503
15 (How to) Profit from Molecular Dynamics-based Ensemble Docking
application of graphics processing units (GPUs) speeding up molecular dynamics
simulations and hence allowing for routine microsecond simulations [15].
As all atomic coordinates as well as the system energy are stored, a conformational ensemble capturing thermally accessible states for subsequent protein–ligand
docking can be extracted from the trajectory. In addition to capturing enthalpic contributions, these thermodynamic ensembles also include entropic effects rendering
a valid free energy for the set of conformations. This fact is especially important,
as flexible regions both close to the binding site as well as at distinct allosteric sites
were shown to influence protein–ligand binding sites [16]. Thus, inherently flexible
regions are of considerable interest for both lead identification and lead optimization.
15.1.3 Molecular Docking
Molecular docking aims at predicting valid protein–ligand geometry (pose) from a
receptor conformation and a set of ligand conformations. To identify a promising
binding mode, the poses are scored according to molecular interactions and steric
characteristics aiming at the identification of the optimal binding pose of a ligand.
More important than the identification of a binding mode is the application of molecular docking on large compound databases. The score is used to rank potential
ligands to obtain an enrichment of potential high affinity binders amongst highest scoring hits. Molecular docking is applied as standard structure-based virtual
screening technique for the identification of early hits in drug discovery as well as
idea generator for de novo design [17]. Another application of docking protocols is
the elucidation of a binding mode for an experimentally verified ligand at atomic
level.
The described rigid docking approach does not account for receptor flexibility,
solely aiming at the identification of the most suitable three–dimensional arrangement of two rigid bodies. The inclusion of side chain flexibility to the receptor poses
additional challenges in form of additional degrees of freedom to be sampled within
the docking workflow [18]. Soft docking applications were described including side
chain flexibility inherently by allowing some degree of overlap between ligand and
receptor atoms [19].
Degrees of freedom are further increased if also the protein backbone is allowed
to show conformational heterogeneity leading to so-called ensemble docking. This
can either be achieved by a united description of several receptor conformations
by a grid averaging approach [20], the remodeling of loop regions after docking
[21], or the explicit treatment of different conformations during docking. This last
technique will be the focus of our review, as it largely resembles a mechanism of
conformational selection.
15 (How to) Profit from Molecular Dynamics-based Ensemble Docking
application of graphics processing units (GPUs) speeding up molecular dynamics
simulations and hence allowing for routine microsecond simulations [15].
As all atomic coordinates as well as the system energy are stored, a conformational ensemble capturing thermally accessible states for subsequent protein–ligand
docking can be extracted from the trajectory. In addition to capturing enthalpic contributions, these thermodynamic ensembles also include entropic effects rendering
a valid free energy for the set of conformations. This fact is especially important,
as flexible regions both close to the binding site as well as at distinct allosteric sites
were shown to influence protein–ligand binding sites [16]. Thus, inherently flexible
regions are of considerable interest for both lead identification and lead optimization.
15.1.3 Molecular Docking
Molecular docking aims at predicting valid protein–ligand geometry (pose) from a
receptor conformation and a set of ligand conformations. To identify a promising
binding mode, the poses are scored according to molecular interactions and steric
characteristics aiming at the identification of the optimal binding pose of a ligand.
More important than the identification of a binding mode is the application of molecular docking on large compound databases. The score is used to rank potential
ligands to obtain an enrichment of potential high affinity binders amongst highest scoring hits. Molecular docking is applied as standard structure-based virtual
screening technique for the identification of early hits in drug discovery as well as
idea generator for de novo design [17]. Another application of docking protocols is
the elucidation of a binding mode for an experimentally verified ligand at atomic
level.
The described rigid docking approach does not account for receptor flexibility,
solely aiming at the identification of the most suitable three–dimensional arrangement of two rigid bodies. The inclusion of side chain flexibility to the receptor poses
additional challenges in form of additional degrees of freedom to be sampled within
the docking workflow [18]. Soft docking applications were described including side
chain flexibility inherently by allowing some degree of overlap between ligand and
receptor atoms [19].
Degrees of freedom are further increased if also the protein backbone is allowed
to show conformational heterogeneity leading to so-called ensemble docking. This
can either be achieved by a united description of several receptor conformations
by a grid averaging approach [20], the remodeling of loop regions after docking
[21], or the explicit treatment of different conformations during docking. This last
technique will be the focus of our review, as it largely resembles a mechanism of
conformational selection.