Silencing of DNMT3s impairs the effects of TGF-β on liver cancer cells. a. HepG2 cells were transfected with 20nM of a non-targeting siRNA, or a pool of siRNAs against DNMT3A, DNMT3B, or a combination of both. One day after transfection cell medium was replaced by standard culture medium or medium containing 10nM TGF-β for each condition. Two days after treatment with TGF-β cells were collected from all conditions for morphological evaluation, CD133 FACS expression analysis, RNA and DNA extraction. All transfections were performed in triplicate wells. b. qRT-PCR was performed to assess the efficiency and specificity of the transfections. Higher efficiency was obtained with siRNA against DNMT3B alone or in combination with DNMT3A. c. Representative Phase contrast images for all conditions (compare with Figure S2c for the morphology of non-transfected HepG2 cells in basal conditions and in response to TGF-β). d. cells were collected after treatment and processed immediately for analysis of CD133 surface expression using FACS. Secondary antibody staining (with no primary CD133 staining) is shown as control in the top panel, and was used as a reference for the gates shown below. The plots show one representative histogram for each condition, and the corresponding percentage of CD133+ cells (compare with Figure 1a and Figure 7e for the basal expression and variation in HepG2 non-transfected cells). e. Mean + SEM for CD133 FACS expression in triplicates of each condition. A non-transfected control was also included in this analysis. Asterisks depict the significance between control and TGF-β -treated cells transfected with non-targeting siRNA, and all other TGF-β -treated conditions compared to non-targeting TGF-β -treated cells. f. DNA was extracted and bisulfite modified for pyrosequencing of TRRAP. (*) indicates P value < 0.05 relative to non-treated cells within each experimental condition.