Related Experiment Video
Updated: Apr 24, 2026

Immunostaining for DNA Modifications: Computational Analysis of Confocal Images
Published on: September 7, 2017
Quantum tunneling in DNA methylation: A computational study using QM/MM and molecular dynamics
Justin Sato1, Annabelle Choi1, Peace U Clement2
1California Academy of Math and Science, Carson, CA, USA; Department of Physics, California State University Dominguez Hills, Carson, CA, USA.
Abstract:
DNA methylation is a central epigenetic modification regulating gene expression, chromatin structure, and disease progression. Although commonly treated as a thermally activated enzymatic process, quantum tunneling can contribute to methyl transfer in DNA methyltransferases (DNMTs). We present a multiscale computational study combining density functional theory (DFT), hybrid quantum mechanics/molecular mechanics (QM/MM), and tunneling formalisms-Wentzel-Kramers-Brillouin (WKB) analysis as a baseline and ring-polymer instanton (RPI) theory with instanton-Hessian refinement-applied to a Morse-fitted reaction coordinate. Near-well curvature and width are extracted from an enzyme-aligned scan and used to interpret bead-converged RPI actions and kinetic isotope effects (KIEs). In the absence of enzymatic facilitation, WKB predicts negligible tunneling probabilities and astronomically long timescales, underscoring the necessity of a quantum mechanical treatment that includes the catalytic environment. Near the minimum, a quadratic fit gives ξe = -0.018 Å with curvature k = 460.4 kcal·mol-1·Å-2 (RMSE: 4.45 kcal, N = 12). A Morse fit with fixed De = 150 kcal·mol-1 yields a = 1.03 Å-1 and re = -0.09 Å (RMSE: 4.6-8.8 kcal depending on weighting), consistent with a moderately steep, slightly displaced well. With DNMT1 preorganization at 298 K, instanton + Hessian analysis of the pre-methylation geometry yields kCH3 ≈ 6.7 × 10-3 s-1 and kCD3 ≈ 8.2 × 10-7 s-1, corresponding to a large KIE (≈8.17 × 103; lnKIE ≈ 9). For the methylated state, RPI calculations (32-128 beads) converge to a modest KIE of 6.28-6.32 (lnKIE ≈ 1.84), with anchor-qualified rates kCH3 ≈ (6.1-6.7) × 10-3 s-1 and kCD3 ≈ (1.0-1.1) × 10-3 s-1. These results indicate that DNMT1 reshapes and narrows the barrier to enable methyl tunneling during the chemical step, while product-like geometries suppress isotope sensitivity after mark installation. Relative quantum observables are robust to prefactor choice.

