Chemistry Reference
In-Depth Information
NMR technique would be the first choice is when one wants to find fragment binders
to a specific site for which there are no known binders, e.g. a newly identified allosteric
site. The introduction of a paramagnetic spin-label either at the site of interest or on a
known binder binding at an adjacent site would be an alternative method. However, the
long distances (15-20 Å) over which spin-labels exert their effect render the method less
precise. Another case when a protein-detected method could be preferred is when the tar-
get protein is relatively small and the structural information needed later in the project is
to come from the use of NMR spectroscopy and not X-ray crystallography. Even though
protein-detected NMR techniques generally are not the first choice for the primary fragment
screening, they could prove very valuable at later stages to providemore precise information
regarding binding sites. In this context, an approach where careful analysis of the protein
chemical shift perturbations leads to a more accurately characterized ligand-binding site
should be mentioned.
[
111
]
Most drug-like molecules contain aromatic rings that have an
effect on the protein chemical shifts upon binding. By quantifying the spatial dependence
of the ligand ring current field, more precise information on the binding geometry can
be obtained.
The ligand-detected NMR techniques for primary fragment screening described above
can all be used when a suitable spy molecule is available. The screening will then detect
fragments able to displace the spy molecule, most likely due to overlapping binding sites.
If there is no suitable spy molecule available, the choice of a ligand-detected NMR tech-
nique for primary fragment screening is in practice limited to four techniques: transverse
relaxation filter, STD, WaterLOGSY or SLAPSTIC. SLAPSTIC requires the covalent
attachment of a spin label on a protein side-chain at or near the binding site of interest.
This leads to several other requirements for the use of SLAPSTIC in primary fragment
screening: (i) detailed knowledge of the 3D structure of the target protein; (ii) the pres-
ence of a protein side-chain in the vicinity of the binding site of interest amenable to
spin labeling; (iii) the attachment of the spin-label should not affect the structural integ-
rity or the binding properties of the target protein. At least for targets for which there
is no detailed structural information, it is difficult to say if these requirements are ful-
filled. Therefore, to select a screening technique where the target protein can be used
without modification will be very tempting. The other three techniques are all straightfor-
ward to use for fragment screening of novel targets. The related STD and WaterLOGSY
techniques appear to be currently the most popular techniques, one major reason being
the relatively low protein consumption. Typical protein and fragment concentrations are
0.5-5 and 50-200 M, respectively. This large ligand excess is a potential drawback since
the ligands then may start to populate nonspecific, low-affinity binding sites. The trans-
verse relaxation filter experiment, on the other hand, does not use a very high ligand
molar excess. Compared with STD and WaterLOGSY, the protein concentrations used
in fragment screening are generally higher, but the fragment concentrations are also sig-
nificantly lower. Typical protein and fragment concentrations are 4-10 and 20-40 M,
respectively. With these concentrations, transverse relaxation-filtered spectra with a suf-
ficient signal-to-noise ratio can be obtained in as short time as 10-15min, which is very
competitive compared with STD andWaterLOGSY. The throughput is, of course, hampered
by the need to collect reference spectra in the absence of protein. However, if using
the same ligand concentrations, the reference spectra may not have to be collected for
every NMR screening campaign. Protein consumption for these techniques may or may
