Computational Drug Discovery and Design

# Terminal residues have different atom names

donor mask @OXT

acceptor mask :1@N :1@H1

acceptor mask :1@N :1@H2

acceptor mask :1@N :1@H3

#

hbond print .05 series hbt time 10 distance 3.5 angle 120.0 \

out hbond.dat solventdonor WAT O solventacceptor WAT O H1 \

solventacceptor WAT O H2

The generic version of the example above is downloadable from

the addresshttp://ambermd.org/tutorials/basic/tutorial3/files/

analyse_hbond.ptraj. Adjust it according to your needs. The output

file (“hbond.dat”) contains information about the formation and

breaking of hydrogen bonds throughout the trajectory. The total

combined information should be used to assess the stability of the

studied complex.

3.7 Estimation
of Ligand Affinity

One of the most attractive features offered by the analysis of an MD

trajectory is the estimation of the protein–ligand interaction energy.

This in turn should be linked to ligand affinity. Unfortunately, this

is a very complex issue and not yet sufficiently resolved in practice.

Theoretically, the intermolecular interaction energy can be calcu-

lated from the difference between the energy of the complex and

the sum of energies of its individual components. The Gibbs free

energy of the system is the sum of the enthalpic and the entropic

terms. The force field-based energy is an approximation to the

enthalpic term of the free energy expression [55–57]. The entropic

term is often as important as the enthalpic one, but there are

enormous difficulties in computing it, and so in practice it is usually

neglected. This simplification nevertheless allows comparison of

similar ligands, for which the entropic terms may not vary too

much. In spite of these difficulties, several methods are frequently

used. The most popular approach in evaluating intermolecular

interaction energy is the molecular mechanics (MM) combined

with the Poisson-Boltzmann (PB) or generalized Born (GB) and

surface area (SA) continuum solvation methods (MM-PBSA and

MM-GBSA) [58]. Other, more sophisticated methods have also

been developed, such as interactive Linear Interaction Energy

(iLIE) approach [59–61], but they require calibration on a known

set of ligands with experimentally determined affinities, which con-

stitutes a major limitation. Moreover, they require more time to

run, which is prohibitive in virtual screening, where affinities for

many ligands have to be determined in a relatively short time. In

what follows we focus on the most widely used MM-PBSA and

MM-GBSA methods, as available in the Amber software suite.

These methods are based on the implicit solvent approach and

need as input the separate trajectories of unsolvated complex, of

Molecular Dynamics in Virtual Screening 169