Figures and data
Structure of m6A bound human METTL3-METTL14 MTase core.
a, Domain architecture of METTL3 and METTL14; and boundaries of each used in crystallization are shown on top. Structure of the complex is shown in cartoon mode for METTL3 in cyan and METTL14 in orange; m6AMP (red) and interacting residues of METTL3 (green) and METTL14 (orange) are shown in stick mode. Blue dot, the position of N6 (in acceptor mode), i.e., ~3Å from the methyl group of the donor SAM. Methylated N6 of m6A is ~16Å away from its acceptor position in the catalytic pocket (blue dot). Black dots, water. Black dashes, h-bonds. The panel on right shows a close-up of the m6A interaction network, including the arginine clasp. b, An alignment of the regions participating in m6A confirms strict conservation of the interaction network throughout the evolution from yeast (Uniprot ID: P41833); arabidopsis (082486) and rice (Q6EU10); fruit fly (Q9VCE6), zebrafish (F1R777), mouse (Q8C3P7), hamster (A0A1U7R3Z3), and monkey (A0A8J8YGJ7); to human (Q86U44). c, Methyltransferase activity results of full-length human METTL3-METTL14 (wild-type, WT) and eight mutant enzymes as derived from three independent experiments, with error bars indicating the range of data points from these experiments (n = 3). d, Quantitative measurement of RNA (red, m6A-RNA; blue, A-RNA) binding (n = 3) by the WT enzyme shown as binding isotherms fitted with a one-site specific binding model. The equilibrium dissociation constant or Kd derived for each mutant enzyme is plotted along with Kd of the WT enzyme (e). ns, not significant (p >0.05), * denotes p ≤ 0.05. Source data for panels c-e are provided.
Enzyme and binding kinetics.
a, Methylation of NEAT2* RNA by full-length METTL3-METTL14 (wild-type, WT and its mutants) at saturating concentrations of SAM and RNA. b-c, Kinetics of RNA binding to the WT and mutant METTL3-METTL14 as measured using surface plasmon resonance. Two RNA oligos (NEAT2* and a single-stranded RNA) comprising substrate A (grey circle) or product m6A (red circle) were probed.
Capture level of different RNA substrates
Kinetic parameters for all evaluated bindings with 1:1 binding model
Kinetic parameters for all evaluated bindings with two state reaction model
Base swiveling and loop orchestration.
a-c, Upper panels show overlays of regions of METTL3 encompassing the catalytic motif, gate loops 1 and 2, and interface loop in METTL3 bound to m6A (red stick), SAM (pink stick), and SAH (orange stick). Arrows indicate the directional movement of loops. Lower panels: The entire region of each overlay is in stick mode. Green dots, the residues that form the m6A interaction network. d, Close-up of an overlay of m6A and apo MTase of METTL3-METTL14 shown in two orientations for clarity. The exit channel between M402 and H474 in the m6A bound conformation becomes wider (up to 8Å) to stabilize m6A and avoid steric clashes with its purine and ribose moieties. e, An overlay of MTase cores of arabidopsis METTL4 (light blue cartoons)/SAH (light blue stick)/Am (blue stick) and METTL3 (cyan)-METTL14 (orange)/m6A (red stick) clarifies the ~ 120º pivot of the base around phosphate. Black dots, water molecules in the m6A structure help stabilize the m6A and compensate for the loss in binding energy in the site emptied by base pivoting. f, Change in emission fluorescence intensity upon titration of increasing concentration of WT (upper panel) and R298P mutant enzymes (lower panel) with 2-aminopurine (2-Ap) containing RNA (n=3). See the methods section and source data for details.
Mode of m6A binding by writer/sensor, eraser, and reader.
Interaction networks of m6A (red) binding to METTL3 (green), and METTL14 (a), 6mA (blue) binding to FTO (b), and m6A binding to YTH domain of YTHDC1 (c). The two nucleotides flanking the flipped methylated base in FTO and YTHDC1 are shown in light blue and grey, respectively. The hydrophobic stacking surface in YTHDC1 can only be aligned by rotating the molecule 180º around the x-axis, suggesting that reader proteins approach RNA from the opposite direction. The m6A pocket of METTL3-METTL14 harbors features that enable it to act as an atypical m6A sensor/reader during its switch from writer to reader. Dashed lines, h-bonds.
Data collection and refinement statistics (molecular replacement)
a, Coomassie-stained SDS-PAGE showing high purity of METTL3-METTL14 MTase core. b, Size-exclusion chromatography coupled with multi-angle scattering estimated a molecular mass of this complex of ~59 kDa, in excellent agreement with theoretical mass (~ 60 kDa). c, crystals of METTL3-METTL14 MTase core used for soaking N6-methyladenosine monophosphate (m6AMP or m6A). d, A green mesh showing an unbiased electron density omit map countered at 2.2σ, confirming the presence of m6A. e-f, A blue mesh representing a 2Fo-Fc map (σ = 1.0) showing the final refinement of the m6A-METTL3-METTL14 structure. The map confirms the excellent agreement between calculated and observed electron density for the region surrounding m6A and for m6A itself (f). g, Network of interaction between m6A and residues from METTL3 (green) and METTL14 (orange). g, Quantitative measurement of 30-mer RNA (left panel, A-RNA; right panel, m6A-RNA) binding (n = 3) to the WT and mutant enzymes shown as binding isotherms fitted with a one-site specific binding model, (Y=Bmax*X/(Kd + X). The equilibrium dissociation constant (Kd) was derived from three independent experiments, with error bars indicating the range of data points (n = 3). m6A replaces the central adenine base within the GGACU motif in A-RNA in m6A-RNA. See the methods section and source data for details.
a, Am (blue stick) binding into ‘DPPW (or motif IV)’ in Arabidopsis METTL4 (green stick, PDB: 7CV6). Blue mesh, 2Fo-Fc electron density map contoured at σ = 0.97. b, BA2 SAM-Adenosine covalent analog sampling two different conformations in human METTL3-14 MTase in complex with BA2 (Corbeski et al., PMID: 38470714). c. Overlay of METTL3-14 MTase core bound to BA2 and BA4 analogs and m6A (this study).
The distance between the center of mass of m6A and the center of mass of the product m6A binding pocket for (a) mutants and (b) nucleotides.
The most stable conformation of the mutants, as identified from SuMD simulations.
The distance between the center of mass of (a) m6A and the center of mass of the product m6A binding pocket; the distance between m6A and (b) H478 and (c) R298 and (d) the plots of interaction energies from the three replicas of SuMD.
The base swiveling mechanism of m6A between the starting conformation (orange) and the end (green) conformation. Different snapshots of m6A are illustrated in cpk sticks. The arrow shows the direction of the swiveling nucleotide.
