Hi all,
If pKa value or binding constant of a metal ion is given so How can I calculate 12-6 Lennard-Jones (LJ) Nonbonded parameter for metal ion. I need to try this parameter for my test experiment of molecular Dynamics in Gromacs.
Any clues or tutorial or article would be much appreciable.
Thank you.
I think the question is aimed directly at force field parameter optimization, for which there are few degrees of freedom in an additive model - charge is fixed, so the only thing that can be tuned is the Lennard-Jones parameters. This is a considerable limitation and is very crude, especially for anything with a +2 or larger charge. Monovalent ions behave reasonably well, but anything divalent or more complex is pretty dicey since polarization and charge-transfer effects (which are prominent) are completely neglected in this representation.
Alternatives would be polarizable force fields or multi-site models. But the parametrization strategy, as always, depends on the parent force field, in this case, the one chosen to represent the protein. Whatever its methods are for deriving ion parameters (surely there will be some that are already compatible for some species to guide the process) are what need to be followed.
But the larger point is that I doubt any additive representation with an integer charge and simple LJ interactions will ever yield correct behavior with respect to pKa shifts (again, simply not possible with a fixed-charge force field unless you're doing constant-pH simulations, which is another boat entirely), polarizable/fluctuating charge models, or QM/MM (in which case you don't need force field parameters for the ion because it will be part of the QM region).
Thank you Justin for fruitful explanation. But I am looking for methods/equations by which I could be able to approximate parameter. Basically I am looking for Zn--Hist residues LJ optimized parameter using gromos FF.
The GROMOS force fields already have Zn2+ parameters, so that interaction should be covered, though I can't vouch for how accurate it is. Again, any additive approximation of such interactions will be extremely crude (QM/MM has shown, for instance, that in His-ligated Zn complexes, the Zn ion only carries about a +0.7 e charge).
If you feel that some refinement is necessary for specific interactions/atom pairs, then you need to define suitable target data (e.g. QM interaction energy for Zn-imidazole, or geometric parameters of a QM-optimized complex) and parametrize that specific interaction. I am not aware of a published source for GROMOS ion parametrization, and the relationship to QM is minimal, at best, since the force field is highly empirical relative to other force fields like AMBER and CHARMM, which have a well-established, direct relationship to QM (which actually makes parametrization much easier).
- Badges
- Science topic