Figure 2

Cell culture-induced 5hmC loss occurs genome-wide and does not reflect changes in 5mC. (a) Dot-blot of genomic DNA (500 ng) from MEF primary tissue and cultured cells derived therefrom with a 5hmC-specific antibody (left). Quantification of global 5hmC levels by densitometric analysis of dot-blots (right). 5hmC values were normalized to the single-stranded DNA (ssDNA) loading control. *P < 0.05, Mann-Whitney U-test. (b) Heatmap showing median 5mC enrichment levels in non-overlapping 2 Mb windows of the mouse genome (chromosomes 14 to 19 and X) from primary MEF tissue and cultured cells derived therefrom, as determined by MeDIP-chip. (c) 5mC-enrichment was highly correlated between MEF tissue and matched cultured cells derived therefrom (Spearman’s rho = 0.98, P < 0.00001). A scatterplot of 5mC enrichment in male replicate ‘A’ before and after cell culture. (d) Autocorrelation analysis of 5mC enrichment of all probes (2.1 million) on the array. (e) Heatmap showing median 5hmC enrichment levels in non-overlapping 2 Mb windows of the mouse genome (chromosomes 14 to 19 and X) from primary MEF tissue and cultured cells derived therefrom, as determined by hMeDIP-chip. (f) A scatterplot of 5hmC enrichment in male replicate ‘A’ before and after cell culture. (g) Autocorrelation analysis of 5hmC enrichment of all probes (2.1 million) on the array revealed a marked reduction in concordance between the 5hmC values of adjacent probes for MEF tissue and matched cultured cells derived therefrom. (h) Clear patterning of 5hmC is lost upon adaptation of MEFs to culture. A schematic representation of the 5mC (left panel, red) and 5hmC (right panel, blue) array profiles over a 40 Mb section of chromosome 14; adapted from the UCSC Genome Browser. Refseq transcripts (transcript) mapped to the region are also shown.