How Human TET2 Enzyme Catalyzes the Oxidation of Unnatural Cytosine Modifications in Double-Stranded DNA

Abstract

Methylation of cytosine bases is strongly linked to gene expression, imprinting, aging, and carcinogenesis. The Ten-eleven translocation (TET) family of enzymes, which are Fe(II)/2-oxoglutarate (2OG)-dependent enzymes, employ Fe(IV)=O species to dealkylate the lesioned bases to an unmodified cytosine. Recently, it has been shown that the TET2 enzyme can catalyze promiscuously DNA substrates containing unnatural alkylated cytosine. Such unnatural substrates of TET can be used as direct probes for measuring the TET activity or capturing TET from cellular samples. Herein, we studied the catalytic mechanisms during the oxidation of the unnatural C5-position modifications (5-ethylcytosine (5eC), 5-vinylcytosine (5vC) and 5-ethynylcytosine (5eyC)) and the demethylation of N4-methylated lesions (4-methylcytosine (4mC) and 4,4-dimethylcytosine(4dmC)) of the cytosine base by the TET2 enzyme using molecular dynamics (MD) and combined quantum mechanics and molecular mechanics (QM/MM) computational approaches. The results reveal that the chemical nature of the alkylation of the double-stranded (ds) DNA substrates induces distinct changes in the interactions in the binding site, the second coordination sphere, and long-range correlated motions of the ES complexes. The rate-determining hydrogen atom transfer (HAT) is faster in N4-methyl substituent substrates than in the C5-alkylations. Importantly, the calculations show the preference of hydroxylation over desaturation in both 5eC and 5vC substrates. The studies elucidate the post-hydroxylation rearrangements of the hydroxylated intermediates of 5eyC and 5vC to ketene and 5-formylmethylcytosine (5fmC), respectively, and hydrolysis of hemiaminal intermediate of 4mC to formaldehyde and unmodified cytosine proceed exclusively in aqueous solution outside of the enzyme environment. Overall, the studies show that the chemical nature of the unnatural alkylated cytosine substrates exercises distinct effects on the binding interactions, reaction mechanism, and dynamics of TET2.

Keywords: DNA repair; QM/MM calculations; TET2 enzyme; demethylation; molecular dynamics; non-heme enzymes; reaction mechanism.

Figure 1:

(a) Protein structure for the human TET2-dsDNA complex derived from the average molecular dynamics (MD) structure of ferryl intermediate simulations with the 5eC substrate. (b) Overlaid structure of the C5-alkylated 5eC, 5vC, and 5eyC unnatural substrates with the natural 5mC substrate. (c) Overlaid structure of the N4-methylated 4mC, 4dmC substrates with the natural C5-methylated 5mC substrate. In both (b) and (c), the Fe center depicts the one of 5mC dsDNA-bound TET2.

Figure 2:

Principal component analysis (a) and dynamic cross correlation (b) of the ferryl complex of TET2 bound to 5eC dsDNA substrate. Residues numbers are as follows: 1–445 (TET protein), 446–448 (Zn), 449 (Fe), 450 (O), 451 (succinate), and 452–475 (DNA). In part (b), residues numbers range 6–17, 157–181, 332–363, and 452–472 on both axes denote the Cys-N, L2, GS-linker, and DNA, respectively. Yellow to blue represents the direction of motion of residues in part (a).

Figure 3:

Principal component analysis of the ferryl complex of TET2 bound to 4mC dsDNA (a) and 4dmC (b) substrates. Residues numbers are as follows: 1–445 (TET protein), 446–448 (Zn), 449 (Fe), 450 (O), 451 (succinate), and 452–475 (DNA). Yellow to blue represents the direction of motion of residues.

Figure 4:

QM/MM potential energy profile for the hydroxylation and desaturation reactions of 5eC dsDNA substrate by TET2, calculated using UB3LYP/def2-TZVP with ZPE. The relative energies are in kcal/mol.

Figure 5:

The transition states structures obtained during hydroxylation and desaturation processes of 5eC dsDNA by TET2. Distances (Å), the spin densities are in black and red, respectively, while the Fe—O—H and C—H—O angles are in degrees.

Figure 6:

Possible reaction pathways for the HAT by the ferryl complex.

Figure 7:

Energy decomposition analysis (EDA) of the residues stabilizing the transition states and the products.

Figure 8:

QM/MM potential energy profile for the hydroxylation and desaturation reactions of 5vC (a) and for the hydroxylation of 5eyC (b) dsDNA substrates by TET2, calculated using UB3LYP/def2-TZVP with ZPE. The relative energies are in kcal/mol.

Figure 9:

QM/MM potential energy profiles for the formation of ketene inside the enzyme without the assistance of water molecule (a) and the ketene formation with the assistance of water (b) inside (green) and outside (black) of the enzyme. The relative energies and the distances are in kcal/mol and Å, respectively.

Figure 10:

Reaction scheme and the QM/MM potential energy profile for the decomposition of hemiaminal intermediate of 4mC hydroxylation inside the enzyme, calculated using UB3LYP/def2-TZVP with ZPE. The relative energies are in kcal/mol.

Scheme 1:

Possible reaction mechanisms pathways explored for the oxidation of the unnatural 5eC, 5vC, and 5eyC dsDNA substrates by TET2.

Scheme 2:

Reaction mechanisms for the hydroxylation of the unnatural 4mC and 4dmC dsDNA substrates by TET2.

Scheme 3:

Reaction scheme for a) formation of ketene inside the enzyme without the assistance of water molecule; b) formation of ketene with the assistance of water inside the enzyme; c) formation of ketene with the assistance of water outside the enzyme in the water solvent; d) formation of 5-formylmethylcytosine (5fmC) outside the enzyme in the water solvent.