Chemistry Reference
In-Depth Information
fragments (i.e. reagents), virtually enumerating these fragments with their target scaffold
and, if a protein active site is known, scoring these virtual compounds in the protein. In
such an application, the conformational space and binding modes of the fragments are
substantially reduced to those most likely to positively impact protein-ligand interactions.
Below, two in-house examples of such virtual screening of fragments have identified active
molecules of interest to Merck's BACE-1 (Beta-site Amyloid precursor protein-Cleaving
Enzyme) inhibitor program.
Even though there has been a paucity of papers specifically focused on virtual screening
of small fragments, there is an increase in papers and presentations by scientists focusing
on identification, scoring and docking of fragments.
[
20, 21
]
Virtual screening is the process of
utilizing a hypothesis generated from existing biological or biophysical data on a given, or
homologous, target and through the application of a computational method generating a list
of compounds for screening in a validated biological assay. The choice of computational
method could range from a simple 2D methodology such as a similarity-to-substructure
method to amore sophisticated algorithmsuch as 3Ddocking of compounds and subsequent
scoring of them in the active site. Hence virtual screening may include docking and scor-
ing, but there are many other approaches, including ligand-based similarity searches and
quantitative structure-activity relationship (QSAR) models.
[
22
]
Researchers are also com-
bining virtual screening techniques with experimental methods, such as crystallography,
to aid in hit identification.
[
23
]
In addition, independent software vendors (ISVs), such as
Schrödinger, are specifically refining their scoring functions to describe fragment-protein
interactions better.
[
6
]
9.4 BACE-1: Identification of Binding Fragments
The identification of a small, brain penetrant BACE-1 inhibitor has been sought after for
nearly a decade. There are patents and publications of brain penetrant BACE-1 inhibitors,
and for a recent review of such compounds, the interested reader is referred to the report
by Hills and Vacca.
[
24
]
A couple of recent papers on BACE-1 fragments highlight complementary approaches
used by Murray
et al
.
[
23
]
The first utilizes a crystallographic approach to identify the place-
ment of the fragment. Even though computational chemistry was not used in the role of
virtual screening to identify these small molecular inhibitors, it was applied in the generation
of the fragment library which was biologically screened.
This particular fragment library was constructed to contain scaffolds and side-chains
which often appear in drugs.
[
25
]
After screening the library, direct binding to the BACE-1
active site was determined by X-ray crystallography, yielding a hit rate around 0.6%. Two
aminopyridine fragments hits are depicted in their binding mode in 2D format in Figure 9.8.
It is important to point out that when examining small fragments and trying to prioritize
them, scientists will often utilize simple metrics such as binding efficiency index (BEI)
[
3
]
and ligand efficiency (LE).
[
4
]
The BEI is calculated using the equation
BEI
=
IC
50
/
MW
(9.1)
