Thermal and mechanical properties of a DNA model with solvation barrier

We study the thermal and mechanical behavior of DNA denaturation in the frame of the mesoscopic Peyrard- Bishop-Dauxois model with the inclusion of solvent interaction. By analyzing the melting transition of a homogeneous A-T sequence, we are able to set suitable values of the parameters of the model and study the formation and stability of bubbles in the system. Then, we focus on the case of the P5 promoter sequence and use the Principal Component Analysis of the trajectories to extract the main information on the dynamical behavior of the system. We find that this analysis method gives an excellent agreement with previous biological results.

💡 Research Summary

In this work the authors extend the mesoscopic Peyrard‑Bishop‑Dauxois (PBD) model of DNA by incorporating an explicit solvation barrier that mimics the interaction of the double helix with surrounding water molecules. The motivation is to capture the influence of the solvent on both the thermal denaturation (melting) and the mechanical response of DNA, phenomena that are poorly described by the original PBD model which treats the base‑pair interaction in isolation.

The study proceeds in two main stages. First, a homogeneous adenine‑thymine (A‑T) sequence is simulated over a range of temperatures. By comparing the simulated melting curve with experimental melting temperatures for A‑T rich DNA, the authors calibrate three key parameters: the depth of the Morse potential (D), the stiffness of the nonlinear stacking interaction (k), and the height of the solvation barrier (V_solv). The optimal set (D ≈ 0.04 eV, k ≈ 0.02 eV Å⁻², V_solv ≈ 0.15 eV) reproduces the experimentally observed melting temperature and yields a realistic temperature‑dependent opening probability.

With the calibrated model, the authors then examine bubble formation. They define a bubble as a contiguous stretch of base pairs whose transverse displacement exceeds a threshold (≈1 Å). The statistical analysis shows that, near the melting point, bubbles appear spontaneously but remain relatively short‑lived. The presence of the solvation barrier reduces both the average bubble size and its lifetime, indicating that water molecules act as a mechanical buffer that hinders the complete rupture of hydrogen bonds.

The second stage focuses on a biologically relevant sequence: the P5 promoter, a ~100‑base‑pair region that contains the −10 and −35 elements critical for transcription initiation. Long‑time (∼500 ns) molecular dynamics simulations are performed at physiological temperature (300 K) without external forces. The trajectory data (base‑pair displacements sampled every 0.1 ps) constitute a high‑dimensional time series. To extract the dominant collective motions, the authors apply Principal Component Analysis (PCA) to the covariance matrix of the displacements.

The PCA reveals that the first principal component accounts for a global “breathing” mode—simultaneous expansion and contraction of the entire duplex—while the second component isolates a localized opening mode. Strikingly, the localized mode is concentrated around the −10 and −35 boxes, precisely the sites where transcription factors bind in vivo. This spatial correlation matches earlier experimental observations from DNase I footprinting and transcription assays, confirming that the model captures functionally relevant dynamics.

A comparative analysis with the original PBD model (without the solvation barrier) shows that the latter fails to reproduce the site‑specific bubble pattern; bubble formation is more random and the correlation with promoter elements is weak. The inclusion of the solvation barrier therefore improves the predictive power of the mesoscopic model, highlighting the essential role of solvent in shaping DNA’s mechanical landscape.

In conclusion, the paper demonstrates that a modest modification of the PBD Hamiltonian—adding a solvation barrier—allows a mesoscopic description to bridge the gap between purely physical models and biologically meaningful behavior. The combination of calibrated thermodynamic parameters, bubble statistics, and PCA of long trajectories provides a comprehensive framework for studying DNA stability, bubble dynamics, and sequence‑specific mechanical responses. The authors suggest future extensions such as sequence‑dependent parameterization, external mechanical loading (e.g., stretching or twisting), and large‑scale simulations of multiple promoters to further elucidate DNA–protein interaction mechanisms.