[gmx-users] Simulation protocol for Protein-DNA-complex

[Justin A. Lemkul]

On 5/25/12 10:01 AM, Matthias Ernst wrote:
> Hi,
>
> I have a question regarding simulation of a protein-DNA-complex where the
> protein encloses the DNA double helix. I did not find a tutorial for a system of
> three rather big molecules like these, that's why I ask. If there is such, I
> would appreciate a hint.
>

In principle, it is no different from the lysozyme tutorial cited below.  The 
workflow starts with pdb2gmx, which can handle all the molecules in the system 
(protein and water) and then you solvate, add ions, and simulate.  There's not 
much very special in this case.

> I want to start with a crystal structure from PDB. When I do the steps in J.
> Lemkuls tutorial "Lysozyme in water", first thing would be an energy
> minimization in vacuo. Unfortunately, doing this I end up with the two strands
> of the DNA double helix being far away from the protein and seperated from each
> other. Can this result from clashes and therefore high energy in the system that
> allows the DNA strands to move "through" the protein or how else can this
> happen? And how can I prevent this?
>

I suspect this is more a result of periodicity than anything else.  Depending on 
how bad the initial geometry is, this in vacuo minimization may not be 
completely necessary.  You can test an in vacuo minimization by using "pbc = no" 
and/or setting an unreasonably large box with:

editconf -f conf.gro -o hugebox.gro -c -d 10

In this instance, with the protein/DNA complex centered in a huge box, there's 
no way you'll get breaks across periodic boundaries during EM.

> I mean, usually the protocol is:
> - minimize system in vacuo
> - add solvent and ions
> - minimize again
> - add thermostat and barostat
> - simulate
> Obviously, I cannot follow this if the first step already does not work. When I
> tried to skip in-vacuo-minimization and to minimize the system in solvent, it
> ended up bei either reaching machine precision without the maximum force being
> small enough or in the simulation, the atom were moving to fast. Would it be a
> good idea to use position restraints for the minimizations? If yes, for which
> part, in which order and in which steps?
>

Position restraints for EM will likely make the outcome worse, as any necessary 
tweaks that may need to happen during EM will be disfavored due to the applied 
restraints.  I never use position restraints during EM for similar systems. 
Perhaps one can think of instances where some restraints might be useful, but I 
don't think this is one of them.

-Justin

-- 
========================================

Justin A. Lemkul, Ph.D.
Research Scientist
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin

========================================