, 10% glycerol, 1 mM DTT, 1× Proteinase Inhibitor Complex and 100 µl Drosophila preblastoderm embryo extract (DREX, prepared as described previously³⁷). Assembly proceeded for 5 h at 26 °C at 300 rpm on a shaking heat block. Then, 250 nM of Spy-tagged proteins were added and allowed to bind for 1 h. Samples were cross-linked with formaldehyde (0.1% final concentration) for 10 min and then quenched by addition of 125 mM glycine. Samples were partially digested by 200 U of micrococcal nuclease (MNase, Sigma) for 2 min. Digestion was stopped by addition of 25 mM EDTA. For immunoprecipitation, samples were precleared on a rotating wheel with 20 µl protein AG beads per 1 µg chromatin for 1 h at 4 °C. Two µl of hIgG1-FcSpyCatcher3 (BioRad TZC009) was added and the reaction was incubated on a rotating wheel at room temperature for 1 h. Then, freshly washed protein AG beads (Helmholtz Centre Munich, monoclonal facility) were added and the incubation continued overnight at 4 °C. The beads were washed 4 times for 5 min with 1 ml of 1× RIPA buffer (1 µg chromatin on 20 µl beads). The beads then were suspended in 100 µl 1× TE buffer and digested with 10 µg RNAse A (Sigma) for 30 min at 37 °C. Then, 100 µg Proteinase K (Qiagen) was added and samples were digested and de-cross-linked overnight at 65 °C while shaking. Beads were pelleted at 1,000g for 1 min and the supernatant was transferred to a fresh tube. DNA was purified by two extractions with phenol:chloroform:isoamyl-alcohol (25:24:1, Sigma Aldrich) precipitation and a 70% ethanol wash and dissolved in 10 mM Tris/NaCl, pH 8. Concentrations were determined using Qubit (Thermo Fisher Scientific).

NGS libraries were prepared using the NEBNext Ultra II DNA Library (New England Biolabs) according to the manufacturer’s instructions and sequenced on an Illumina NextSeq1000 sequencer. About 20 million paired-end reads were sequenced per sample for each of the ChIP replicates. Replicates were performed using a separate batch of purified proteins and DREX extracts. Base calling was performed by Illumina’s RTA software, v.1.18.66.3.

Experiments involving mouse tissue collection were approved by the Texas A&M University Institutional Animal Care and Use Committee. Adult male mice were maintained at a constant temperature of 22–23 °C and relative humidity of 50–60%, with a 12-h light:12-h dark cycle. Wild-type (Charles River strain 027) and Bmal1^−/− (BMKO; Jackson Laboratory strain 009100) mice were both in a C57BL/6Crl background and were euthanized in the middle of the day at ZT6 by isoflurane anaesthesia followed by decapitation. Livers were collected, briefly washed in ice-cold 1× PBS, snap-frozen in liquid nitrogen and stored at −80 °C until further use. Nuclei were extracted as described previously¹⁰⁵. In brief, frozen mouse liver was grained into powder under liquid nitrogen in a mortar and homogenized in 4 ml of ice-cold 1× PBS. Liver homogenate was mixed with 25 ml of ice-cold sucrose homogenate solution (2.2 M sucrose, 10 mM HEPES pH 7.6, 15 mM KCl, 2 mM EDTA, 1 mM PMSF, 0.15 mM spermine, 0.5 mM spermidine and 0.5 mM DTT). After incubation on ice for 10 min, the liver homogenate sucrose solution was carefully poured on the top of a sucrose cushion solution (2.05 M sucrose, 10% glycerol, 10 mM HEPES pH 7.6, 15 mM KCl, 2 mM EDTA, 1 mM PMSF, 0.15 mM spermine, 0.5 spermidine and 0.5 mM DTT) and centrifuged for 45 min at 24,000 rpm (100,000g) at 4 °C using a Beckman SW32Ti rotor. Nuclei were resuspended in SMF wash buffer (10 mM Tris pH 7.5, 10 mM NaCl, 2 mM MgCl₂ and 0.1 mM EDTA) and washed once with the same buffer.

The SMF protocol was adapted from ref. ¹⁰⁶ and optimized for mouse liver. For each sample, 250,000 nuclei were washed once with M.CviPI wash buffer (50 mM Tris pH 8.5, 50 mM NaCl and 10 mM DTT) and resuspended in 1 mL of 1× M.CviPI reaction buffer (50 mM Tris pH 8.5, 50 mM NaCl, 300 mM sucrose and 10 mM DTT). Then, 18.75 µl of 32 mM SAM and 200 U of M.CviPI (NEB-M0227L; 50 µl) were added, and the reaction was incubated at 37 °C for 7.5 min in a water bath. The reaction was supplemented with 100 U of M.CviPI (25 µl) and 128 µmol of SAM (4 µl) for a second incubation round of 7.5 min at 37 °C. The methylation reaction was stopped by adding 350 µl of SDS-containing buffer (20 mM Tris, 600 mM NaCl and 1% SDS 10 mM EDTA) and 20 µl of Proteinase K (20 mg ml⁻¹), and the mixture was incubated overnight at 55 °C. Genomic DNA was isolated by phenol-chloroform purification and isopropanol precipitation, resuspended in 10 mM Tris pH 7.5 and treated with RNAse A at for 1 h at 37 °C. Two micrograms of genomic DNA were used for bisulfite conversion using the Epitect bisulfite conversion kit (QIAGEN 59124). Ten to twelve nanograms of bisulfite-converted DNA were used to amplify a distal enhancer of the gene Por (chr. 5:135,674,788–135,675,224; Mus musculus mm10 genome version), using the KAPA HiFi Uracil+ kit (Roche) as in ref. ¹⁰⁶ (forward primer: GGTTTTTTGAGYATAGAATTTTTTTTTT; reverse primer: CCATCTTCTCTCACTTCTRCCCAAT). PCR products were purified with 1.5× SPRI beads, and around 20 ng was used to generate sequencing libraries using the NEBNext Ultra II Kit. Libraries from three biological replicates of wild-type ZT6 and three biological replicates of BMKO ZT6 were pooled together and sequenced with a MiSeq v.2 Nano Reagent kit (paired-end 250 bp).

The PairwiseAligner function in the Bio.Align Python package was used for sequence alignment. The matched, mismatched and gapped alignment conditions were given a score of 1.0, −0.2 and −0.5, respectively. The sum of the alignment score at each position divided by the total alignment length was defined as the final alignment score. Sequences in the paired-end fastq files were pre-selected by aligning the first around 25-nt query sequences to both forward and reverse primer sequences. Reads with a primer final alignment score higher than 0.8 were selected, and full-length paired-end query sequences were aligned to bisulfite-converted target sequence (HCH replaced by HTH, GC replaced by GY, and CG replaced by YG, with Y = pyrimidine, and H = not G). Paired-end sequences with a final alignment score higher than 0.7 were selected to reconstitute the full-length enhancer sequence based on the alignment result (in the overlapping region, nucleotides having a higher quality score were used). Next, PCR duplicates were removed, and an equal number of reads were randomly selected in each sample for downstream analysis (n = 1,052 reads per sample to match that of the sample with the lowest amount of unique reads). The methylation information at cytosines of all GCH positions (GpC positions that are not followed by a G, to avoid conflicts with endogenous CpG methylation) was extracted, using 0 or 1 to represent unprotected or protected cytosines, respectively. Reads from all six samples were then clustered using the Binary Matrix Decomposition clustering algorithm¹⁰⁷, and then parsed according to their relative cluster and genotype. Raw data (fastq) reads are available at Mendeley Data: 10.17632/t7xj4rc62t.1.

Wild-type mouse Bmal1 or mutants (Uniprot: Q9WTL8↗) were cloned into the mammalian lentiviral expression backbone (Addgene plasmid, 73320) with a modification to include a stop codon in-frame with the EGFP to prevent expression of the fusion protein (TWIST Biosciences). Recombinant lentiviral particles were produced in HEK293T cells (ATCC) using Pax2 and pMD2.5 packaging plasmids. The resulting supernatant was used to transduce Bmal1^−/− PER2::LUC fibroblasts as previously¹⁰⁸. For selection, 1 μg ml⁻¹ puromycin was applied for one week with medium changes every 48 h.

Successfully transduced cells were grown to confluence in 12-well dishes in high-glucose (27.8 mM), glutamax-containing DMEM (GIBCO) supplemented with 10% serum (HyClone FetalClone III, Thermo Fisher Scientific) and penicillin–streptomycin. Reconstituted lines also had 0.5 μg ml⁻¹ puromycin to maintain selection. Confluent cultures were kept for up to 4 weeks with the medium refreshed every 7–10 days. Before the start of recording, cells were synchronized by the addition of 100 nM dexamethasome for 1 h and then changed to MOPS-buffered ‘air medium’ (bicarbonate-free DMEM, 5 mg ml⁻¹ glucose, 0.35 mg ml⁻¹ sodium bicarbonate, 0.02 M MOPS, 100 μg ml⁻¹ penicillin–streptomycin, 1% Glutamax, 1 mM luciferin, pH 7.4, 325 mOsm (ref. ¹⁰⁹). Cells were then transferred to an Alligator system (Cairn Research), in which bioluminescent activity was recorded at 15-min intervals using an electron multiplying charge-coupled device (EM-CCD) at constant 37 °C.

Bioluminescent traces of cells were fitted with damped cosine waves using the following equation:\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$y=mx+c+{\rm{Amplitude}}\cdot e-kx\cdot \cos (2\pi (x-{\rm{phase}}){\rm{period}})$$\end{document}y=mx+c+Amplitude⋅e−kx⋅cos(2π(x−phase)period)where y is the signal, m is the gradient of the detrending line, c is the y intercept of this detrending line, x is the corresponding time, amplitude is the height of the peak of the waveform above the trend line, k is the decay constant (such that 1/k is the half-life), phase is the shift relative to a cos wave and the period is the time taken for a complete cycle to occur.

Samples were run on AnyKD Mini-PROTEAN TGX gels (BioRad) using the manufacturer’s protocol with a Tris-Glycine SDS buffer system. Protein transfer to nitrocellulose was performed using the Trans-Blot Turbo Transfer system (BioRad), with a standard or high-molecular weight protocol as appropriate. Nitrocellulose was washed briefly, and then blocked for 30 mins at room temperature in 5% w/w non-fat dried milk (Marvel) in Tris-buffered saline/0.05% Tween-20 (TBST). Membranes were then incubated, rocking, with 1:4,000 primary antibody (M2 anti-Flag, Sigma F3165) to detect CLOCK-BMAL1 and anti-GAPDH (Santa Cruz Biotechnologies sc-365062) was used as a loading control at a dilution of 1:3,000 in blocking buffer (5% milk, TBST) overnight at 4 °C. The following day, the membrane was washed for a further 3 × 10 min in TBST and incubated again for one hour with anti-mouse HRP secondary antibody (Sigma, A9917, 1:5,000). A further 3 × 10-min washes in TBST were performed before chemiluminescence detection using Immobilon reagent (Millipore), which was imaged using a ChemiDoc XRS+ imager (BioRad). Quantification was performed using Image Lab Software 6.0 (BioRad).

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

Supplementary InformationThis file contains Supplementary Table 1 (DNA sequences and primers used in this study) and Supplementary Figure 1 (Raw gels).Reporting SummarySupplementary Table 2Cross-linking mass spectrometry data for MYC-MAX and CLOCK-BMAL1–nucleosome complexes.Supplementary Table 3Processed DNA sequencing reads for single-molecule footprinting of the Por enhancer locus in the mouse liver.