[Prev][Next][Index][Thread]
Binding energy calculations in water solvent (repeat) (fwd)
Sender: bruce_bush@merck.com (Bruce Bush)
Subject: WSN: Binding energy calculations in water solvent
I would be grateful for information, anecdotes, citations, etc.
on attempts to calculate ENZYME-INHIBITOR BINDING FREE ENERGY
with *simplified* or *continuum* solvent models,
while allowing the solutes (ligand and receptor) to move dynamically.
This query is not directed especially toward full-fledged
thermodynamic perturbation methods (FEP) that
employ all-atom water models,
simulate a chemical mutation of the solute ligand,
yield a single value of delta-free-energy,
and require (or assume) that the simulation has sampled
all relevant water networks and conformations of the
ligand + receptor, such as sidechain roatmers.
I'm more interested for the present in SIMPLIFIED WATER models
that may sample wider conformational changes of solutes.
Examples of such models (in approx. order of increasing sophistication):
Simplified empirical functions of atom pairs in the contact zone
(taking "solvent exposure" into account) a la Bohacek and McMartin
Surface-area-based energies
Group-constant methods
Binding site mapping incorporating desolvation, 'hydrophobic' binding
Continuum dielectric models based on surface boundary elements or
finite differences (DelPhi, COSMO, GRASP, etc.)
Effective potential: nonadditive solvent exposure effect (GB/SA etc.)
Mean-field discrete water (Langevin Dipole: POLARIS etc.)
(This list is not meant to be complete, nor to imply endorsement!)
A major challenge is to estimate ddG(bind) for a large series of ligands
without adjusting regression coefficients. Has anyone accomplished this
over a diverse series of substituents: hydrophobes, polar groups, small large?
Information on GOOD TRIES, NEAR-MISSES, PUZZLING BEHAVIOR (not only SUCCESSES)
would be of interest to the whole community, I'm sure.
Thank you -- / Bruce /
bruce_bush@merck.com (908) 594-6758