Fig 2: DNMT3A and -3B have different preferences for flanking sequences for CpG and non-CpG sites. Box plot of methylation levels in NCGN context after incubation with (A) DNMT3A or (B) DNMT3B for 240 min, ranked by median methylation level. Sequence logo of the 1000 most methylated 10-mer CG sequences after incubation with (C) DNMT3A or (D) DNMT3B for 30 min. Sequence logo of the 1000 most methylated 10-mer non-CpG sequences after incubation with (E) DNMT3A or (F) DNMT3B for 120 min.

Fig 3: Effect of curcumin on DNMT3B and role of DNMT3B in PKM splicing. (a-b) Methyltransferase inhibition activity of curcumin using an in-vitro methyltransferase-assay kit, with (a) nuclear-extracts of the HNC cells, (b) purified DNMT3B enzyme and (c-e) RPS16 normalized qRT-PCR in shDNMT3B transfected cells versus shcontrol using the indicated primers for (c) DNMT3B and (d-e) PKM gene. (f) Western blot showing the protein level of DNMT3B, PKM2, and PKM1 in shDNMT3B transfected cells versus shControl in H157 cells, GAPDH act as a loading control. (g) MeDIP in shDNMT3B transfected cells versus shcontrol in H157 cells and qRT-PCR relative to input and control IgG. (h-i) ChIP in H157 cells transfected with shDNMT3B versus shcontrol using (h) BORIS and (i) RNA Pol II antibody, followed by qRT-PCR relative to input. Three independent experiments were conducted with mean values ± SD. P value using two-tailed Student’s t-test, * P < 0.05, ** P < 0.01, *** P < 0.001, ns = non-significant

Fig 4: Methylation signatures in the mouse and human genome depend on the presence of DNMT3A or DNMT3B. Sequence logos of the most methylated sequence contexts in chromosome 1 of mouse stem cells. (A) Top methylated 10-mer CG sequences: 100% methylation for WT (left; N = 325622), >60% methylation for TKO with Dnmt3a (middle; N = 628), and >40% for TKO with Dnmt3b (right; N = 399). (B) Top methylated 10-mer CH sequences: >40% methylation for WT (left; N = 16392), >30% for TKO with Dnmt3a (middle; N = 533), and >20% for TKO with Dnmt3b (right; N = 177). Sequence logos of the most methylated (>20% methylation) CH sequence contexts in chromosome 1 of hESCs. (C) WT hESCs (N = 163050). (D) DNMT3A-KO hESCs at early passage 7 (left; N = 28101) and late passage 22 (middle; N = 59323). (E) DNMT3B-KO hESCs at passage 6 (left; N = 24327) and passage 22 (middle; N = 20603). (F) DNMT3A/B double-knockout hESCs at passage 7 (left; N = 12736) and passage 22 (right; N = 6859). (G) Ratio between CAC and CAG methylation in mouse Dnmt-TKO cells with Dnmt3a/Dnmt3b reintroduced and hESCs with either DNMT3A or DNMT3B knocked out.

Fig 5: In vitro methylation assay using recombinant full-length human DNMT3A and -3B. (A) Average level of methylation introduced by DNMT3A, DNMT3B, or DNMT3L in CpG and non-CpG sites. Box plots of methylation levels at NCN sequences after incubation with (B) DNMT3A or (C) DNMT3B for 240 min, ranked by median methylation level. Sequences are written in 5′ to 3′ order.