3,N4-Etheno-5-methylcytosine blocks TET1-3 oxidation but is repaired by ALKBH2, 3 and FTO

Abstract

5-Methyldeoxycytidine (5mC) is a major epigenetic marker that regulates cellular functions in mammals. Endogenous lipid peroxidation can convert 5mC into 3,N4-etheno-5-methylcytosine (ϵ5mC). ϵ5mC is structurally similar to the mutagenic analog 3,N4-ethenocytosine (ϵC), which is repaired by AlkB family enzymes in the direct reversal repair (DRR) pathway and excised by DNA glycosylases in the base excision repair (BER) pathway. However, the repair of ϵ5mC has not been reported. Here, we examined the activities against ϵ5mC by DRR and BER enzymes and TET1-3, enzymes that modify the 5-methyl group in 5mC. We found that the etheno modification of 5mC blocks oxidation by TET1-3. Conversely, three human homologs in the AlkB family, ALKBH2, 3 and FTO were able to repair ϵ5mC to 5mC, which was subsequently modified by TET1 to 5-hydroxymethylcytosine. We also demonstrated that ALKBH2 likely repairs ϵ5mC in MEF cells. Another homolog, ALKBH5, could not repair ϵ5mC. Also, ϵ5mC is not a substrate for BER glycosylases SMUG1, AAG, or TDG. These findings indicate DRR committed by ALKBH2, 3 and FTO could reduce the detrimental effects of ϵ5mC in genetics and epigenetics and may work together with TET enzymes to modulate epigenetic regulations.

Graphical Abstract

Figure 1.

Chemical structure of ϵ5mC and generally accepted mechanisms of TET1-3/ALKBH2/3/FTO-mediated enzymatic reactions on related alkyl-DNA modifications (22,34,37).

Figure 2.

High resolution MS spectra of oligonucleotides containing 5mC and ϵ5mC and oxidations to them by TET1 and ALKBH2. (A) 5mC, (B) 5mC + TET1, (C) ϵ5mC, (D) ϵ5mC + TET1, (E) ϵ5mC + ALKBH2 and (F) ϵ5mC + ALKBH2 + TET1.

Figure 3.

Quantification of etheno deoxyribonucleosides in wild type and ALKBH2 knockout MEF cells exposed to different concentrations of CAA (0.0, 62.5 and 125.0 μM in triplicate) in 0.1% DMSO and analyzed by dynamic MRM analysis on Triple Quadrupole LC/MS. Levels of (A) ϵA and (B) ϵ5mC in cells. The P values were calculated using two-way ANOVA analysis (* P< 0.05, ** P< 0.01, *** P< 0.001 and **** P< 0.0001).

Figure 4.

Assay of DNA glycosylases with known substrates and ϵ5mC. Reactions consisted of 100 nM radiolabeled ds-DNA in a tris buffer at 37°C for 6 h with (+) or without (–) 5 μM DNA glycosylase. Denaturing PAGE analysis shows intact substrate (red boxes) and expected location of excised product (blue boxes). A C:G base pair (lanes 3 and 4 for each enzyme) was used as a negative control.