Fig 2: Binding to U-rich elements of H19 inhibits SAHH activity in vitro.(a) Sequences of human and mouse H19 fragments used in EMSA and in vitro SAHH activity assays, with the U-rich stretches underlined. Numbers in parentheses mark the nucleotide positions of the sequences relative to the transcription start sites of H19. (b) EMSA results using biotin-labelled hURE, mURE or hNSE with increasing concentrations of rSAHH. The positions of unbound RNA and RNA–protein complexes are marked to the right. Band intensities were quantified from three independent experiments and used to calculate the Kd values displayed in the bottom panel. (c) EMSA results with biotin-labelled hURE and 2 mM of rSAHH. The amounts of unlabelled competitors added in relative molar excess are indicated on top. Band intensities were calculated and presented as percentage of fraction bound. Numbers are average of two independent experiments. (d) In vitro rSAHH activity assay showing dose-dependent inhibition of the hydrolytic activity of rSAHH by the UREs and poly(U). Student's t-tests were used to compare differences between quantitative variables. Numbers are mean±s.d. (n=4). **P<0.01, *P<0.05.

Fig 3: A model for H19/SAHH-dependent regulation of gene methylation.(a) Myotubes were transfected with siCon or siH19, followed by RNA extraction and RT–qPCR analysis 48 h later. Relative H19 level is presented. Student's t-tests were used to compare differences between quantitative variables. Numbers are mean±s.d. (n=3). **P<0.01. (b) DNMT3B ChIP at the DMRs of the indicated genes. ns, not statistically different. (c) In the absence of H19 (left panel), SAHH hydrolyses SAH to homocysteine and adenosine, and DNMT3B is active in DNA methylation. When H19 is present (right panel), SAHH activity is attenuated because of its association with H19. This leads to accumulation of SAH, which binds to DNMT3B and prevents it from methylating DNA.

Fig 4: DNMT3B contributes to SAHH-mediated, H19-dependent methylation change in CME.(a) Myotubes were transfected with siCon or siH19, followed by RNA extraction and RT–qPCR analysis 48 h later. Relative RNA levels of H19 and Sahh are presented. (b) In vivo SAHH activity assessed 48 h post-siRNA transfection, as described in a. (c) Relative DMR methylation assessed using QMSP. (d) Sequence of the DMR amplified using the QMSP primers, with the differentially methylated cytosine highlighted in red. Numbers above the sequence mark positions of the indicated nucleotides in the chromosome. Numbers below indicate percentage of methylation in siH19 (left) versus siCon (right) cells, as determined using genome-wide methylation analysis. (e) Myotubes were transfected with siCon, siH19 or siH19 plus siDnmt3b, followed by RNA extraction and RT–qPCR analysis 48 h post transfection. Relative RNA levels of H19 and Dnmt3b are presented. (f) Western blot results representative of three independent transfection experiments (performed as described in e) are shown. Antibodies specific for DNMT3B and beta-tubulin were used on the top and bottom blots, respectively. Numbers underneath the blots indicate DNMT3B protein levels after normalization against TUBB loading controls with that in siCon cells arbitrarily set as 100%. (g) Myotubes were transfected with siCon, siH19 or siH19 plus siDnmt3b. DMR methylation was assessed 48 h post transfection using QMSP. (h) DNMT3B occupation at CME. Myotubes were transfected with siCon or siH19. ChIP analysis was performed 48 h later using anti-DNMT3B antibody to evaluate binding of DNMT3B at the DMR. ns, not statistically different. (i) Association of activated RNAP with Nctc1 and H19. Myotubes were transfected as described in h. ChIP analysis was performed using anti-RNAP antibody. (j) Relative RNA levels of H19 and Nctc1 following siRNA transfection. Student's t-tests were performed to compare differences between quantitative variables. Numbers are mean±s.d. (n=3). *P<0.05, **P<0.01.

Fig 5: SAHH interacts with H19 in RNPs.(a) Schematic outline of purification of H19-associated RNPs and protein component identification. (b) Image of a Commassie staining gel with protein size markers (in kDa) labelled on the left. The SAHH protein band is marked with a red arrow. (c) pH19-S1 or pH19 were transfected into HEK293 cells, followed by affinity purification as outlined in a. Purified RNPs were resolved using 10% SDS–PAGE, followed by western blot analysis using rabbit polyclonal antibodies specific for SAHH (top), HuR (middle) or beta-tubulin (bottom). Ten percent input was loaded on lanes 1 and 2. Molecular marker positions in kDa are shown on the right. (d) RIP with mouse monoclonal anti-SAHH or pre-immune IgGs from myotube extracts. RNA levels in immunoprecipitates were determined using reverse transcription and qPCR. Levels of H19 and GAPDH mRNA are presented as fold enrichment in anti-SAHH relative to IgG immunoprecipitates. Student's t-tests were used to compare differences between quantitative variables. Numbers are mean±s.d. (n=3). **P<0.01. (e) Relative RNA levels of H19 and GAPDH in myotubes. (f) Immunoprecipitation using monoclonal anti-SAHH (lane 2) or IgG (lane 3), followed by western blot analysis using a rabbit polyclonal anti-SAHH. Five percent input was loaded in lane 1. The SAHH band is marked. (g) Cartoon depiction of the Igf2-H19-Nctc1 locus (not drawn to scale). The three exons of Nctc1 are marked as red boxes, and the CME as a blue dot. The transcription start site of Nctc1 is set as 0 kb.