crystal structures of Dvl-PDZ bound to an organic compound and
snapshots from molecular dynamics simulations of the Dvl-PDZ/
peptide complex guided the creation of pharmacophore models.
These models combined with a virtual screening of a large chemical
library allowed the identification of compounds that could mimic
the binding mode of reference molecules. Fluorescence spectroscopy and NMR experiments confirmed the binding of several compounds at the Dvl-PDZ–CXXC5 interface, and the best compound
had a K d value of 22 μM.
Using a combination of NMR, quantitative structure–activity
relationship (QSAR) and structure-based pharmacophore filtering,
Shan et al. identified and optimized inhibitors for the Dvl-PDZ
domain [48]. This series of compounds essentially consists of
merging a benzoic acid moiety with two protein residues. The
best compound from the series exhibited a K i value of 1.5 μM in
the fluorescence polarization assay. Potential binding modes were
predicted using the Glide docking tool [30] and matched those
from endogenous partners.
Hori et al. reported new inhibitors for the Dvl-PDZ domain
using the “NMR/Docking Performance Index” (NMR-DPI) protocol, which relies on both NMR and molecular docking experiments [49]. Several reference inhibitors were investigated with
GOLD as the docking engine [50, 51] to select the best scoring
scheme using different scoring functions (ChemScore, GoldScore,
and ChemPLP) and with and without consensus scoring. The best
scoring protocol was then employed for virtual screening with a
focused library (approximately 5 K compounds). In total, 13 compounds were selected to be experimentally tested, and several of
them showed partial proliferation inhibition activity against a triplenegative breast cancer cell line.
Saupe et al. reported in 2011 a study about the Shank3-PDZ
domain [10]. The ChemBioNet library was first screened using a
fluorescence polarization assay, and one natural product-like scaffold (cyclopentyl-tetrahydroquinoline-carboxylates) emerged as a
PDZ inhibitor. SAR studies around this core were performed to
optimize its potency. The best compound analog exhibited a K i
value in the 10 μM range, and the binding of the compound within
the PDZ domain was confirmed by NMR experiments. Then, X-ray
crystallography studies were used to determine the structure of
Shank3-PDZ/inhibitor complex (PDB ID: 3O5N) and to confirm
its ability to mimic the C-terminal end of the protein partner.
Kegelman et al. disclosed one chemical inhibitor, which targets
the first PDZ domain of syntenin (Synt-PDZ1), using an integrated
strategy involving an NMR-based screening of fragment-like compounds, SAR studies, and molecular modeling [13]. More precisely, approximately 5 K fragments were initially evaluated using
an NMR-based screening. Two nonoverlapping fragments were
identified, and a structure-based linking strategy was employed to
284
Laurent Hoffer et al.
snapshots from molecular dynamics simulations of the Dvl-PDZ/
peptide complex guided the creation of pharmacophore models.
These models combined with a virtual screening of a large chemical
library allowed the identification of compounds that could mimic
the binding mode of reference molecules. Fluorescence spectroscopy and NMR experiments confirmed the binding of several compounds at the Dvl-PDZ–CXXC5 interface, and the best compound
had a K d value of 22 μM.
Using a combination of NMR, quantitative structure–activity
relationship (QSAR) and structure-based pharmacophore filtering,
Shan et al. identified and optimized inhibitors for the Dvl-PDZ
domain [48]. This series of compounds essentially consists of
merging a benzoic acid moiety with two protein residues. The
best compound from the series exhibited a K i value of 1.5 μM in
the fluorescence polarization assay. Potential binding modes were
predicted using the Glide docking tool [30] and matched those
from endogenous partners.
Hori et al. reported new inhibitors for the Dvl-PDZ domain
using the “NMR/Docking Performance Index” (NMR-DPI) protocol, which relies on both NMR and molecular docking experiments [49]. Several reference inhibitors were investigated with
GOLD as the docking engine [50, 51] to select the best scoring
scheme using different scoring functions (ChemScore, GoldScore,
and ChemPLP) and with and without consensus scoring. The best
scoring protocol was then employed for virtual screening with a
focused library (approximately 5 K compounds). In total, 13 compounds were selected to be experimentally tested, and several of
them showed partial proliferation inhibition activity against a triplenegative breast cancer cell line.
Saupe et al. reported in 2011 a study about the Shank3-PDZ
domain [10]. The ChemBioNet library was first screened using a
fluorescence polarization assay, and one natural product-like scaffold (cyclopentyl-tetrahydroquinoline-carboxylates) emerged as a
PDZ inhibitor. SAR studies around this core were performed to
optimize its potency. The best compound analog exhibited a K i
value in the 10 μM range, and the binding of the compound within
the PDZ domain was confirmed by NMR experiments. Then, X-ray
crystallography studies were used to determine the structure of
Shank3-PDZ/inhibitor complex (PDB ID: 3O5N) and to confirm
its ability to mimic the C-terminal end of the protein partner.
Kegelman et al. disclosed one chemical inhibitor, which targets
the first PDZ domain of syntenin (Synt-PDZ1), using an integrated
strategy involving an NMR-based screening of fragment-like compounds, SAR studies, and molecular modeling [13]. More precisely, approximately 5 K fragments were initially evaluated using
an NMR-based screening. Two nonoverlapping fragments were
identified, and a structure-based linking strategy was employed to
284
Laurent Hoffer et al.