[gmx-users] Binding energy of membrane protein to the membrane lipids

[Justin Lemkul]

On 2/23/14, 10:36 AM, sojovictor wrote:
> Dear all,
>
> I would like to find the difference in the energy of binding or insertion of
> a protein into two different types of phospholipid membranes (plus water and
> ions). i.e., my hypothesis is that the energy should be lower with one type
> of membrane than with the other.
>
> I have considered an alchemical transformation from one system to the other,
> gradually replacing a lipid with the other one, but I've been unable to find
> any example of people switching off something as large as a protein; all the
> examples I find are of ligands, single aminoacid residues, or solutes, much
> smaller particles than what I'm attempting. This makes me wonder whether
> what I'm planning even makes sense, or if such a simulation would never
> converge or give reliable results due to the large vacuum in the middle of
> the membrane created upon switching off the protein.
>
> Taking guidance from Mobley,Chodera & Dill's 2006 paper on J.Chem.Phys., and
> from Justin Lemkul's tutorial, I've devised the following, potentially
> impossible, thermodynamic cycle:
> <http://gromacs.5086.x6.nabble.com/file/n5014735/Thermodynamic_Cycle.png>
>
> Clockwise from the top left, I would do this in seven steps:
> 1) (Position?)Restrain the protein in the first membrane.
> 2) Decouple coulombic interactions in the protein (I assume annihilating
> would not be a good idea here).
> 3) Decouple Lennard-Jones interactions in the protein. This will leave me
> with a system that is effectively equivalent to an independent
> fully-restrained/non-interacting protein, and a free membrane.
> 4) Change membrane to the second type, keeping both interactions off and
> protein restriction on.
> 5) Switch on Lennard-Jones interactions.
> 6) Switch on coulombic interactions.
> 7) Remove protein restriction.
>
> I believe it wouldn't be necessary to calculate energies for the protein in
> solution, since my interest is not the free energy of binding to one
> membrane, but the change that would be experienced upon changing from one
> membrane to the other.
>

This sort of contradicts what you said in your first sentence.  I take it you're 
interested more in a transfer free energy between the two lipid types?

> Now, my initial impression is that this would not work, but I can't think of
> anything better, so I'd very much welcome input.
>
> Alternatively, I have considered umbrella sampling: I would pull the protein
> out of one of the membranes, separately do the same with the other membrane,
> and see what's the difference, but I can't think of anything I could keep as
> the reference system for the centre of mass (the membrane does not seem to
> be a good idea).
>

Why not?  This seems to be the clearest approach to me, which also has the 
upside that it avoids the possible unphysical outcomes of trying to alchemically 
transform a protein.  The complex system of restraints that would be necessary 
would, in my mind, seriously complicate the interpretation of the outcome.  A 
one-dimensional reaction coordinate along the normal to the bilayer seems 
intuitive and straightforward to calculate.

-Justin

> In general, any advice on how to proceed with this calculation, a reference
> to the literature, or wild guesses, would be very welcome!
>
> Thanks, all.
>
>
> Victor Sojo
>
> --
> View this message in context: http://gromacs.5086.x6.nabble.com/Binding-energy-of-membrane-protein-to-the-membrane-lipids-tp5014735.html
> Sent from the GROMACS Users Forum mailing list archive at Nabble.com.
>

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

Justin A. Lemkul, Ph.D.
Ruth L. Kirschstein NRSA Postdoctoral Fellow

Department of Pharmaceutical Sciences
School of Pharmacy
Health Sciences Facility II, Room 601
University of Maryland, Baltimore
20 Penn St.
Baltimore, MD 21201

jalemkul at outerbanks.umaryland.edu | (410) 706-7441
http://mackerell.umaryland.edu/~jalemkul

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