Each of the five models was inserted into a heterogeneous neuronal plasma membrane (Table 1) using the CHARMM-GUI membrane builder [19,20]. A box of 250 Å × 250 Å was established with periodic boundary conditions. The system was solvated with 150 mM NaCl at a net charge of zero using the TIP3P water model [21]. The simulations were run in GROMACS [22] using the University of Cambridge High Performance Computing Resources. Long-range electrostatic interactions were calculated using the Particle-mesh Ewald method with the Coulomb and van der Waals interaction cut-offs set to 12 Å [23]. The LINCS algorithm was used to constrain molecular bonds. All systems were run using the CHARMM36m force field. Following steepest descent minimisation, all systems were subjected to six series of a 125 picosecond NPT equilibration ensemble with temperature coupling with velocity rescaling and pressure coupling using the Parrinello–Rahman method [24,25,26] for all simulations. Simulation time steps were set at 2 fs, and, as RMSDs plateaued within 10 ns for all models, total simulation runtimes were 50 ns (receptor models) or 25 ns (docked receptor models); simulations were run 3–5 times for each model.
Lipid composition of the neuronal plasma membrane.
Do you have any questions about this protocol?
Post your question to gather feedback from the community. We will also invite the authors of this article to respond.