[gmx-users] Free Energy of Binding Question

Hannes Loeffler Hannes.Loeffler at stfc.ac.uk
Tue Sep 20 15:57:09 CEST 2016

On Tue, 20 Sep 2016 13:27:06 +0000
Abdülkadir KOÇAK <kocak at gtu.edu.tr> wrote:

> What I am doing is exactly as you said; decoupling the ligand once
> from the WT and once from the MUT. And now I see that I am getting
> "absolute" free energies in each case. I did not run MD for the
> proteins in the absence of the ligand. As you mentioned even for the
> absolute binding free energies, the value is too high.

Just to clarify: of course you would need to decouple the ligand in
solution as well if you really wanted to get the absolute binding free
energy.  Otherwise decoupling only from the protein would just mean
transforming the ligand to the vacuum.  But obviously that's the same
procedure for WT and MUT and so would not need to be carried out if you
are really only interested in the _relative_ stability of the two

For the absolute BFE you would do the decoupling in solution (only the
ligand in the solvation box) only _once_ and subtract the result from
the protein decoupling step. That's essentially what you mentioned in
your original email.

So the transformation for obtaining the relative BFE in solution would

Lig-WT -> WT
Lig-MUT -> MUT

where Lig is decoupled from its protein environment.

For the absolute BFE for either Lig-WT or Lig-MUT also

Lig -> "nothing" (in solution)

The alternative procedure would be

Lig-WT -> Lig-MUT

where you do a relative transformation between the affected side-chains.
This is the same as the first cycle but run in vertical direction.  May
be difficult if that is not a point mutation.

> I don't know whether you could help answering for that with the .mdp
> file options used in the production MD step (I attached). I have a
> Zn2+ containing enzyme with charged ligand. The co-crystal structures
> for both WT and MUT with ligand are available.

Attachments don't work with this mailing list.

> As the protocol, I first energy minimized (in two step 1- steep 2-
> cg, 5000 total steps) , NVT (200 ps) and NPT (1 ns) equilibrated the
> systems (For each lamda values of course). Then the equilbrated
> structures were run 10 ns in final MD... For the decoupling
> parameters I used the same lambda values in the .mdp for each
> simulation step.

Ok, but how do you handle the vanishing end states?  See e.g.

> ________________________________
> From: gromacs.org_gmx-users-bounces at maillist.sys.kth.se
> <gromacs.org_gmx-users-bounces at maillist.sys.kth.se> on behalf of
> Hannes Loeffler <Hannes.Loeffler at stfc.ac.uk> Sent: Tuesday, September
> 20, 2016 3:45:51 PM To: gromacs.org_gmx-users at maillist.sys.kth.se Cc:
> gmx-users at gromacs.org Subject: Re: [gmx-users] Free Energy of Binding
> Question
> It usually helps to draw a thermodynamic cycle to inform yourself what
> exactly you are doing.
> What you seem to have been doing is to decouple the ligand, once from
> the WT and once from the MUT.  This should give you the "absolute"
> binding free energy.  Computing the difference between those values
> will give you the relative binding free energy.  It is unclear to me
> what exactly you mean with dG=1134 kJ/mol but if this is the absolute
> binding free energy it seems to be excessively high.  But without
> knowing the details regarding the setup and simulation protocol it
> will not be possible to find out what could have gone wrong.
> Of course, you could also compute the relative free energy by
> transforming WT to MUT, once in presence of the ligand and once
> without the ligand (side-chain mutation). Principally, both
> approaches should give you comparable results.
> On Tue, 20 Sep 2016 11:48:21 +0000
> Abdülkadir KOÇAK <kocak at gtu.edu.tr> wrote:
> > Dear GMX Community,
> >
> > I am aiming to compare the relative binding energy (BE) of a ligand
> > to wild type (WT) vs mutant (MUT) protein and thus trying to run a
> > Free Energy Calculation for the binding energy of the ligand to both
> > proteins (WT and MUT) using Bennett Acceptance Ratio (BAR).
> >
> > As the first step, I calculated decoupling of the ligand from both
> > proteins in two seperate MD runs by first turning off the coulombic
> > interaction and then the van der waals interaction. The next step
> > would be the solvation free energies of the ligand in order to get
> > the correct BE. However, the ligand is the same for both WT and MUT,
> > so in terms of getting relative BE, should we still calculate the
> > ligand in water or will the ligand solv energies will cancel in the
> >
> > Besides, from the decoupling of the Protein-ligand complex, I am
> > getting very high DG values (1134 kJ/mol). Is this meaningful or
> > not?
> >
> > Thx  
> --
> Gromacs Users mailing list
> * Please search the archive at
> http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before
> posting!
> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
> * For (un)subscribe requests visit
> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or
> send a mail to gmx-users-request at gromacs.org.

More information about the gromacs.org_gmx-users mailing list