Dear Francois,
I am afraid you have to go for a nanosecond-long dynamics in explicit
solvent if you want to obtain a change in fold of the enzyme upon mutation.
As for dG, you also have to run MD and then post-process it by MM-PBSA, for
example or run even more computationally demanding thermodynamic integration
or free energy perturbation.
Cheer up!! The results will pay off! (Ca vaut la peine!!)
Regards,
Martin Lepsik
> I would like to calculate the "dG" (free energy value) between a wild type
> protein and one of its mutant (simple residue substitution, let's say for
> instance: Ala45Pro). AMBER can estimate such values as a sum of the MM
> energy, non-polar and electrostatic solvation (entropy might or might not
be
> computed). If I well understood, the mutant is constructed from the wild
> type fold. My feeling is that using this wild type fold to construct the
> mutant might be wrong in particular if the experimental mutant fold is
> highly different from the wild type one. However, if the experimental
mutant
> fold is unknown, it is (for me) the unique solution. To get this dG value,
> comparative molecular dynamics or minimization strategies are used. It is
> clear that running two different dynamics on the wild type and mutant
> proteins will provide better information on the dG value in particular
(for
> me) when the mutant fold is different from the wild type one.
>
> Now if we apply the 'simple' minimization strategy on both proteins, is it
> wrong to say that the dG value obtained show 'how much' the mutation is
> unfavorable in the wild type fold (when the mutant fold is different from
> the wild type one) or do we have to run molecular dynamics in this case
(the
> simulation time is not the same !) ?
>
> Thanks, Sincerely,
> Francois
Received on Wed Dec 19 2001 - 06:43:25 PST
