Ion–Lipid Interactions in Biological Membranes: Insights from Combined Molecular Dynamics and Quantum Chemical Calculations

Understanding ion-lipid interactions at biomembrane interfaces is fundamental to deciphering biological processes and designing biomimetic systems. While classical MD simulations provide valuable insights into ion-lipid interactions, they sometimes fail to reproduce experimental observations due to the limitations inherent in the force field accuracy. In this study, we employ high-level ab initio calculations along with molecular dynamics (MD) simulations to elucidate the interplay between cations (Na+, K+, Ca2+, and Mg2+), as well as Cl- counterions, with phosphatidylcholine lipid head groups. Optimized configurations reveal that Na+ and K+ preferentially interact with phosphatic oxygen (OP) rather than carbonyl oxygen (OC), whereas Ca2+ and Mg2+ exhibit dual-site binding at OP and OC. Notably, all cations except Mg2+ bind directly to the lipid head groups by shedding hydration water. Due to its high hydration energy, Mg2+ retains its hexahydrated state, interacting indirectly with lipid head groups via a bridging water molecule. Similarly, Cl- avoids direct interactions with choline groups and instead binds via water-mediated bridges. Energetic analysis at the DLPNO-CCSD(T)/cc-pVTZ//B3LYP-D3/6-31G** level of theory reveals that Na+ exhibits interaction energies significantly weaker than those of Ca2+, aligning well with experimental observations, whereas classical MD often overestimates the binding affinity of Na+. A trend of increasing binding strength with lipid cluster size is observed, with divalent cations displaying a biphasic binding trend, where the largest increase in interaction energy occurs between monomers and dimers, suggesting near-saturation of binding sites in dimeric lipid clusters. This observation aligns with experimental findings that Ca2+ preferentially bridges two lipid headgroups, a feature often misrepresented in classical force fields. These findings provide crucial insights into ion-lipid interactions in biological membranes, highlighting the limitations of classical force fields and emphasizing the need to incorporate ab initio-derived parameters to improve the accuracy of MD simulations.