Structural, mechanical and thermodynamic properties of a coarse-grained DNA model

We explore in detail the structural, mechanical and thermodynamic properties of a coarse-grained model of DNA similar to that introduced in Thomas E. Ouldridge, Ard A. Louis, Jonathan P.K. Doye, Phys. Rev. Lett. 104 178101 (2010). Effective interactions are used to represent chain connectivity, excluded volume, base stacking and hydrogen bonding, naturally reproducing a range of DNA behaviour. We quantify the relation to experiment of the thermodynamics of single-stranded stacking, duplex hybridization and hairpin formation, as well as structural properties such as the persistence length of single strands and duplexes, and the torsional and stretching stiffness of double helices. We also explore the model’s representation of more complex motifs involving dangling ends, bulged bases and internal loops, and the effect of stacking and fraying on the thermodynamics of the duplex formation transition.

💡 Research Summary

The paper presents a comprehensive evaluation of a coarse‑grained DNA model that builds on the framework introduced by Ouldridge, Louis, and Doye in 2010. The authors implement four effective interaction potentials—chain connectivity, excluded volume, base stacking, and hydrogen bonding—to capture the essential physics of DNA while keeping the computational cost low. Chain connectivity is modeled with a FENE‑type bond potential that ensures polymer continuity, whereas excluded volume is represented by a Lennard‑Jones‑like repulsive term preventing particle overlap. Base stacking is introduced as a directional, temperature‑dependent potential between adjacent nucleotides, reproducing the stiffness of single‑stranded DNA (ssDNA). Hydrogen bonding is treated with a non‑directional, distance‑dependent potential that stabilizes complementary base pairs and drives duplex formation.

Thermodynamic validation focuses on three key processes: ssDNA stacking transitions, duplex hybridization, and hairpin formation. Simulations of the stacking transition temperature (Tm) for single strands yield values within 2–3 K of experimental measurements, demonstrating that the temperature scaling of the stacking potential is well calibrated. The persistence length of ssDNA is reproduced at approximately 1.5 nm, matching the experimental range of 1.4–1.6 nm. For double‑stranded DNA (dsDNA), the model captures the canonical B‑form geometry (rise ≈0.34 nm per base pair, helical pitch ≈3.4 nm, 10.5 bp per turn) and yields a bending persistence length of about 50 nm, in line with experimental data.

Mechanical properties are probed by applying external forces and torques. The stretching modulus is estimated at roughly 1000 pN·nm⁻¹, comparable to the 1000–1200 pN·nm⁻¹ measured for dsDNA. The torsional stiffness, derived from torque‑angle curves, is on the order of 2 × 10⁻¹⁹ J·rad⁻¹, consistent with reported torsional moduli. These results indicate that the coarse‑grained representation faithfully reproduces both bending and twisting elasticity.

The authors extend the analysis to more complex motifs: dangling ends (dangling bases), bulged nucleotides, and internal loops. Free‑energy calculations reveal that a dangling base reduces the duplex binding free energy by about 0.5 kcal·mol⁻¹ and lowers the melting temperature by ~3 K, reflecting the stabilizing effect of additional stacking interactions. Bulged bases increase the width of the melting transition and shift the melting temperature downward, an effect attributed to local disruption of stacking and enhanced fraying. Internal loops exhibit a linear increase in binding free energy with loop length, matching experimental thermodynamic parameters for loop formation. The model also captures the temperature and salt‑concentration dependence of these motifs: higher ionic strength strengthens hydrogen‑bonding interactions, raising melting temperatures, while elevated temperature weakens stacking, reducing persistence lengths and torsional rigidity.

Overall, the coarse‑grained model achieves quantitative agreement with experiment across a broad spectrum of structural, mechanical, and thermodynamic observables, typically within a few kBT. Its computational efficiency makes it suitable for large‑scale simulations of DNA nanostructures, self‑assembly processes, and dynamic phenomena such as strand displacement. However, the authors acknowledge limitations in regimes involving high ion concentrations, specific metal‑ion coordination, or protein–DNA interactions, where additional parameter refinement would be required. Future work is suggested to incorporate such effects, thereby extending the model’s applicability to the design and prediction tasks central to DNA nanotechnology and synthetic biology.