[gmx-users] Deformation of monomer in pulling simulation

[Robert Dole]

Hello all,

I'm trying to pull two peptides (coordinates are from their dimer crystal structure) apart using the distance option for the pull geometry and harmonic potential using GMX 4.5.3 using OPLS-AA to define the protein and tip3p for the water model with NaCl used @ 140 nM for ions. The box is 20 nm long and the pulling is done along the longest axis of the cuboid (i.e., Z-direction) such that the peptide move much less than half the periodic distance of the box over the course of the simulation. I've run the pulling simulation at 0.0005 and 0.001 nm/ps and with pull_k1 set to 1000 or 5000 to qualitatively see whether a higher force constant made a difference (which is doesn't).

As the simulation runs, there is a large deformation of the monomer being pulled in which the pulled peptide (initially looks like a "V" with the hydrophobic core in the "middle") is stretched. The hydrophobic N-terminal stays very close to the reference peptide for the majority of the simulation until it finally comes "off" towards the end of the pull run.

I've gone through some of the literature and it seems as though this is an expected phenomenon (i.e., deformation of the pulled structure), but the deformation is so severe in my case, that it loses ~25-35% of its secondary structure! Any suggestions as to how I can remedy this? I thought of using distance restraints on the pulled peptide for the pulling simulation after which I would remove them for simulations of the windows, but I thought I would ask before trying this as I've nearly reached my monthly quota on the HPC cluster I work on.

Thanks,

Rob

GROMPP commands:
grompp -f PULL.mdp -c NPT.pdb -p topol.top -n index.ndx -t NPT.trr -o PULL.tpr

MDP options:
integrator    = md        ; leap-frog integrator
nsteps        = 5000000    ; 10 ns
dt        = 0.002        ; 2 fs
.
.
.
pull            = umbrella    ; notes to self
pull_geometry   = distance      ; simple distance increase 
pull_dim        = N N Y        ; pull peptide B away fromAB in the Z direction (the long side of the cuboid)
pull_start      = yes           ; 
pull_ngroups    = 1        ; only one group is getting pulled
pull_group0     = Chain_A 
pull_group1     = Chain_B 
pull_rate1      = 0.0005          ; 0.5 nm per ns = 5 A per nanosecond 
pull_k1         = 5000          ; run 1 - 1000, run 2 - 5000
 		 	   		  
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20110918/d08f4fc1/attachment.html>