Modulation of DNA methylation by L-ascorbate and 5-aza-2’-deoxycytidine in murine embryonic stem cells

Abstract

Cytosine methylation, normally found on cytosine residues adjacent to guanine (i.e., a CpG dinucleotide), is one means by which long-term gene repression occurs. Immediately after semi-conservative DNA replication, CpG dinucleotides on the replicated daughter strand are unmethylated, giving the “hemimethylated” state, where one DNA strand is methylated and one unmethylated. Hemimethylation is usually corrected through complementary methylation by Dnmt1 maintenance methyltransferase, but oxidative stress can inhibit Dnmt1, and so replicated DNA will remain hemimethylated. The Morison laboratory had new evidence that this aberrant hemimethylated DNA is “corrected” by the Tet family of enzymes, which actively catalyse the conversion of parent strand 5-mC to 5-carboxylcytosine (5-cC), that is subsequently replaced with cytosine by base excision repair.

This study developed a murine embryonic stem cell model through which the molecular basis of active demethylation could be investigated. It was hypothesised that Tet family enzymes actively demethylate DNA, following oxidative stress-mediated inhibition of Dnmt1; i.e., after the generation of hemimethylated DNA. To model oxidative stress-induced hemimethylation the Dnmt1 inhibitor decitabine was used. The effect of ascorbate on Tet activity, both alone and in conjunction with decitabine was also assessed. Ascorbate increases Tet activity by increasing regeneration of Fe2+, whilst decitabine (5-aza-2’-deoxycytidine) inhibits Dnmt1 by binding and sequestering it. We hypothesise that when Tet is activated using ascorbate and Dnmt1 is inhibited by decitabine the proportion of unmethylated DNA should increase, due to the creation of hemimethylated DNA by decitabine and the upregulation of Tet activity by ascorbate. To perform this research, murine embryonic stem (ES) cells were maintained and manipulated in cell culture. Cell lines were synchronised to G1 by thymidine block. ES cells received differential treatments: control culture, + decitabine, + ascorbate, and + decitabine / + ascorbate. Samples were extracted at 2, 4 and 6 hours post-release from thymidine synchronisation. Hairpin linkers were used to maintain the connection between complementary DNA strands throughout PCR amplification, allowing comparison of parent-daughter strand methylation to identify hemimethylated sequences. Hairpin linkers were synthesised for three highly methylated genes: Asz1, Ckt2 and Kcnv2. Two next-generation sequencing libraries were prepared, containing 144 and 288 samples respectively, and sequenced using the high-throughput Illumina MiSeq platform. Bisulfite-converted sequences were aligned using BiQ Analyzer HT software, and methylation symmetry in complementary DNA strands determined using RStudio. Methylation proportions were averaged and/or correlated with results for each replicate, and plotted. This project identified that ascorbate induced marked demethylation in ES cells, whilst decitabine caused large increases in hemimethylation, but no increase in demethylation. When decitabine was added in conjunction with ascorbate, increasing demethylation was observed. These findings demonstrate that decitabine has the potential to induce marked hemimethylation, even in wild-type cells, in ascorbate-deficient culture. There was also evidence to suggest that in the presence of hemimethylated target sequence, ascorbate is necessary to allow Tet activity. This was seen in Tet-Triple knockout ES cells, where ascorbate failed to induce demethylation. Overall, the results support the hypothesis that hemimethylated DNA has the potential to induce Tet enzyme activity.