GROMACS version: 2023
GROMACS modification: No
In a molecular dynamics simulation, position restraints are applied to the peptide backbone heavy atoms (4000 kJ/mol/Ų) and lipid phosphate groups (1000 kJ/mol/Ų) during a 1 ns NVT equilibration phase.
Are position restraints used in the NVT phase?
What do you mean specifically? From the text you report, they have been used in those simulations (probably as part of the equilibration phase). Position restraints are commonly used to equilibrate systems (e.g. restraining a protein while you thermalize the box) or to enforce some kind of configuration, e.g. if you are using enhanced sampling methods such as umbrella sampling or restraining some parts of the phase space while you sample others.
Thank you for your reply.
I mean, based on the text above, are position restraints applied during the NVT phase, or during both the NVT and NPT phases?
To me it sounds only during NVT of 1 ns.
The original text is this one: “The membrane was energy minimized and equilibrated in water with 0.15 mol/L of NaCl by running a 250 ps NVT and a 125 ps NPT equilibration with a semi-isotropic pressure scaling, followed by three 500 ps NPT runs with decreasing position restraints on lipid phosphate groups from 1000 to 0 kJ/mol/Å2. Then, the peptide and the lipid bilayer were assembled in a rectangular box, solvated, energy minimized, and equilibrated with 1 ns NVT and 1 ns NPT runs by decreasing the position restraints on peptide backbone heavy atoms and lipid phosphate groups from 4000 and 1000 to 50 and 0 kJ/mol/Å2, respectively.”
Would you please explain more?
This is a relatively standard way to equilibrate a membrane/protein. The idea is to apply position restraints to specific parts of the system to help the rest of it equilibrate and, finally, remove these restraints and let the full system evolve freely. Regarding lipids, what they are doing is putting a restraint on the lipid head groups (phosphate atom) so that they are kept planar; regarding the protein, they are restraining the heavy atoms of the backbones so that the side chains can equilibrate/reorient before letting the whole protein “move”. Usually these processes are done gradually, i.e., you start from harder spring constants (4000 kJ/mol/Å2) and gradually decrease them to zero during a few NVT/NPT simulations.
This is what the paper is summarizing. From what I see, they are running 1ns NVT with restraint on backbone atoms at 4000 kJ/mol/Å2 and 1000 kJ/mol/Å2 on lipid phosphates and then an NPT of 1ns with 50 kJ/mol/Å2 on the protein backbone and 0 kJ/mol/Å2 (so no restraints) on the lipid heads.
Thanks a lot for your response.