Thymine DNA glycosylase recognizes the geometry alteration of minor grooves induced by 5-formylcytosine and 5-carboxylcytosine

Abstract

The dynamic DNA methylation-demethylation process plays critical roles in gene expression control and cell development. The oxidation derivatives of 5-methylcytosine (5mC) generated by Tet dioxygenases in the demethylation pathway, namely 5-hydroxymethylcytosine (5hmC), 5-formylcytosine (5fC), and 5-carboxylcytosine (5caC), could impact biological functions by altering DNA properties or recognition by potential reader proteins. Hence, in addition to the fifth base 5mC, 5hmC, 5fC, and 5caC have been considered as the sixth, seventh, and eighth bases of the genome. How these modifications would alter DNA and be specifically recognized remain unclear, however. Here we report that formyl- and carboxyl-modifications on cytosine induce the geometry alteration of the DNA minor groove by solving two high-resolution structures of a dsDNA decamer containing fully symmetric 5fC and 5caC. The alterations are recognized distinctively by thymine DNA glycosylase TDG via its finger residue R275, followed by subsequent preferential base excision and DNA repair. These observations suggest a mechanism by which reader proteins distinguish highly similar cytosine modifications for potential differential demethylation in order to achieve downstream biological functions.

Fig. 1. Schematic diagram of 5fC-dsDNA and 5caC-dsDNA crystal structures. (a) Overall structure of 5fC-dsDNA. Two complementary strands are colored in green and cyan, respectively. The fully symmetric 5fC bases are shown as sticks and their C5m atoms are shown as purple spheres. The distance between two C5m atoms is shown as yellow dashes. (b) The hydrogen bond networks around 5fC base pairing. The water molecules are shown as red dots, and the carbonate ions are shown as magenta sticks. The hydrogen bonds are shown as yellow dashes. (c) The overall structure of 5caC-dsDNA. (d) The hydrogen bond networks around 5caC base pairing.

Fig. 2. Comparison of the base-step and groove rigid body parameters of 5C-, 5mC-, 5hmC-, 5fC- and 5caC-dsDNA with canonical A-form and B-form dsDNA. (a–c) The comparison of slide, shift, and roll base pair parameters. (d, e) The major and minor groove width comparison.

Fig. 3. The dihedral angle rotation analysis against 5mC (a), 5hmC (b), 5fC (c), and 5caC (d). The modified bases are shown as sticks, and atoms in the modified group are labeled. The potential energy calculated during O5/H-C5m-C5-C6 dihedral angle rotation is shown as curves.

Fig. 4. Comparison of the structural parameters of 5C-, 5mC-, 5hmC-, 5fC- and 5caC-dsDNA with canonical A-form and B-form dsDNA from molecular dynamics simulation. (a–c) The comparison of slide, shift, and roll base pair parameters. (d and e) The major and minor groove width comparison. (f and g) The electric potential energy of the minor groove and the midpoint of C5-P (phosphate in the backbone).

Fig. 5. Schematic diagram of TDG glycosylase activity assay and single turnover assay results. (a) The glycosylase activity assay of wild type and R275A mutants against various substrates containing uracil, thymine, 5fC, and 5caC, respectively. (b) The single turnover assay against 5fC or 5caC containing substrates (representative of three independent experiments).