Chemistry Reference
In-Depth Information
magnetization transfer in the free state. For the binding compounds, the contribution from
the bound state will dominate provided that the ligand excess is not extremely high and/or
the interaction very weak. Consequently, the resulting WaterLOGSY spectrum will consist
of positive signals from compounds binding to the target protein and negative signals from
compounds not binding to the protein. To detect also very weak binders, WaterLOGSY
spectra should be collected in both the presence and absence of target protein. The signals
from a very weak binder may be negative, but less negative compared with when the target
protein is absent. As with the STD experiment, efficient detection of reversibly binding
compounds is based on the long
T
1
values of the free ligand protons and the use of an
excess of ligand leading to a build-up of the population of ligands that has experienced
magnetization transfer during the mixing time.
Both STD and WaterLOGSY are very popular fragment screening techniques and share
many features. However, STD is an incoherent technique, i.e. it relies on transfer of satura-
tion (incoherence), whereas WaterLOGSY takes advantage of the same physical processes
but in a coherent way, i.e. polarization transfer. Both techniques use a relatively high lig-
and:protein ratio and the sensitivity with respect to both protein consumption and detectable
dissociation constant range is comparable. However, for targets with a low proton density,
e.g. RNA, the use of WaterLOGSY is clearly advantageous.
[
103
]
Targets with a low proton
density suffer from inefficient spin diffusion, leading to low STD sensitivity. WaterLOGSY,
on the other hand, is much less dependent on spin diffusion efficiency since the method
mainly relies on magnetization transfer from the water molecules surrounding the target.
Nevertheless, the low spin diffusion efficiency inDNAcan be exploited in STD experiments
to obtain information on the binding site of ligands. By saturating at different frequencies
(and therefore different DNA regions) and observing the relative STD responses, base-pair
intercalators have been distinguished from minor groove binders.
[
104
]
For the best sensi-
tivity, the WaterLOGSY experiment should be performed in H
2
O with only small amounts
of D
2
O for field-frequency locking. The intense H
2
O signal, however, raises the need for
efficient water suppression schemes and is a source of potential spectral artifacts. Since sig-
nals from nonbinding compounds appear in the spectrum as negative peaks, the spectrum
will be more complex than the corresponding STD spectrum.
Line broadening and relaxation filter experiments.
The spin-spin relaxation times (
T
2
)of
protons in a slowly tumbling entity such as a target protein are much shorter than those of
protons from a faster tumbling small organic compound. Since the
1
H signal linewidth at
half-height is proportional to (
πT
2
)
−
1
, a simple approach is to measure the linewidths of the
ligand
1
H signals in the absence and presence of a target protein. The method requires that
the magnet field homogeneity is identical for the samples. This is achieved by employing
a gradient shimming step for each sample and by checking that the shimming quality is
identical for all samples by observing the linewidth of an inert compound present in all
samples, e.g. the calibration standard DSS (2,2-dimethyl-2-silapentane-5-sulfonic acid).
Upon binding of the ligand to the target protein, the tumbling rate of the small molecule
is decreased. This results in a decreased
T
2
, which is manifested by both a broadening and
a concomitant decrease of ligand
1
H peak heights. Further, if fast exchange kinetics are
assumed and possible exchange contributions to the linewidths are small or negligible, it is
also straightforward to estimate the binding affinity.
[
105
]
However, since the primary NMR
screening is often performed on cocktails of fragments, spectral overlap both from other
