Skip to main content
Fig. 3 | BMC Genetics

Fig. 3

From: CGGBP1-regulated cytosine methylation at CTCF-binding motifs resists stochasticity

Fig. 3

Cytosine methylation changes caused by CGGBP1 depletion are less stochastic at CTCF-binding RFM than at MFR. a CTCF-binding sites in HEK293T are enriched in low methylation bins with no strong differences in MeDIP signals between CT and KD. b MeDIP reads in HEK293T CT and KD are equally enriched in peaks (dashed lines) that are positive for CTCF motifs (continuous lines-circles). In agreement with the cytosine methylation-sensitivity of CTCF binding to DNA, the occurrence of peaks and CTCF motifs decline with an increase in MeDIP signal. c In GM02639 the concentration of CTCF-binding sites at low methylation bins is enhanced by CGGBP1 depletion. d The occurrence of peaks in the GM02639 MeDIP reads (dashed lines) and their CTCF motif positivity (continuous lines-circles) regresses at a higher rate in KD showing that the cytosine methylation sensitivity of CTCF-binding at motifs in GM02639 is stronger in absence of CGGBP1. e and f Cytosine methylation patterns at CTCF-binding sites in CT (e, n = 42,978 CTCF peaks) or KD (f, n = 47,632 CTCF peaks) are disrupted upon CGGBP1 depletion as revealed by MeDIP signals in the 10 kb flanks of the peak centres (X-axes). g and h GM02639 MeDIP signals in the 10 kb flanks of peak centres (X-axes) at the same CTCF-binding sites (as shown in e and f) highlight three differences from the pattern observed in HEK293T (g and h compared with e and f respectively); overall low cytosine methylation levels, a stronger loss of methylation in KD compared to CT, and a larger disturbance in methylation caused by CGGBP1 depletion. A comparison of e and f with g and h respectively reinforces the findings that the cytosine methylation difference between CT and KD is less stochastic (Fig. 2a) and the CTCF-binding motifs in KD are recused to a low methylation status in GM02639 (Fig. 2c and d, and Fig. 3c and d). i The cytosine methylation changes observed in HEK293T (a, b, e and f) affect the repeat-derived (MFR) and motif-derived (RFM) CTCF-binding sites differently with a slight net GoM at RFM. j The cytosine methylation changes in GM02639 CT and KD were strongly different between MFR (stochastic GoM and LoM with a normal distribution of Diff/Sum) and RFM (strong GoM and LoM with a multimodal distribution of Diff/Sum). A comparison of i and j shows that the cytosine methylation at RFM, unlike MFR, is specifically regulated by CGGBP1. The CGGBP1-dependence of RFM cytosine methylation is higher in GM02639 than HEK293T. k PCA analyses reveal the different levels of stochasticity of methylation changes at RFM and MFR. The major component of variance (PC1) shown on the X-axis represented stochastic changes as it failed to segregate the CT and KD samples of the two cell types when RFM and MFR were analyzed separately or together. Commensurate with the previously described findings, the PC1 (X-axis) accounted for the least stochastic variance in RFM (77.9%) and highest (90.6%) for MFR. The PC2 (Y-axis) accounted for 11.2% of variance between GM02639 CT and KD RFM only. For MFR and all sites, the PC2 (Y-axis) accounted for variances between the cell types but not CT versus KD. Thus, at RFM the MeDIP signals have a GM02639-specific dependence on CGGBP1 whereas the same at MFR follow a cell type-specific pattern predominantly

Back to article page