AMBER Archive (2003)

Subject: MM-PBSA results

From: Thomas Steinbrecher (thomas.steinbrecher_at_physchem.uni-freiburg.de)
Date: Fri Mar 21 2003 - 07:15:25 CST

Dear AMBER users,

I have performed MM-PBSA calculations with AMBER7 on a
receptor with varying ligands. I simulated 1ns of dynamics
and extracted 100 snapshots for each complex.

The calculations were performed as in the MM-PBSA tutorial
distributed with AMBER, electrostatics were treated with
the Delphi program.

The calculations resulted in Binding Free Energies of -21
to -30 kcal/mole which are far to negative to be correct.
For a molecule that is know not to bind to the receptor -18
kcal/mole were obtained. Experimental data suggest Binding
Free Energies of about 7 kcal/mole.

My question is, can the lack of agreement between the
calculations and experimental data be attributed to a
structural change in the protein from a ligand bound to
unbound conformation that is to slow to be observed in my
simulations?
Or are there other possible causes for such an offset
between the calculated and exerimental values?

>From several papers I consulted, I expected the MM-PBSA
method to produce Binding Free Energies that are in good
accordance with experimental data, so I wonder if I did
something wrong or if my system could be inappropriate for
treatment by the MM-PBSA method.

Thank you in advance for any suggestions/comments,

Thomas Steinbrecher