A Semiempirical Quantum Model for Hydrogen-Bonded Nucleic Acid Base Pairs

Area of Science:

• Computational Chemistry
• Biophysics
• Molecular Modeling

Background:

• Accurate modeling of hydrogen bonding in nucleic acids is crucial for understanding DNA/RNA structure and function.
• Existing semiempirical methods often lack the precision required for biological interactions.
• High-level quantum chemical calculations are computationally expensive for large systems.

Purpose of the Study:

• To develop a novel semiempirical Hamiltonian, PM3BP, specifically for modeling hydrogen bonding in nucleic acid base pairs.
• To achieve computational efficiency comparable to traditional semiempirical methods while maintaining high accuracy.
• To provide a cost-effective tool for studying biological interactions.

Main Methods:

• Developed PM3BP by reparametrizing the PM3 Hamiltonian using nonlinear optimization algorithms.
• Interfaced the new Hamiltonian with a d-orbital semiempirical program.
• Validated PM3BP against experimental data and benchmark density-functional theory (DFT) calculations for nucleic acid dimers and trimers.

Main Results:

• PM3BP accurately reproduces experimental base pair dimer enthalpies and DFT results.
• Outperforms AM1, PM3, MNDO, and MNDO/H in modeling dimer and trimer structures and interaction enthalpies.
• Achieves accuracy rivaling DFT for dimer interaction enthalpies at a significantly reduced computational cost (over 3 orders of magnitude faster).

Conclusions:

• PM3BP offers a significant advancement in modeling hydrogen bonding in nucleic acids.
• Provides a balance between accuracy for biological interactions and computational efficiency.
• Highlights the limitations of conventional semiempirical methods for complex biological systems.