Expression and purification of DNMT3A1 and JARID2 proteinsThe gene fragment encoding residues 126–223 of human DNMT3A1 (a.k.a. DNMT3A1 UDR herein) was PCR amplified and inserted into a modified pRSFDuet-1 vector, preceded by a His6-SUMO tag and Ubl-specific protease (ULP) 1 cleavage site. The His6-SUMO-DNMT3A1 UDR fusion protein was expressed in Escherichia coli BL21 DE3 (RIL) cells, followed by purification using a Ni2+-NTA column. Subsequently, the His6-SUMO tag was removed by ULP1 cleavage and the DNMT3A1 UDR protein was sequentially purified via ion-exchange chromatography on a Heparin column (GE Healthcare), nickel affinity chromatography on a Ni2+-NTA column and size-exclusion chromatography on a HiLoad 16/600 Superdex-75 pg column pre-equilibrated with buffer containing 20 mM HEPES (pH 7.5), 50 mM NaCl and 5 mM dithiothreitol (DTT). The purified DNMT3A1 UDR sample was concentrated and flash-frozen in liquid nitrogen and stored at −80 °C before use. DNMT3A1 UDR mutants were generated using site-directed mutagenesis and purified in the same way as that for WT DNMT3A1 UDR.For the JARID2−H2AK119ub1 binding assay, the DNA fragment encoding residues 1–60 of human JARID2 (referred to hereafter as JARID21-60) was synthesized by Integrated DNA Technologies and inserted into pRSFDuet-1 (for production of tag-free JARID21-60) or pGEX-6P-1 (for production of GST-JARID21-60) vector. Tag-free JARID21-60 was purified using the same method as DNMT3A1 UDR, except that no ion-exchange chromatography was involved and the buffer for size-exclusion chromatography contained 500 mM instead of 50 mM NaCl. GST-tagged JARID21-60 was purified via GST affinity chromatography, ion exchange and size-exclusion chromatography. The purified GST-tagged JARID21-60 sample was stored in buffer containing 20 mM HEPES (pH 7.5), 500 mM NaCl and 5 mM DTT at −80 °C before use. The free GST tag used for the BLI assay was obtained from the purified GST-JARID21-60 after cleavage by preScission protease, a second GST affinity chromatography, and size-exclusion chromatography on a HiLoad 16/600 Superdex-75 pg column pre-equilibrated with buffer containing 50 mM Tris-HCl (pH 8.0) and 300 mM NaCl. The protein sample was concentrated and stored at −80 °C before use.For DNA methylation assays, full-length DNMT3A1 and DNMT3L were each cloned into an in-house MBP-tagged vector and expressed in E. coli BL21 DE3 (RIL) cells. The DNMT3A1- and DNMT3L-expressed cells were then co-lysed and the DNMT3A1-DNMT3L complex was purified using the same method as DNMT3A1 UDR, except that the MBP tag was removed by Tobacco Etch Virus (TEV) protease, followed by size-exclusion chromatography on a Superdex 200 increase 10/300 GL column in buffer containing 20 mM HEPES (pH 7.5), 5 mM DTT, and 250 mM NaCl. Full-length DNMT3A1 mutants were generated using site-directed mutagenesis and produced in the same approach as that for WT protein.Expression and purification of histone proteins and H2AK119ub1 modificationPreparation of Xenopus H3, H4, H2A and H2B followed previously published protocols6,71 except that H2A and H2B were cloned in tandem into the pRSFDuet-1 vector and expressed in a soluble form using E. coli BL21 DE3 (RIL) cells. The H2A and H2B proteins were then co-purified using a Ni2+-NTA column, followed by ULP1 treatment for removal of His6-SUMO tag and ion-exchange chromatography on a Heparin column (GE Healthcare). His6-SUMO-tagged ubiquitin G76C was inserted into pRSFDuet-1 vector, expressed in E. coli BL21 DE3 (RIL) cells and purified using Ni2+-NTA column. His6-SUMO-tagged ubiquitin I36A/G76C or L71A/G76C mutant was generated using site-directed mutagenesis and purified in the same way as that for His6-SUMO-tagged ubiquitin G76C. The H2AK119ub1 modification was generated through conjugating G76C-mutated ubiquitin with histone H2AK119C via dichloroacetone crosslinker. In essence, His6-SUMO-tagged ubiquitin G76C, I36A/G76C or L71A/G76C mutant and H2AK119C-H2B were mixed in 6:1 molar ratio in 20 mM Tris-HCl (pH 7.5), 600 mM NaCl, 5% Glycerol and 5 mM tris (2-carboxyethyl) phosphine (TCEP), followed by addition of crosslinker 1,3-dichloroacetone with the amount equal to one-half of the total sulfhydryl groups. The reaction continued overnight, before being quenched by 5 mM β-mercaptoethanol (ME), which was then removed via dialysis against buffer containing 20 mM Tris-HCl (pH 7.5), 600 mM NaCl and 5% Glycerol. Next, the H2AK119C-ubiquitin conjugate (denoted as H2AK119ub1 herein)-H2B sample was sequentially purified through Ni2+-NTA affinity chromatography, ion-exchange chromatography on a Heparin column (GE Healthcare), removal of His6-SUMO tag via ULP1 cleavage, and a second round of Ni2+-NTA affinity chromatography. The purified H2AK119ub1-H2B protein sample was concentrated and stored at −80 °C before use.Reconstitution of nucleosome core particles (NCP)The 601 nucleosomal DNA was generated by PCR amplification (forward primer F1: 5′-GCTCTCTACGTAAACATCCTGGAGAATCCCGGTGC-3′; reverse primer R1: 5′-CGAAGTGGGTAAGTCACAGGATGTATATATCTGACACG-3′) of the 147-bp 601 DNA72 and purified using a Mono-Q column (GE Healthcare). For Biolayer Interferometry (BLI) assay, the 601 DNA was also generated by PCR using the forward primer containing a biotin tag at the 5′ end (forward primer F2: 5′-biotin-CTCTCTCCGTAAACATGCTGGAGA-3′; reverse primer R1). For in vitro DNA methylation assay, the nucelosomal DNA was generated by PCR using the forward primer containing multiple CpG sites (forward primer F3: 5′-GCATGCGCCGTCGTTAAGCGCCCCGTGTCGAGAATCCCGGTGCCGAGG-3′; reverse primer R1).The histone H2A-H2B tetramer, H3 and H4 were mixed in a ratio of 1.1:1:1 and dialyzed against buffer containing 20 mM Tris-HCl (pH 7.5), 7 M Urea at 4 °C overnight, followed by dialysis against buffer I containing 20 mM Tris-HCl (pH 7.5), 7 M Guanidine-HCl and 10 mM DTT for 3 hr, buffer II containing 20 mM Tris-HCl (pH 7.5), 2 M NaCl and 5 mM β-ME for 4 hr, and buffer II with additional 1 mM EDTA overnight. After concentration, the histone octamer was further purified using a HiLoad 16/600 Superdex-200 pg column (GE Healthcare). The purified histone octamer sample was concentrated and stored at −80 °C before use.The histone octamer and nucleosomal DNA were mixed in a molar ratio of 0.9:1, and further mixed with 4 M KCl in equal volume. A salt gradient method was used for histone octamer reconstitution with nucleosome DNA. In essence, the samples were first dialyzed against the refolding buffer containing 10 mM Tris-HCl (pH 7.5), 2 M KCl, 1 mM DTT, and 1 mM EDTA. The concentration of KCl was then gradually decreased from 2 M to 250 mM over a period of 18 hr. Finally, the samples were dialyzed towards buffer containing 10 mM HEPES (pH 7.5) and 1 mM DTT at 4 °C overnight. The reconstituted nucleosomes were kept at 4 °C before use.Assembly of the complex between DNMT3A1 UDR and H2AK119ub1-modified nucleosomeThe H2AK119ub1-modified nucleosome and DNMT3A1 UDR were mixed at a molar ratio of 1:8, and incubated on ice for 30 min. The sample was subject to chemical crosslinking via GraFix method73. An amount of 200 pmol of sample was loaded onto the top of a 10–30% linear glycerol gradient in a buffer (50 mM HEPES, pH 7.5, 50 mM NaCl, 30% Glycerol) with a gradient of 0–0.05% (v/v) of crosslinker glutaraldehyde in each ultracentrifuge tube for GraFix, followed by centrifugation at 40,000 × g for 18 hr using a Beckman SW41 rotor at 4 °C. Then the samples were separated from top to bottom at a volume of 500 µL each for 25 fractions. The fractions were examined by 6% native PAGE gel and negative-staining electron microscopy (EM). Fractions with good homogeneity and reasonable particle size were pooled and buffer exchanged to 10 mM Tris-HCl (pH 7.5), 50 mM NaCl, 5% Glycerol and 1 mM DTT using a PD10 column (Cytiva). The samples were concentrated and stored on ice before use.Cryo-EM sample preparation and data collectionThe complex of DNMT3A1 UDR with H2AK119ub1-modified NCP was adjusted to a concentration of 3.5 µM. A volume of 4 µL of the complex sample was applied on Copper 300 mesh grids (Quantifoil R 1.2/1.3 100 holey carbon films) after fresh glow discharge at 20 mA for 1 min. Subsequently, the sample on grids was vitrified and plunge-frozen in liquid ethane using Vitrobot (thermos scientific) with the chamber equilibrated at 100% humidity, 6.5 °C. The blot time, wait time and drain time were 4 s, 10 s and 0 s respectively and a blot force of 1 was used.The cryo-EM data collection for the complex between DNMT3A1 UDR and H2AK119ub1-modified NCP was performed on a Titan Krios microscope operated at 300 kV at Nation Center for CryoEM Access and Training (NCCAT). Micrographs were collected in counting mode with an energy filter (slit width of 10–20 eV) at an apoF-calibrated pixel size of 0.926 Å. The exposures were recorded with an accumulated total dose of 50 e/Å2 over 40 frames and a defocus range of −0.8 µm to –2.5 µm.Cryo-EM image processingThe cryo-EM data were processed using cryoSPARC (v4.0.1)74. The multi-frame movies were motion-corrected and dose-weighted using the patched motion correction module. Initial particle picking was performed using the TOPAZ method75, resulting in a total of 3.5 million particles extracted from 9467 micrographs, with a downscaled pixel size of 2.6 Å. After several rounds of 2D classifications, 1.60 million particles from classes with discernible features of nucleosomal DNA and histone octamer were selected for further analysis. The initial models were created using 3D ab initio reconstruction and subject to heterogeneous refinement with C1 symmetry being applied. The particles associated with the class exhibiting the most clearly defined DNMT3A segment and H2AK119ub1 were re-extracted with a pixel size of 1.23 Å, followed by removal of duplicates and non-uniform refinement. Using a mask protecting density around the ubiquitin and adjacent DNA segment, particle subtraction and subsequent alignment-free 3D classification were performed using Relion 4.176. Two classes with relatively strong ubiquitin density were combined and subject to another round of focused classification. Finally, 86,685 particles from the class with traceable density for both ubiquitin and DNMT3A segment were selected for final non-uniform refinement, local CTF correction and focused refinement of the ubiquitin moiety. A composite map was generated by Phenix software package (v1.20.1)77 using combined densities from the consensus map and the local refinement map for ubiquitin.Model building and refinementFor model building, coordinates extracted from PDB entry 6WKR were used as a template for nucleosome and coordinates of a ubiquitin molecule predicted by AlphaFold78 were used as a template for ubiquitin. The templates of nucleosome and ubiquitin were then fit into the overall and local refinement cryo-EM maps, respectively, using ChimeraX (v1.5)79. Subsequent model building of all components was performed using Coot (v0.9.6)80, followed by real-space refinement over the composite map using Phenix. The reported resolution was based on the gold-standard Fourier shell correlation curve (FSC) at 0.143 criterion.Biolayer Interferometry (BLI) assayBinding affinities for DNMT3A, WT or mutants, GST tag, and GST-JARID21-60 with nucleosomes were measured on a Gator BLI platform instrument using streptavidin (SA) XT biosensors. All steps were performed at 30 °C, 1000 rpm. Reagents were formulated in the buffer containing 20 mM Tris-HCl (pH 7.5), 50 mM NaCl, 1 mM DTT and 0.01% Tween-20. The biosensors were equilibrated in the buffer for 10 min. Biotin-labeled nucleosomes were immobilized on the biosensors, followed by washing for 2 min. Then the nucleosome-loaded biosensors were submerged in solutions of DNMT3A WT or mutants, GST tag, or GST-JARID21-60, for 3 min and transferred to the buffer for 6 min to measure the association and dissociation kinetics. Data was analyzed on GatorOne software (v2.10.4) and results were plotted in GraphPad Prism v6.01.In vitro DNA methylation assayIn vitro DNA methylation assays were carried out for full-length DNMT3A1 (WT or mutant)–DNMT3L complex on H2AK119ub1-modified NCP with one of the linker DNAs containing multiple CpG sites, as described above. A 20-μL reaction mixture contained 0.2 µM H2AK119ub1-modified NCP, 0.1 µM DNMT3A1–DNMT3L tetramer, 2 μM S-adenosyl-L-[methyl-3H] (3H-SAM) in reaction buffer containing 50 mM Tris-HCl (pH 8.0), 100 mM NaCl, 0.05% β-ME, 5% glycerol and 200 μg/mL BSA. The reaction mixtures were incubated at 37 °C for 0, 15 or 30 min before being quenched. Ten µL of each reaction mixture was loaded onto the positively charged Nylon transfer membrane and air dried, followed by sequential washes by 0.2 M ammonium bicarbonate, Milli Q water and ethanol. After being aired-dried again, the Nylon transfer membrane was transferred to scintillation vials filled with the counting cocktail (RPI) and subject to radioactivity detection using a Beckman LS6500 counter for the tritium radioactivity. For control, the methylation assay included samples containing nucleosome and 3H-SAM only in the reaction buffer, which gave basal radioactivity <70 counts-per-minute (C.P.M.). The source data are provided as a Source Data file.Electrophoretic mobility shift assay (EMSA)DNMT3A1 UDR sample was incubated with nucleosomes for 20 min at 4 °C in a binding buffer (pH 7.5, 50 mM NaCl, 5% Glycerol, 0.05% β-ME) and run on a native 6% polyacrylamide gel. The electrophoresis was performed using a Tris-borate buffer (25 mM Tris, 12.5 mM boric acid, pH 8.8) at a constant voltage at 100 V for 2 hr at 4 °C. Nucleosomal DNA was detected by staining with SYBR Gold and visualized by gel imager.Establishment of stable cell linesThe murine DNMT-TKO ESCs were transfected with the pPyCAGIZ vector (a kind gift of J. Wang, Columbia University) carrying GFP-tagged DNMT3A1, either WT or mutant, using a transfection agent of linear polyethylenimine (PEI, Sigma). Two days post-transfection, cells were subject to drug selection in the medium with 100 µg/mL zeocin (Invitrogen) for at least 10 days, followed by immunoblotting to verify DNMT3A1 expression using anti-DNMT3A antibody (abcam, ab2850).Chromatin fractionation assayChromatin fractionation assay was performed as described previously81. In brief, 5 million of cells were harvested and resuspended in 300 µL of ice-cold buffer A (10 mM HEPES [pH 7.9], 10 mM KCl, 1.5 mM MgCl2, 0.34 M sucrose, 10% glycerol, 1 mM DTT, and 0.1% Triton X-100, freshly added with the cocktail of protease inhibitors), and incubated for 10 min on ice. After incubation, samples were centrifuged at 1,300 g for 5 min at 4 °C. An aliquot (30 µL) of the supernatant was taken as the cytoplasmic fraction. The pellet was then washed with 500 µL of buffer A and then resuspended in 30 µL of volume buffer A. 10 x volume of buffer B (3 mM EDTA, 0.2 mM EGTA, 1 mM DTT, and protease inhibitors) was added to each sample. After brief vortex, samples were incubated for 30 min on ice and centrifuged at 1,700 g for 5 min at 4 °C. An aliquot (30 µL) of the supernatant was taken as the nucleoplasmic fraction. Pellets were washed in 500 µL of buffer B and then resuspended in 30 µL of SDS lysis buffer (50 mM Tris-Cl [pH 7.5], 2 mM EDTA, 2% SDS), which represents the chromatin fraction (fraction Chr). All fractions were then mixed with an equal amount of SDS sample loading buffer, boiled for 10 min and subject for western blotting. DNMT3A, histone H3 and GAPDH were detected using anti-DNMT3A (abcam, ab2850), anti-H3 (Cell Signaling Technology, 4499 S) and anti-GAPDH (Cell Signaling Technology, 2118 L) antibodies, respectively. The source data are provided as a Source Data file.Immunofluorescence (IF)IF assay was performed as described previously82. Briefly, cells were fixed in 4% of paraformaldehyde for 10 min at room temperature, followed by incubation in 1 × PBS containing 0.2% of Triton X-100 for 10 min to permeabilize the cells. Fixed cells were stained with primary antibody. Anti-GFP antibody (abcam, ab290) was used for GFP-tagged DNMT3A Isoform 1 (DNMT3A1), followed by staining with the anti-rabbit Alexa-488 conjugated secondary antibody (Invitrogen, #A-11008). Nuclei were finally stained with 4,6-diamino-2-phenylindole (DAPI, 0.1 g/ml, Invitrogen, D3571). Images were taken with an FV1000 confocal microscope (Olympus; available at UNC Imaging Core). Signal colocalization analyses were performed using EzColocalization plugin of FIJI49. 20 cells of each sample were used for IF signal colocalization analysis of DNMT3A1 and DAPI. The source data are provided as a Source Data file.Quantifications of the levels of 5-methyl-2′-deoxycytidine (5-mdC) in genomic DNA by liquid chromatography-tandem mass spectrometry (LC-MS/MS)The levels of 5 mdC in genomic DNA were measured as previously described24. Briefly, 500 ng of genomic DNA was digested with 0.1 unit of nuclease P1 (Sigma) and 0.4 unit of turbo DNase (Thermo Fisher Scientific) in a buffer containing 30 mM sodium acetate (pH 5.6) and 1 mM ZnCl2 at 37 °C for 24 h. The mixture was then treated with 1 unit of Quick CIP (New England Biolabs) and 0.002 unit of phosphodiesterase I (Sigma) in 500 mM Tris-HCl (pH 8.9) at 37 °C for another 4 h. The solution was subsequently neutralized with 1 M formic acid. To 1/10 volume of the resulting nucleoside mixture was added with 0.6 pmol of [13C5]-5-mdC and 16.2 pmol of uniformly 15N-labeled 2′-deoxyguanosine (dG). Enzymes were then removed from the digestion mixture via chloroform extraction, the aqueous layer was dried in vacuo, and the dried residues were reconstitution in 20 μL of doubly distilled water for LC-MS/MS analysis.LC-MS/MS experiments were conducted using a TSQ-Altis triple-quadrupole mass spectrometer (Thermo Fisher Scientific, San Jose, CA) equipped with a Nanospray Flex source and coupled with a Dionex UltiMate 3000 UPLC for separation (Thermo Fisher Scientific, San Jose, CA). The samples were loaded onto a μ-precolumn (C18 PepMap 100, 5 μm in particle size, 100 Å in pore size, Thermo Fisher) at a flow rate of 2 μL/min within 8.5 min and then eluted onto an in-house packed Zorbax SB-C18 analytical column (5 μm in particle size, 200 Å in pore size, Michrom BioResource, Auburn, CA, 75 μm × 20 cm) at a flow rate of 300 nL/min. Formic acid (0.1%, v/v) in water and acetonitrile were used as solution A and B, respectively. A gradient of 0–95% B in 20 min, and 95% B for 10 min was employed. The TSQ-Altis triple-quadrupole mass spectrometer was operated in the positive-ion mode. The voltage for electrospray, capillary temperature, collision energies were 2.0 kV, 325 °C and 20 V. Q1 and Q3 resolutions were 0.7 and 0.4 Th full-width at half-maximum (FWHM), respectively. Fragmentation in Q2 was conducted with 1.5 mTorr argon. Multiple-reaction monitoring (MRM) transitions corresponding to the neutral loss of a 2-deoxyribose (116 Da) from the [M + H]+ of 5-mdC (m/z 242 → 126), [13C5]-5-mdC (m/z 247 → 126), dG (m/z 268 → 152) and [15N5]-dG (m/z 273 → 157) were monitored (Supplementary Data 1). Quantification of 5-mdC and dG (in moles) in the nucleoside mixtures were performed using peak area ratios for the analytes over their corresponding stable isotope-labeled standards, along with the fixed amounts of the added internal standards (in moles), and the calibration curves. The levels of 5mdC, reported as percentages of 5mdC relative to dG, were calculated by dividing the molar quantifies of 5-mdC with those of dG. The p-values were calculated using two-tailed, unpaired Student’s t-test. (EV: n = 6; WT: n = 12; R181A: n = 6; F190A: n = 6). The source data are provided as a Source Data file.Statistics and reproducibilityThe two-tailed Student t-test was performed to compare distributions between different groups. And the p-value <0.05 was considered to be statistically significant. The chromatin fractionation experiment was performed twice. No statistical method was used to predetermine sample size. No data were excluded from the analyses.Reporting summaryFurther information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
