Exploring the Self-Assembly of Glycolipids into Three-Dimensional Network Phases

Glycolipids are sugar-based amphiphiles that play crucial roles in many biological processes. Under thermotropic and lyotropic conditions, glycolipids self-assemble into a variety of mesophases, including cocontinuous network phases, such as the double gyroid. In this work, a two-stage molecular dynamics simulation workflow is developed to probe network formation for solvent-free amphiphiles at different temperatures. In the first stage, the structural evolution of systems initiated in lamellar and hexagonally packed cylinder arrangements is examined as an indicator of the likelihood of network formation. In the second stage, initial configurations for four network phases are obtained by applying an external guiding field to overcome any nucleation barriers, and the stability of these network phases after switching off the guiding field is investigated. A convolutional neural network is trained and applied to assign a morphology to each snapshot from the simulation. Three anomerically pure and water-free glycolipids (2-decyl-tetradecyl-d-maltoside, 2-decyl-tetradecyl-d-cellobioside, and 2-decyl-tetradecyl-d-galactoside) were examined in this study, and the predicted phase diagrams show good alignment with experimental observations from differential scanning calorimetry and temperature-dependent small-angle X-ray scattering. The workflow presents a facile approach to probe network phase stability and paves the way for the discovery of new network-forming amphiphiles.