acquired. Analysis of the data is then performed, and compounds
prioritized for further validation depending on the confidence
levels of the singleton LO-NMR data.
3.8 Orthogonal
Biophysical Validation
After the primary screen and the singleton validation steps, a set of
putative fragment ligands is identified. However, these compounds
may still contain a number of false positives, and further validation
steps are strongly advised before embarking on a medicinal chemistry campaign.
Typically, an orthogonal biophysical technique is used as a
validation step at this point. A number of biophysical techniques
can be used at this point, the most common being SPR, X-ray
crystallography, or protein-observed NMR [17]. A crystal structure
of the ligand bound to the protein target is often regarded as a
prerequisite for further progression of a fragment, although a range
of methods exist that allow fragment evolution in the absence of a
crystal structure [43]. However, discussion here will be limited to
validation via protein-observed NMR (PO-NMR) since other techniques lie firmly outside of the remit of this chapter.
In order to validate the putative ligand via PO-NMR, the
compound is titrated onto the isotope-labeled protein; perturbations of the NMR spectrum that occur in a dose–response manner
are taken to be indicative of binding (see Note 6). If possible,
separate samples should be used with a constant concentration of
DMSO in order to reduce the potential for false positives resulting
from the concomitant titration of DMSO alongside the compound.
Alternatively, a separate DMSO titration can be performed against
protein to serve as a control for a series of putative ligands. Additionally, if any of the compounds are charged, then a control
experiment where a simple acid or base is titrated onto the protein
is strongly advised in order to identify possible pH-related artifacts.
Further validation of a putative ligand can be obtained via
examination of the pattern of observed chemical shift perturbations
(CSPs), particularly where
15 N–
1
H or
13
C–
1 H correlation spectra
are acquired. A CSP pattern that is localized to a region of the
protein (where sequence specific assignments are available), or
which is similar to that observed for a known ligand or substrate
(where sequence specific assignments are not available), is a strong
indicator that the putative ligand interacts with a defined binding
site on the protein.
15
N–
1 H correlation spectra typically show more
widespread CSP patterns than do
13
C–
1 H spectra, owing to the
propagation of chemical shift perturbations along and across secondary structure elements, but both types of spectra can be reliably
used to distinguish true ligands from false positives.
Fragment Screening by NMR
265
prioritized for further validation depending on the confidence
levels of the singleton LO-NMR data.
3.8 Orthogonal
Biophysical Validation
After the primary screen and the singleton validation steps, a set of
putative fragment ligands is identified. However, these compounds
may still contain a number of false positives, and further validation
steps are strongly advised before embarking on a medicinal chemistry campaign.
Typically, an orthogonal biophysical technique is used as a
validation step at this point. A number of biophysical techniques
can be used at this point, the most common being SPR, X-ray
crystallography, or protein-observed NMR [17]. A crystal structure
of the ligand bound to the protein target is often regarded as a
prerequisite for further progression of a fragment, although a range
of methods exist that allow fragment evolution in the absence of a
crystal structure [43]. However, discussion here will be limited to
validation via protein-observed NMR (PO-NMR) since other techniques lie firmly outside of the remit of this chapter.
In order to validate the putative ligand via PO-NMR, the
compound is titrated onto the isotope-labeled protein; perturbations of the NMR spectrum that occur in a dose–response manner
are taken to be indicative of binding (see Note 6). If possible,
separate samples should be used with a constant concentration of
DMSO in order to reduce the potential for false positives resulting
from the concomitant titration of DMSO alongside the compound.
Alternatively, a separate DMSO titration can be performed against
protein to serve as a control for a series of putative ligands. Additionally, if any of the compounds are charged, then a control
experiment where a simple acid or base is titrated onto the protein
is strongly advised in order to identify possible pH-related artifacts.
Further validation of a putative ligand can be obtained via
examination of the pattern of observed chemical shift perturbations
(CSPs), particularly where
15 N–
1
H or
13
C–
1 H correlation spectra
are acquired. A CSP pattern that is localized to a region of the
protein (where sequence specific assignments are available), or
which is similar to that observed for a known ligand or substrate
(where sequence specific assignments are not available), is a strong
indicator that the putative ligand interacts with a defined binding
site on the protein.
15
N–
1 H correlation spectra typically show more
widespread CSP patterns than do
13
C–
1 H spectra, owing to the
propagation of chemical shift perturbations along and across secondary structure elements, but both types of spectra can be reliably
used to distinguish true ligands from false positives.
Fragment Screening by NMR
265