3.3 Development
of NMR Fragment
Screening Assay
Although many NMR binding experiments have been described in
the literature, this chapter will focus on the application of the
commonly used set of 1D
1 H NMR, STD, water-LOGSY, and
relaxation-filtered 1D with solvent suppression experiments as discussed above. These ligand-observed NMR binding experiments
should be implemented and tested as described elsewhere—see
Note 2 and Ref. 16.
Depending on the details of the protein and spectrometer, one
or more of these experiments may be excluded. For example, the
STD experiment has relatively low signal/noise, and is consequently often the longest of the acquisitions, and thus may not be
suitable to be run in a reasonable time on a lower field instrument
with a room temperature probe. The water-LOGSY experiment has
also been observed to give rise to anomalous spectra for some
protein:buffer combinations and may not be suitable to be acquired
in all circumstances [42]. In general, however, it is usually better to
acquire more experiments rather than fewer. For each sample,
during assay development, trial screening, and screening of the
full fragment library, the same set of experiments should be
acquired.
The simple
1 H 1D with solvent suppression should always be
acquired since this experiment allows inspection of almost all components present in the assay. As such, it is advisable to acquire a 1D
1 H spectrum with sufficient signal/noise to observe the protein
resonances since this enables identification of situations where the
protein has denatured, refolded, or precipitated during the
experiment.
3.3.1 Binding of a
Low-Affinity Tool
Compound
The first step in developing the NMR fragment screening assay is to
establish a positive control, i.e., to confirm that binding of a known
low-affinity ligand (tool compound) can be observed using the
LO-NMR experiments (see Note 5). Binding in this case is taken
as a positive signal in the STD spectrum, a shift toward a more
positive signal in the water-LOGSY spectrum and a reduction in
signal in the relaxation-filtered spectrum. If binding is not observed
for a known low-affinity ligand, the first parameter to change is to
increase the concentration of protein in the sample; if this does not
result in observation of binding, the concentration of the
low-affinity ligand should be increased. Additional parameters
(such as STD or water-LOGSY mixing times) can also be modified,
although it is also important at this point to verify that the
low-affinity ligand is binding to the target protein through the
use of orthogonal methods such as protein-observed NMR or
another biophysical technique.
If no low-affinity tool compound is available, the trial screen
(see Subheading 3.3.4) should be run using “default” parameters,
and the hit rate analyzed. Conditions that give hit rates of 2% or less
should be modified as discussed for situations where binding of the
low-affinity tool compound is not observed.
258
Ben J. Davis
of NMR Fragment
Screening Assay
Although many NMR binding experiments have been described in
the literature, this chapter will focus on the application of the
commonly used set of 1D
1 H NMR, STD, water-LOGSY, and
relaxation-filtered 1D with solvent suppression experiments as discussed above. These ligand-observed NMR binding experiments
should be implemented and tested as described elsewhere—see
Note 2 and Ref. 16.
Depending on the details of the protein and spectrometer, one
or more of these experiments may be excluded. For example, the
STD experiment has relatively low signal/noise, and is consequently often the longest of the acquisitions, and thus may not be
suitable to be run in a reasonable time on a lower field instrument
with a room temperature probe. The water-LOGSY experiment has
also been observed to give rise to anomalous spectra for some
protein:buffer combinations and may not be suitable to be acquired
in all circumstances [42]. In general, however, it is usually better to
acquire more experiments rather than fewer. For each sample,
during assay development, trial screening, and screening of the
full fragment library, the same set of experiments should be
acquired.
The simple
1 H 1D with solvent suppression should always be
acquired since this experiment allows inspection of almost all components present in the assay. As such, it is advisable to acquire a 1D
1 H spectrum with sufficient signal/noise to observe the protein
resonances since this enables identification of situations where the
protein has denatured, refolded, or precipitated during the
experiment.
3.3.1 Binding of a
Low-Affinity Tool
Compound
The first step in developing the NMR fragment screening assay is to
establish a positive control, i.e., to confirm that binding of a known
low-affinity ligand (tool compound) can be observed using the
LO-NMR experiments (see Note 5). Binding in this case is taken
as a positive signal in the STD spectrum, a shift toward a more
positive signal in the water-LOGSY spectrum and a reduction in
signal in the relaxation-filtered spectrum. If binding is not observed
for a known low-affinity ligand, the first parameter to change is to
increase the concentration of protein in the sample; if this does not
result in observation of binding, the concentration of the
low-affinity ligand should be increased. Additional parameters
(such as STD or water-LOGSY mixing times) can also be modified,
although it is also important at this point to verify that the
low-affinity ligand is binding to the target protein through the
use of orthogonal methods such as protein-observed NMR or
another biophysical technique.
If no low-affinity tool compound is available, the trial screen
(see Subheading 3.3.4) should be run using “default” parameters,
and the hit rate analyzed. Conditions that give hit rates of 2% or less
should be modified as discussed for situations where binding of the
low-affinity tool compound is not observed.
258
Ben J. Davis