Optimising a workflow to study the permeation of small molecules through biological membranes using molecular simulation
File(s)
Author(s)
Jabeen, Saher
Type
Thesis
Abstract
ABSTRACT
This study presents an optimised computational workflow to elucidate the molecular mechanisms governing small-molecule permeation across lipid membranes via free energy profiling. Steered molecular dynamics (SMD) and umbrella sampling (US) simulations were employed to calculate free energy profiles for representative permeants (lactic acid, benzoic acid, and toluene) within a 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC) bilayer using all-atom simulations with the CHARMM36M force field. Free energy reconstruction from biased simulations was achieved using the weighted histogram analysis method (WHAM). The workflow was further validated using coarse-grained Martini 3 simulations of the DOPC–toluene system, demonstrating good agreement with literature data across both atomistic and coarse-grained resolutions.
The optimised protocol was subsequently applied to increasingly complex Gram-negative bacterial membrane models: (1) lipid A, (2) lipid A with core sugars, and (3) an asymmetric Escherichia coli outer membrane model. Initial simulations of symmetric models revealed anomalous shifts in free energy barriers, attributed to membrane perturbations arising from SMD initialisation. Refinement of the pulling protocol, specifically reducing the pulling rate to 0.01 nm ns⁻¹, mitigated these artefacts and yielded improved free energy profiles.
In contrast, the asymmetric membrane system exhibited greater sensitivity to simulation parameters, reflecting its structural complexity. While slower pulling velocities reduced nonphysical energy contributions, results remained inconsistent with simpler models. Notably, reconstruction of independent free energy profiles improved reproducibility, indicating that alternative sampling strategies may be required for heterogeneous membranes.
Overall, the workflow reliably reproduces permeation energetics in simple phospholipid bilayers at multiple resolutions; however, its application to complex bacterial membranes highlights the need for further methodological refinement.
This study presents an optimised computational workflow to elucidate the molecular mechanisms governing small-molecule permeation across lipid membranes via free energy profiling. Steered molecular dynamics (SMD) and umbrella sampling (US) simulations were employed to calculate free energy profiles for representative permeants (lactic acid, benzoic acid, and toluene) within a 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC) bilayer using all-atom simulations with the CHARMM36M force field. Free energy reconstruction from biased simulations was achieved using the weighted histogram analysis method (WHAM). The workflow was further validated using coarse-grained Martini 3 simulations of the DOPC–toluene system, demonstrating good agreement with literature data across both atomistic and coarse-grained resolutions.
The optimised protocol was subsequently applied to increasingly complex Gram-negative bacterial membrane models: (1) lipid A, (2) lipid A with core sugars, and (3) an asymmetric Escherichia coli outer membrane model. Initial simulations of symmetric models revealed anomalous shifts in free energy barriers, attributed to membrane perturbations arising from SMD initialisation. Refinement of the pulling protocol, specifically reducing the pulling rate to 0.01 nm ns⁻¹, mitigated these artefacts and yielded improved free energy profiles.
In contrast, the asymmetric membrane system exhibited greater sensitivity to simulation parameters, reflecting its structural complexity. While slower pulling velocities reduced nonphysical energy contributions, results remained inconsistent with simpler models. Notably, reconstruction of independent free energy profiles improved reproducibility, indicating that alternative sampling strategies may be required for heterogeneous membranes.
Overall, the workflow reliably reproduces permeation energetics in simple phospholipid bilayers at multiple resolutions; however, its application to complex bacterial membranes highlights the need for further methodological refinement.
Version
Open Access
Date Issued
2024-12-23
Date Awarded
2026-05-01
Copyright Statement
Attribution-NonCommercial 4.0 International Licence (CC BY-NC)
License URL
Advisor
Müller, Erich
Jackson, George
Sponsor
Imperial College London
Publisher Department
Department of Chemical Engineering
Publisher Institution
Imperial College London
Qualification Level
Doctoral
Qualification Name
Master of Philosophy (MPhil)