intermolecular interactions can be robustly identified with few
artifacts arising from the detection method. This is particularly
important when screening for low-affinity interactions since these
can be readily masked or distorted by modification of the target’s
environment. Additionally, NMR observes almost all species present in solution, allowing verification of identity and behavior of the
compounds, protein and buffer and further reducing the risk of
errors which can give rise to erroneous results.
NMR detects intermolecular interactions by observing one or
more spectral parameters (such as chemical shift, relaxation rates, or
the transfer of magnetization). These parameters are modulated by
interactions that occur when a bimolecular complex is formed,
giving rise to spectra whose features differentiate binding from
nonbinding species and thus can be used to identify ligands.
Numerous experiments have been proposed to identify intermolecular interactions by detecting various effects, although a relatively
small set has found widespread use [16, 21]. As always, care must be
taken so as not to be misled by experimental artifacts, but in
general, these widely used experiments are sensitive, reliable, and
robust.
In the context of FBS, most NMR experiments are categorized
by the primary species observed—typically either the small molecular components (“ligand-observed NMR binding experiments”)
or macromolecular species (usually “protein-observed NMR binding experiments,” although nucleic acids can also be readily studied
by using different experiments). As mentioned previously, various
NMR experiments are used to identify intermolecular interactions;
the most common of these are summarized in Fig. 2 and described
briefly below. The different NMR experiments are best suited to
different types of target, and it is necessary to consider carefully
which approach will be taken before embarking on an NMR-based
fragment screen.
Protein-observed NMR (PO-NMR) was the first approach
used to experimentally identify fragments binding to proteins
[22] and remains widely used. Effects other than ligand-binding
that may perturb the protein spectrum (such as changes in pH or
the binding of DMSO [15]) must be excluded, but if these controls
are performed, then PO-NMR remains one of the most reliable
NMR methods used for FBS. Various spectra of the protein can be
acquired in order to detect ligand-induced chemical shift perturbations (changes in frequency of spectral peaks), ranging from simple
1 H 1D spectra (under highly favorable circumstances) through the
commonly used
15 N–
1 H HSQC spectra (suitable for smaller proteins below 20–25 kDa) to the more rarely employed TROSY-type
spectra suitable for application to larger systems of up to 50–80 kDa
(or, under some circumstances, even higher [23, 24]). In addition
to identifying intermolecular interactions, PO-NMR can be used to
identify both the K D (via titration of the ligand onto the protein)
Fragment Screening by NMR
249
artifacts arising from the detection method. This is particularly
important when screening for low-affinity interactions since these
can be readily masked or distorted by modification of the target’s
environment. Additionally, NMR observes almost all species present in solution, allowing verification of identity and behavior of the
compounds, protein and buffer and further reducing the risk of
errors which can give rise to erroneous results.
NMR detects intermolecular interactions by observing one or
more spectral parameters (such as chemical shift, relaxation rates, or
the transfer of magnetization). These parameters are modulated by
interactions that occur when a bimolecular complex is formed,
giving rise to spectra whose features differentiate binding from
nonbinding species and thus can be used to identify ligands.
Numerous experiments have been proposed to identify intermolecular interactions by detecting various effects, although a relatively
small set has found widespread use [16, 21]. As always, care must be
taken so as not to be misled by experimental artifacts, but in
general, these widely used experiments are sensitive, reliable, and
robust.
In the context of FBS, most NMR experiments are categorized
by the primary species observed—typically either the small molecular components (“ligand-observed NMR binding experiments”)
or macromolecular species (usually “protein-observed NMR binding experiments,” although nucleic acids can also be readily studied
by using different experiments). As mentioned previously, various
NMR experiments are used to identify intermolecular interactions;
the most common of these are summarized in Fig. 2 and described
briefly below. The different NMR experiments are best suited to
different types of target, and it is necessary to consider carefully
which approach will be taken before embarking on an NMR-based
fragment screen.
Protein-observed NMR (PO-NMR) was the first approach
used to experimentally identify fragments binding to proteins
[22] and remains widely used. Effects other than ligand-binding
that may perturb the protein spectrum (such as changes in pH or
the binding of DMSO [15]) must be excluded, but if these controls
are performed, then PO-NMR remains one of the most reliable
NMR methods used for FBS. Various spectra of the protein can be
acquired in order to detect ligand-induced chemical shift perturbations (changes in frequency of spectral peaks), ranging from simple
1 H 1D spectra (under highly favorable circumstances) through the
commonly used
15 N–
1 H HSQC spectra (suitable for smaller proteins below 20–25 kDa) to the more rarely employed TROSY-type
spectra suitable for application to larger systems of up to 50–80 kDa
(or, under some circumstances, even higher [23, 24]). In addition
to identifying intermolecular interactions, PO-NMR can be used to
identify both the K D (via titration of the ligand onto the protein)
Fragment Screening by NMR
249