Structures of human ALKBH5 demethylase reveal a unique binding mode for specific single-stranded N6-methyladenosine RNA demethylation

Abstract

N(6)-Methyladenosine (m(6)A) is the most prevalent internal RNA modification in eukaryotes. ALKBH5 belongs to the AlkB family of dioxygenases and has been shown to specifically demethylate m(6)A in single-stranded RNA. Here we report crystal structures of ALKBH5 in the presence of either its cofactors or the ALKBH5 inhibitor citrate. Catalytic assays demonstrate that the ALKBH5 catalytic domain can demethylate both single-stranded RNA and single-stranded DNA. We identify the TCA cycle intermediate citrate as a modest inhibitor of ALKHB5 (IC50, ∼488 μm). The structural analysis reveals that a loop region of ALKBH5 is immobilized by a disulfide bond that apparently excludes the binding of dsDNA to ALKBH5. We identify the m(6)A binding pocket of ALKBH5 and the key residues involved in m(6)A recognition using mutagenesis and ITC binding experiments.

Keywords: Crystal Structure; Dioxygenase; Isothermal Titration Calorimetry (ITC); RNA Catalysis; RNA Methylation; Tricarboxylic Acid Cycle (TCA Cycle) (Krebs Cycle).

FIGURE 1.

View of the overall structure of the ALKBH5 catalytic domain (amino acids 74–294). A, sequence alignment of the conserved jelly roll core of the AlkB family members, including human ALKBH1–8, FTO, and E. coli AlkB. The sequence of ALKBH5 (β4-β11) is on the top and labeled with its secondary structure. The metal ion binding residues, His-204, Asp-206, and His-266, are in green. The 2OG binding residues, Asn-193, Tyr-195, Arg-277, and Arg-283 are in blue. The RKK motif in ALKBH2 (Arg-241 to Lys-243) and the long loop region of FTO are both in red. B, overall structure of ALKBH5 catalytic domain is shown in a blue cartoon representation. C, the overall structure of the ALKBH5 catalytic domain is shown in a schematic form, including all α-helices and β-strands.

FIGURE 2.

Crystal structure of ALKBH5 complexed with 2OG and Mn(II). A, m⁶A demethylation requires 2OG and Fe(II). B, overall structure of ALKBH5 (gray cartoon) in complex with 2OG-Mn(II). The protein is in a gray cartoon. 2OG-binding and Mn(II)-binding residues are in blue and green sticks, respectively. C, detailed interactions of ALKBH5 with 2OG and Mn(II). Hydrogen bonds are shown as black dashes. Bonds formed with the metal ion are shown as red dashes. The residues are shown in the same color as in Fig. 2B. The F_o − F_c omit map of 2OG-Mn(II), contoured at 2.8 σ, is prepared using the DELFWT and PHDELWT coefficients calculated by REFMAC. D, ITC binding curves of ALKBH5 WT (left panel), R130A (middle panel), and Y139A (right panel) binding to Mn(II) and 2OG.

FIGURE 3.

Enzyme assays for the ALKBH5 catalytic domain wild type enzyme and variants. A, MALDI spectra showing the demethylation status (−14 Da) of 20 μm of 5-mer ssRNA (5′-GGm⁶ACU-3′) (1580 Da). From top to bottom, the panels show: without enzyme, in the presence of wild type ALKBH5_74–294, ALKBH5_74–294, R130A mutant, ALKBH5_74–294Y139A mutant, and ALKBH5_74–294E153A mutant after a 7-min reaction. In the bottom panel, the activity of wild type ALKBH5_74–294 (wt.), ALKBH5_74–294R130A, ALKBH5_74–294Y139A, and ALKBH5_74–294E153A, relative to wild type are shown. The activity of E153A is similar to that of wild type, whereas the R310A and Y139A variants are not active within the limit of detection (<5%). The error bars were derived from standard deviations of triplicate experiments (n = 3). B, ALKBH5 also catalyzes the ssDNA: dGdG(m⁶A)dCdT. The activity of ALKBH5 toward ssRNA was set to 100% for comparison. The activity of ALKBH5 toward ssDNA was 168 ± 390%. No activity was detected for ALKBH5 when Fe(II) was replaced by Mn(II). C, K_d values of ALKBH5 (wild type or variant) binding to modified ssDNA/ssRNA as determined by ITC.

FIGURE 4.

View from a structure of ALKBH5 complexed with citrate. A, detailed interactions between ALKBH5 and citrate. The citrate and citrate binding residues of ALKBH5 are shown as yellow and blue sticks, respectively. Intermolecular and intramolecular hydrogen bonds are shown as black and red dashes, respectively. The F_o − F_c omit map of the citrate, contoured at 2.8 σ, is prepared using the DELFWT and PHDELWT coefficients calculated by REFMAC. B, detailed interactions between FTO and citrate (PDB entry 4IE7). The citrate and citrate binding residues of FTO are shown as yellow and green sticks, respectively. Bonds formed with the metal ion (Mn(II)) are shown as red dashes. C, ITC binding analysis of ALKBH5 and citrate. D, dose-response curve of 5-mer ssRNA (5′-GGm⁶ACU-3′) demethylation reaction by ALKBH5_66–292 in the presence of different concentrations of citrate. The IC₅₀ value of citrate calculated from this curve is 488 μm. The error bars were derived from standard deviations of triplicate experiments (n = 3).

FIGURE 5.

Structural comparison of ALKBH5 (pale cyan cartoon) and ALKBH2 (gray cartoon)-dsDNA (green cartoon, PDB entry 3H8O). The Arg-241 to Lys-243 motif and the β3-β4 strands of ALKBH2 are colored in orange. The loop between β7 and β8 of ALKBH5 is a blue cartoon. A, overall comparison of ALKBH5 and ALKBH2. B, the Arg-241 to Lys-243 motif (in stick mode) contacts the backbone of the nucleotides in the complementary DNA strand (green sticks). C, β3-β4 of ALKBH2, which is missing in ALKBH5, contacts the DNA A-T pair (green sticks) in dsDNA via F102 (orange sticks). Hydrogen bonds between A and T are shown as red dashes. D, the unique disulfide bond, formed between Cys-230 and Cys-267 of ALKBH5 (blue sticks), immobilizes the loop conformation to exclude the binding of complementary strand nucleic acids.

FIGURE 6.

Sequence alignment of ALKBH5 orthologs. Hs, Homo sapiens; Ga, Gallus gallus; Xs, Xenopus; Dr, Danio rerio; Hv, Hydra vulgaris. Metal binding residues are in green. The two cysteines that form the conserved disulfide bond are in orange. Arg-130, Tyr-139, Arg-277, and Arg-283 are marked.

FIGURE 7.

Modeled structure of ALKBH5 complexed with an m⁶A-containing substrate compared with analogous complexes for other AlkB family members. Substrates and the residues involved in substrate binding are shown in stick mode. Hydrogen bonds are shown as black dashes. Bonds with the metal ion are shown as red dashes. A, structure of E. coli AlkB (cyan) and m¹A containing ssDNA (yellow) derived from PDB entry 2FD8. B, structure of ALKBH2 (green) and m¹A containing dsDNA (yellow) derived from PDB entry 3BUC. C, structure of FTO (red) and m³T (yellow) derived from PDB entry 3FLM. D, electrostatic surface of ALKBH5 binding to m⁶A (yellow). E, modeled structure of ALKBH5 (gray) and m⁶A (yellow). Residues involved in recognizing m⁶A and cofactor/cosubstrate are shown as blue and gray sticks, respectively.