Chemistry Reference
In-Depth Information
compounds and from the target protein may render accurate lineshape analysis difficult.
In order to detect weakly binding ligands reliably, the molar excess of ligand over target
protein must be kept fairly low.
Rather than measuring the ligand linewidths, a more common setup is to apply a relaxa-
tion filter, i.e. a spinlock (e.g. a CPMG sequence)
[
106
]
directly after the 90 degree detection
pulse and before the acquisition of the signal.
[
75
]
The relaxation filter experiments are
performed on two samples that are identical except that the target protein is present in
one and absent in the reference sample. The spectra collected on these two samples are
compared for changes in NMR signal intensities of the small molecules; see Figure 4.4
for an example. A long enough relaxation filter will eliminate the target protein signals.
With the assumptions that the exchange contributions to the linewidths are negligible
and that the ligands bind with a similar on-rate (e.g. a diffusion-controlled on-rate) and
to one site only, the signals for a higher affinity ligand will decrease more than for a
lower affinity ligand with a given time length of the spinlock. Therefore, in general, it
is possible to use the outcome of the relaxation-filtered experiment to rank ligands with
respect to affinity.
[
24, 107
]
A convenient way to affinity rank ligands is to collect relaxation-
filtered spectra with different durations of the spinlock. The duration of the spinlock for
a given target determines the detection cut-off and, therefore, when longer spinlock times
are used, weaker binding ligands are detected. For example, simulations have sugges-
ted that for a relatively small target protein of MW
15 kDa, a spinlock of 400 ms
will eliminate the signals from ligands binding with an affinity of 500 M or tighter for
equimolar amounts of protein and ligand.
[
24
]
Here, it can be noted that it is often erro-
neously claimed that fast exchange is always a requirement for direct detection of binding
by ligand-detected methods. That is not the case for the transverse relaxation filter tech-
nique provided that no molar excess of ligand over target protein is used. In the extreme
case of a covalent binder, for example, all ligand molecules would then be bound to the
target protein during the spinlock time and no NMR signal from the ligand would pass the
relaxation filter.
It is also possible to utilize the differences in spin-lattice relaxation (
T
1
) between a
large protein and small organic molecule to detect binding by applying an inversion-
recovery pulse sequence.
[
76
]
However, detection of binding is only possible for
selective
T
1
measurements, i.e. the inversion of fragment
1
H signals must be performed by the
use of frequency-selective pulses. As a consequence, it is not practical to use this tech-
nique for primary fragment screening if the screened fragment signals are to be observed,
since a separate inversion pulse would have to be designed for every fragment in the
library. Instead, a weakly binding 'spy'molecule should be employed in competition exper-
iments where only selectively inversed signals from the 'spy' molecule are repeatedly
monitored.
[
76
]
∼
Paramagnetic spin labels.
Aparamagnetic spin label is an atom with an unpaired electron,
which will enhance the relaxation of nearby protons due to the strong electron-proton
dipole-dipole interaction. By covalently attaching a paramagnetic spin label on selected
target protein side-chains, the transverse
1
H relaxation times of any fragments binding in the
vicinity of the spin label will decrease dramatically. For detection, a transverse relaxation
filter experiment is usually applied. This technique has been dubbed SLAPSTIC (spin labels
attached to protein side-chains as a tool to identify interacting compounds).
[
81, 82
]
The strong
