Chemistry Reference
In-Depth Information
4.3 Lead Identification
Several recent approaches have investigated the combination of NMR and molecu-
lar docking for identifying inhibitors for specific proteins. Typically, these
approaches apply one of two methodologies: (1) a virtual screen of a large com-
pound library followed by validation of potential binders by NMR or (2) a frag-
ment-based screen using NMR followed by the use of molecular docking to
generate a protein-ligand co-structure for optimization.
Virtual screening followed by NMR validation is perhaps the most commonly
used combination of these two techniques. Several recent studies have highlighted
the use of this approach [
166
-
169
]. Branson et al
.
[
166
] used a virtual screen with
NMR to identify inhibitors of lupindiadenosine 5
0
,5
000
-P
1
,P
4
-tetraphosphate (Ap
4
A)
hydrolase. These proteins are found in eukaryotes, prokaryotes, and archaea and
have been proposed to be involved in several biological functions, ranging from
apoptosis, DNA repair, to gene expression. In bacteria, it has also been shown to be
involved in pathogenesis, which makes this a potential target for developing
antimicrobial agents. There is also a significant difference in sequence between
the bacterial and animal forms of the protein, which makes this even more attractive
as a drug target. In this study, a virtual screen using DOCK 4 [
41
] was performed
on Ap
4
A hydrolase from
Lupinusangustifolius
with a database of ~120,000
compounds. The docked poses from DOCK were reranked according to consensus
scoring using six different scoring functions, where the top 100 ranked ligands were
selected and then filtered again to remove all compounds with a logP of 3 or greater
in order to select for compounds likely to be water soluble. The result was seven
compounds, of which six were commercially available. These six compounds were
then subjected to isothermal titration calorimetry to identify any inhibition of
hydrolase activity. From that analysis, one compound (NSC51531), which contains
a 1,4-diaminoanthracene-9,10-dione core, showed significant binding affinity
(~1
M
K
D
) and was chosen for analysis by 2D
1
H-
15
N HSQC. The NMR analysis
showed CSPs consistent with the ATP binding site of the protein. In addition,
introducing NSC51531 to the human Ap
4
A hydrolase showed non-specific binding
and had no apparent toxic effects against human fibroblasts. This is likely due to
structural differences between the binding sites of the lupin and human forms of
Ap
4
A hydrolase. Potentially, a scaffold based upon NSC51531 could result in an
inhibitor with specificity towards the bacterial form of the protein leading to an
effective microbial agent (Fig.
9a
).
Veldcamp and coworkers [
169
] utilized a similar method that targeted the
chemokine CXCL12, which activates the CXCR4 receptor shown to be involved
with cancer progression. In this approach, nearly 1.5 million compounds from the
ZINC database [
170
] were screened using DOCK 3.5 [
171
] against the region of
CXCL12 that interacts with CXCR4. Specifically, a sulfotyrosine (sY21) was
targeted since it was anticipated to be an important residue for the CXCL12-
CXCR4 interaction. The top 1,000 hits were manually inspected to identify five
compounds with a favorable interaction with sY21. These five compounds were
m
