Biomedical Engineering Reference
In-Depth Information
techniques are frequently combined with WaterLOGSY. A nice example of
this combination came from the Novartis group who used WaterLOGSY in
association with T1r relaxation
33
to screen a small library of 500 compounds
for binding to Abl kinase where the active site was blocked with imatinib.
34
Interestingly, this approach yielded biophysically validated ligands that
targeted the myristate site. However, the compounds were non-functional
inhibitors. The reason for this is discussed below in the section on protein-
ligand structures. Another recent example demonstrates the capability of the
combination of virtual screening and WaterLOGSY to generate validated hits
for protein-protein interaction targets.
35
The calcium-binding protein S100B
binds the C-terminus of p53 where it is thought to inhibit the transcription
regulation function and may play a role in the progression of cancer. A
collection of 123 000 commercially available compounds was screened using
three different in silico approaches; importantly however, the ensemble of
NMR structures of S100B was used as input for each. 280 hits from the virtual
screen were purchased and binding to S100B was assayed using WaterLOGSY.
The most interesting compounds from WaterLOGSY were titrated into
15
N-
labelled protein samples to confirm binding and determine the binding site (yet
another example of the use of protein-observed NMR to validate screening
hits). This approach successfully yielded five selective inhibitors that bound at
the p53 binding site with reasonable ligand efficiency. The most potent
inhibitor was soaked into a crystal of S100B and indeed bound to the same site
as a peptide from p53.
In the original implementation of WaterLOGSY, water magnetisation is
destroyed prior to acquisition. Subsequently one must wait in order for
equilibrium magnetisation to be restored. Jahnke and coworkers recently
described a polarisation-optimised version (PO-WaterLOGSY)
72
in which (a
portion of) the water magnetisation is selectively returned to the z-axis,
partially negating the need for a relaxation delay. Thus the PO-WaterLOGSY
is said to speed throughput by a factor of three to five. Aroma-WaterLOGSY,
which selects that aromatic signals of ligands, effectively achieves similar
results to PO-WaterLOGSY using a slightly different approach.
36
Just as STD can provide details of the ligand-protein interaction that can
help define the ligand-binding mode, WaterLOGSY can provide information
on the ligand-water interaction. Solvent Accessibility and protein Ligand
binding studies by NMR spectroscopy (SALMON)
37
utilises the difference in
the sign of the NOE for free and protein-bound ligands to determine which
portions of the ligand are solvent-exposed. This information was sufficient to
differentiate two possible ligand-binding orientations that were compatible
with the electron density derived from X-ray diffraction experiments.
Interligand NOEs for PHARmacophore MApping (INPHARMA) is a
method closely related to trNOE (see below) with the twist that the NOE is
relayed by the protein.
30
INPHARMA transfers magnetisation between two
small molecules that compete for binding to the same site via an NOE to or
from the protein. INPHARMA allows indirect mapping of the binding pocket
