Fig 2: Global gene expression profiling indicates ALKBH5 downstream targets related to DNA damage repair are induced by ROS. (A, B) The frequency distribution of m6A peaks across the length of mRNA transcripts shown by metagene in control cells (A) or in cells expressing ALKBH5 K86/321R (B) with or without H2O2 treatment. Each region of the 5′ untranslated region (5′ UTR), coding region (CDS), and 3′ untranslated region (3′ UTR) was split into 100 segments, and the percentage of m6A peaks that fall within each segment was determined. (C, D) The density (line) and frequency (histogram) of m6A peaks in control samples (C) or SD-mutant overexpressed samples (D) with or without H2O2 treatment. (E) The adjusted density (line, top) and distribution (histogram, bottom) of hyper peaks from (C) across different mRNA regions in control samples with or without H2O2 treatment. (F) Enrichment analysis for significantly increased peaks of m6A modification from (C) in control samples with or without H2O2 treatment. (G) Differentially expressed genes shown in volcano figure in control samples with or without H2O2 treatment. There were 1051 genes withy significantly reduced expression (log2FC < 0, P < 0.01), 1106 genes with significantly increased expression (log2FC > 0, P < 0.01) and 15 102 genes without statistically significant changes in expression. (H) Differentially expressed genes shown in volcano figure in SD-mutant ALKBH5 overexpressing cells with or without H2O2 treatment. Almost all the transcripts show negligible expression changes. (I) m6A peak visualization of key transcripts in DNA repair in control samples with or without H2O2 treatment. (J) qRT-PCR analyses showing that ROS markedly up-regulates transcription of the three selected target genes (FAAP20, ATRX and RRC2). (K) mRNA m6A methylation validation of the three selected target genes (FAAP20, ATRX and RRC2) by MeRIP analysis. (L) ALKBH5 RIP analyses showing that H2O2-induecd ROS dramatically decreases ALKBH5 enrichment at FAAP20, ATRX and RFC2 mRNAs. (M–O) mRNAs half-life of the three selected target genes (FAAP20, ATRX and RRC2), with or without ROS treatment.

Fig 3: ALKBH5 SUMOylation is mainly responsible for ROS-induced global increase in mRNA m6A methylation. (A) Dot-blot assay showing that SUMOylation-deficient mutant ALKBH5 overexpression completely blocks ROS-induced global increase in mRNA m6A methylation. Strep tagged wild-type or SD-mutant ALKBH5 was transfected into HEK293T cells. Half of cells were used for western blot analysis to confirm the expressions of ectopically expressed plasmids, and the rest of cells were used for dot-blot analysis. (B, C) LC–MS/MS analyses indicating ROS induces global mRNA m6A methylation via ALKBH5 SUMOylation. HEK293T cells stably expressing vector, wild-type, or SUMOylation-deficient mutant ALKBH5 were treated with or without H2O2 (B) or CCCP (C) for 6 h, and subjected to LC–MS/MS analysis of mRNA m6A methylation. (D, E) γ H2A.X immunostaining analysis showing that SUMOylation-deficient mutant ALKBH5 overexpression dramatically delays H2O2–induced DNA damage repair. HeLa cells stably expressing Vec (D), or ALKBH5 K86/321R (E) were treated with H2O2, which was removed after 6 h. γ H2A.X immunostaining analyses were performed at time intervals between 10 and 30 min. (F) γ H2A.X foci quantitative data for D, E and Supplementary Figure S17. (G, H) Western blot analysis showing that SUMOylation-deficient mutant ALKBH5 overexpression significantly inhibits H2O2–induced DNA damage repair. (I) Cell apoptosis analyses showing ALKBH5 K86/321R overexpression markedly increases ROS-induced cell apoptosis. Apoptosis analyses were performed in HEK293T cells transfected with Strep tagged vector, or vectors expressing ALKBH5 or Strep tagged ALKBH5 K86/321R in the presence or absence of H2O2. (J) Histograms showing the summary and statistical analysis of Figure 3I.

Fig 4: mRNA m6A modification enhances mRNA stability of FAAP20 and ATRX by IGF2BPs. (A, B) RT-PCR analysis showing that knockdown of IGF2BP1/2/3 completely blocks H2O2-induecd upregulation of FAAP20 and ATRX. (C, D) IGF2BP2 RIP analysis showing that H2O2 treatment significantly increases IGF2BP2 enrichment at mRNAs of FAAP20 and ATRX. (E-G) Western blot analyses indicating that knockdown of FAAP20, ATRX and RFC2 markedly increases CCCP-induced DNA damage in HEK293T cells. (H) Annexin V staining analysis showing that CCCP-induced cell apoptosis can be dramatically promoted by knockdown of FAAP20, ATRX and RFC2.

Fig 5: ROS stress activates the ERK/JNK/ALKBH5 PTMs/m6A methylation axis in mouse HSPCs in vivo. (A–E) CCCP or vehicle (DMSO) was intraperitoneally injected into three pairs of mice. And, all the mice were sacrificed 12 h after injection. (A, B) FACS analysis showing the effect of CCCP on endogenous ROS in mouse bone marrow stem/progenitor cells. (C) Western blot and dot-blot analysis showing phosphorylation of H2A.X and global mRNA m6A methylation levels in mouse bone marrow progenitor cells with or without CCCP treatment respectively. (D, E) Denaturing IP analysis indicates that ALKBH5 phosphorylation and SUMOylation can be dramatically induced by CCCP injection in mouse bone marrow progenitor cells. (F) Western blot analysis showing CCCP-induced DNA damage can be significantly enhanced by wild-type but not enzymatic mutant ALKBH5 overexpression in HPC-7 cells. (G) Annexin V staining analysis showing that ALKBH5 overexpression markedly facilitates CCCP-induced cell apoptosis in HPC-7 cells. (H) Histograms showing the summary and statistical analysis of G. (I) Working model of ROS mediated regulation of DNA repair genes. ROS stress activates ERK/JNK signaling resulting in phosphorylation of ALKBH5 at serine residues S87 and S321. ALKBH5 phosphorylation facilitates its SUMOylation at lysine residues K86 and K321, leading to the inhibition of its m6A demethylase activity and increase m6A modification in mRNAs of DNA damage repair related genes, such as FAAP20, ATRX and RFC2. As a consequence, increased expression of DNA repair genes protects cells from ROS- induced DNA damage.