simulations [22]. The snapshots extracted from the MD trajectory
represent multiple conformers of the protein structure, involving
both backbone and side chain dynamics. These extracted confor-
mations can reveal hidden spots within a crystal structure.
This chapter focuses on binding site identification approaches,
with an emphasis on structure-based methods. It also discusses the
use of MD simulations in identifying these sites, understanding
their dynamicity and evaluating their druggability. Finally a few
case studies are summarized, followed by a summary and conclu-
sion of our findings. We hope this chapter would shed light on
recent advances in this hot area and will be of great use for inter-
ested researchers.2 Binding Site Identification and Druggability Evaluation Methods
As mentioned above, the prediction of druggable binding sites
involves two stages. Firstly, one should identify all potential binding
sites within and on the surface of the target structure. This is
followed by the ranking of these sites in terms of their druggability.2.1 Binding Site
Identification Methods
The last few years witnessed the development of a few reliable
structure-based methods to identify binding sites [23]. These
structural methods can be categorized into two main classes,
geometric-based methods (e.g., PASS [24], POCKET [25], Lig-
Site [26]), and energy-based methods (e.g., GRID [27],
Q-SiteFinder [28]) (seeNote 1). Several reviews comprehensively
describe these two approaches [10, 11, 23, 29], though they will be
briefly summarized below.2.1.1 Geometric-Based
Methods
Geometric-based methods recognize a binding site based on its
geometric parameters. Two example parameters that are commonly
used in this aspect are the depth and surface area of the binding site.
In this context, Hajduk et al. defined a term, called pocket com-
pactness, as the ratio of the pocket volume to the pocket surface
area [3]. The optimal value for this parameter is usually in the range
of 0.4. Larger values correspond to more spherical shaped pockets
and smaller values represent more elongated shaped pockets. The
residual composition of a binding site, which includes polarization,
charges, and H-bonds, are also important geometrical parameters
to characterize binding sites. By using these various parameters one
can identify potential binding sites and provide further druggability
assessment (Fig.2).
For protein–protein interactions (PPIs), the geometric descrip-
tors are usually smaller compared to those of catalytic/active sites,
which are usually formed by major, large, and deep binding clefts.
An important study in this regard is the work by Bourgeas et al.
They extracted the best descriptors, geometrical parameters, andPrediction of Druggable Binding Sites 89