Ddg calculation for a metalloprotein using APBS

[Anonymous]

Hi all,

I am trying to calculate Ddg for a protein complex, which is made of 3 chains: A, B and C.

The chains B and C have a zinc-finger each, which are a zinc ion complexed by four CYS side-chains.

To calculate the Ddg I am using the SetupPoissonBoltzmannPotential mover followed by the Ddg filter, in order to separate chain B from chain A+C and calculate the binding energy.

The problem I’m having is that to calculate Ddg, the SetupPoissonBoltzmannPotential mover creates a .pqr file for the bound form and another .pqr file for the unbound form. As you can see in the attached figure, in the unbound.pqr the chain B is separated from the others, but the zinc ion that belongs to chain B is not shifted along with chain B.

I have tried to use SetupMetalsMover to create a distance constraints in the metal binding sites, but without success. 

Is there any mover specific for metalloprotein to calculate binding energy?

 

Thanks,

Danilo.

[Anonymous]

There are two possible cases, here:

1.  The metal is bound solely by residues in chain B, or solely by residues in chains A and C.

2.  The metal is at the B/AC interface, and is bound by residues in both B and in (A or C).

If case 1, metals in Rosetta are usually a separate chain — let’s call it D.  You can either:

– Try to set up the mover that you’re using to calculate the Ddg between BD and AC (if B binds the metal) or between B and ACD (if AC binds the metal), or

– Delete the metal from your structure.  If the metal is not at or near the interface, this will presumably have the same effect on both the bound and unbound states, and no effect on the delta-G of binding.

If case 2, Rosetta is the wrong tool for the job, since coordinate covalent bonds must be broken to separate the complex, and you need a quantum mechanics software package to accurately model the energetics of that.  I recommend GAMESS.

[Anonymous]

Hi Dr. Vikram,

Thanks for the quick response. I’ll try your recomendations and will give a feedback on the results.

Best,

Danilo.

 

[Author]

Posts