Fig 2: Cytosines in TFF1 and GREB1 enhancers are epigenetically unmodified. a Comparing TDG binding at sites regulated by E2 reveals that TDG binds to hypomethylated locations. b Heatmap comparing methylation signal at sites of TDG binding globally reveals that TDG binds to hypomethylated sites. c MCF7 DNase signal intensity, revealing that, globally, TDG binding occurs preferentially at sites which are hypomethylated and “open.” d Bisulfite sequencing and MAB-Seq at TFF1 after cells were treated with scrambled siRNA (siControl) or siRNA targeting TDG (siTDG) and then with or without E2 treatment. TFF1 enhancer is devoid of any methylation or active demethylation metabolites and remains so in response to E2 or TDG depletion

Fig 3: TDG peaks overlap with sites involved in promoter-enhancer looping. TDG binding was compared to public datasets containing E2-dependent ERa localization and ChIA-PET performed using an antibody against ERa. a Heatmap showing ChIA-PET signal at sites of TDG binding (left) and aggregate plot showing global average (right). b Overlap of TDG, ER and ChIA-PET signal at specific sites reveals TDG, ER and looping occur at precisely the same locations at these sites. c Schematic of GREB1 showing approximate locations of looping as identified by publically available data. d ChIP using ER in the presence and absence of TDG as well as - or + E2, showing that ER binding is unaltered during depletion of TDG. e Loss of TDG prevents enhancer–promoter looping at the GREB1 locus. MCF7 cells were treated with siControl or siTDG, and then treated with 100 nM E2 for 1 h. 3C, semiquantitative method of measuring the looping between the GREB1 enhancer and promoter, revealed that E2-driven looping of the enhancer and promoter is disrupted upon TDG knockdown

Fig 4: Effect of TDG knockout on the activity of GC box containing 5-mC/5-hmC/5-fC/5-caC. (a) Strategy for TDG knockout in Hela cells. The TDG gene was targeted simultaneously at two different sites by a pair of single guide RNAs enclosing the codon for the catalytic R140 residue. The Cas9 cut sites are indicated by arrowheads. (b) Validation of TDG negative clones by Western blotting. Clone F3 further used as a transfection host for the gene expression analyses is marked (*). (c) Expression time course of constructs containing 5-mC/5-hmC/5-fC/5-caC in the purine-rich (left group of plots) or the pyrimidine-rich GC box strand (right). Isogenic clonal TDG knockout (upper row) and NTH1 knockout (lower row) cell lines were compared with the parental HeLa cell line (overlaid in plots). All cell lines were transfected in parallel with the same sets of reporter constructs. Results of n = 3 independent experiments (mean ± SD).

Fig 5: TDG depletion sensitizes MCF7 cells to Tamoxifen and produces proliferation and attachment deficiencies. a Growth curves examining responsiveness of CRISPR-mediated (right panel) and siRNA-mediated (left panel) TDG knockout and knockdown, respectively. MCF7 cells were grown for 48–72 h in charcoal-stripped phenol red-free media prior to treatment (left panel) (n = 3, p value < 0.05, error bars show ± 1 standard deviation). b To examine migration and adhesion, MCF7 cells were treated with either scrambled siRNA or siRNA targeting TDG and grown in regular media for 2 days prior to performing the cell-to-cell adhesion assay. We found that MCF7 s with depleted levels of TDG form significantly less aggregates with one another (n = 4, error bars indicate standard deviation ± 1). c Treatment of MCF7 cells with varying concentration of trypsin revealed that those depleted of TDG detached from tissue-plate substratum at lower concentrations (top panel) and upon resuspension TDG-depleted cells demonstrated re-attachment deficiencies, remaining loosely attached and spherical while cells treated with siControl became stabilized with visible “flattening” indicative of cell-substratum contacts(bottom panel). d To measure the effect of TDG on the migration and invasion capacity, TDG was depleted in MCF7 cells using siRNA which were treated with or without Tamoxifen for 3d. Cells were harvested using 3 mM EDTA, and a suspension of 100,000 cells was incubated at 37 °C for 72 in gelatin or Matrigel to test migration and invasion, respectively. Cells were counted using ImageJ (n = 4). MCF7s depleted of TDG show a drastically more aggressive profile with cells migrating and invading gel at a significantly higher rate than siControl cells. Tamoxifen treatment had an impact on the increased aggressiveness, although failed to restore cells to basal capacity