The amphipathic nature of the lipid molecule (hydrophilic head and hydrophobic tails) enables it to act as a barrier between fluids with various properties and to sustain an environment where the processes critical to life may proceed. While computer simulations of biomolecules primarily investigate protein conformation and binding to drug-like molecules, these interactions often occur in the context of a lipid membrane. Chemical specificity of lipid models is essential to accurately represent the complex environment of the lipid membrane. This review discusses the development and performance of currently used chemically specific lipid force fields (FF) such as the CHARMM, AMBER, GROMOS, OPLS, and MARTINI families. Considerations in lipid FF development including lipid diversity, temperature dependence, phase behavior, and effects of atomic polarizability are considered, as well as methods and goals of parametrization. Applications of these FFs to complex and diverse models for cellular membranes are summarized. Lastly, areas for future development, such as efficient inclusion of long-range Lennard–Jones interactions (significant in transitions from polar to apolar media), accurate transmembrane dipole potential, and diffusion under periodic boundary conditions are considered.

The barrier function of the stratum corneum (SC) is intimately related to the structure of the lipid matrix, which is composed of ceramides (Cer), cholesterol (Chol), and free fatty acid (FFA). In this study, the all-atom CHARMM36 (C36) force field is used to simulate bilayers of N-palmitoylsphingosine (Cer16), N-lignoceroylsphingosine (Cer24), Chol, and lignoceric acid (LA) as simple models of the SC. Equimolar mixtures of Cer, Chol, and LA are replicated from experiment for comparison and validation of the C36 force field, and the effects of lipid diversity and temperature are studied. The presence of Chol and LA have effects on nearly all membrane properties including surface area per lipid, area compressibility moduli, chain order, Chol tilt, bilayer thickness, interdigitation, hydrogen bonding, and lipid clustering, while temperature has a more moderate effect. In systems containing Cer16, there is a profound difference in interdigitation between pure Cer and mixed systems, while systems containing Cer24 are relatively unaffected. Increasing temperature has the potential to shift hydrogen bonding pairs rather than uniformly decrease bonding, which can lead to greater Cer-Cer bonding at higher temperatures. Comparison with deuterium order parameter experiments demonstrates good agreement, which supports further use of this class of lipids and fatty acids for development of more complex SC models.

Permeation of many small molecules through lipid bilayers can be directly observed in molecular dynamics simulations on the nano-and microsecond timescale. While unbiased simulations provide an unobstructed view of the permeation process, their feasibility for computing permeability coefficients depends on various factors that differ for each permeant. The present work studies three small molecules for which unbiased simulations of permeation are feasible within less than a microsecond, one hydrophobic (oxygen), one hydrophilic (water), and one amphiphilic (ethanol). Permeabilities are computed using two approaches: counting methods and a maximumlikelihood estimation for the inhomogeneous solubility diffusion (ISD) model. Counting methods yield nearly model-free estimates of the permeability for all three permeants. While the ISD-based approach is reasonable for oxygen, it lacks precision for water due to insufficient sampling, and results in misleading estimates for ethanol due to invalid model assumptions. It is also demonstrated that simulations using a Langevin thermostat with collision frequencies of 1/ps and 5/ps yield oxygen permeabilities and diffusion constants that are lower than those using Nose-Hoover by statistically significant margins. In contrast, permeabilities from trajectories generated with Nosé-Hoover and the microcanonical ensemble do not show statistically significant differences. As molecular simulations become more affordable and accurate, calculation of permeability for an expanding range of molecules will be feasible using unbiased simulations. The present work summarizes theoretical underpinnings, identifies pitfalls and develops best practices for such simulations.

The all-atom CHARMM36 (C36) force field is used to simulate bilayers of pure palmitoylsphingomyelin (PSM) as well as binary mixtures of PSM and stearoylsphingomyelin (SSM) at various cholesterol concentrations (X) and temperatures. C36 simulation data is in good agreement with experimental deuterium order parameters and previous computational results, providing evidence of the utility of the force field for potentially studying more complex membranes. The area compressibility modulus is shown to achieve a large value of 2.82 ± 0.08 N/m in cholesterol-rich membranes (X = 0.50). Surface area per lipid (SA/lip), tilt angle, membrane thicknesses, and acyl chain ordering are shown to have strong dependencies on cholesterol concentration. Relaxation times also indicate cholesterol dependence and show a strong preference for rotational axial motion over wobbling motion. Radial distribution functions and lipid clustering indicate strong relationships between lateral ordering and hydrogen bonding, which is long lived in SM membranes. These interactions lead to strong self-association of cholesterol at high concentrations, causing shielding from further SM-cholesterol interactions. The importance of a ternary component on SM-SM hydrogen bonds is revealed in light of previous results and is consequential in the modeling of lipid rafts.

Recent studies in lipid raft formation and stratum corneum permeability have focused on the role of ceramides (CER). In this study, we use the all-atom CHARMM36 (C36) force field to simulate bilayers using N-palmitoylsphingosine (CER16) or α-hydroxy-N-stearoyl phytosphingosine (CER[AP]) in 1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC) or 1-palmitoyl-2-oleoylphosphatidylcholine (POPC), which serve as general membrane models. Conditions are replicated from experimental studies for comparison purposes, and concentration (X) is varied to probe the effect of CER on these systems. Comparisons with experiment based on deuterium order parameters and bilayer thickness demonstrate good agreement, thus supporting further use of the C36 force field. CER concentration is shown to have a profound effect on nearly all membrane properties including surface area per lipid, chain order and tilt, area compressibility moduli, bilayer thickness, hydrogen bonding, and lipid clustering. Hydrogen bonding in particular can significantly affect other membrane properties and can even encourage transition to a gel phase. Despite CER's tendency to condense the membrane, an expansion of CER lipids with increasing X is possible depending on how the balance between various hydrogen-bond pairs and lipid clustering is perturbed. Based on gel phase transitions, support is given for phytosphingosine's role as a hydrogen-bond bridge between sphingosine ordered domains in the stratum corneum.