hits to a point where a degree of saturation of the binding curve is
observed (typically 3–4 times K D ). This can often require compounds to be soluble to the low mM range.
The requirements for compounds for a
19 F-NMR FBS are
rather different from those discussed above. Obviously, the compound must contain at least one
19
F atom; typically, only one C–F
or –CF3 group per molecule is preferred in order to reduce spectral
overlap. –CF2 groups can be used, although fluorine coupling can
be a problem in this instance. However, compounds are typically
screened at 25–50 μM, and as a result the solubility requirement is
significantly less demanding than is the case for LO- or PO-NMR.
2.3 Protein
For all NMR FBS, relatively large amounts of pure protein are
required. Assuming a 2000 compound fragment library and a
25-kDa protein, approximately 1 mg (
19 F-NMR), 30 mg (
1 H
LO-NMR), and more than 200 mg (PO-NMR) are required. For
PO-NMR, labeling with a stable isotope (usually
13
C or
15
N) is
typically required. This protein should be as homogeneous and
pure as possible and importantly should be stable for the duration
of the NMR screen (often several days).
2.4 Tool Compounds
If possible, a potent competitor molecule should be identified,
which binds to the site of interest. Although not absolutely
required, observation of displacement of weak binding fragments
by a potent competitor increases the confidence level of the fragment screen. This competitor is typically a compound known from
the literature, although proteins, peptides, or other tight binding
molecules can also be used effectively. The K D for this competitor
molecule should be μM or better.
If available, a low-affinity molecule with a K D of 100 μM or
greater can be used to identify conditions where weak binding can
be reliably identified. This molecule can be a fragment of a larger
known ligand, or another molecule such as a substrate or small
peptide.
2.5 Analysis
Software
Several software tools are available which assist in the analysis of
NMR binding experiments. The most widely used are FragmentBased Screening (Bruker, https://www.bruker.com), MNova Binding (Mestrelab, http://mestrelab.com), and ACD/Labs (https://
www.acdlabs.com). Each of these can assist in the organization and
analysis of the extensive data generated during an NMR FBS campaign, with visualization and guided spectral interpretation being
common features. However, it is also possible to analyze the NMR
data acquired manually using freely available NMR tools, although
this requires careful and consistent use of a database in order to
track the results from the screen.
254
Ben J. Davis
observed (typically 3–4 times K D ). This can often require compounds to be soluble to the low mM range.
The requirements for compounds for a
19 F-NMR FBS are
rather different from those discussed above. Obviously, the compound must contain at least one
19
F atom; typically, only one C–F
or –CF3 group per molecule is preferred in order to reduce spectral
overlap. –CF2 groups can be used, although fluorine coupling can
be a problem in this instance. However, compounds are typically
screened at 25–50 μM, and as a result the solubility requirement is
significantly less demanding than is the case for LO- or PO-NMR.
2.3 Protein
For all NMR FBS, relatively large amounts of pure protein are
required. Assuming a 2000 compound fragment library and a
25-kDa protein, approximately 1 mg (
19 F-NMR), 30 mg (
1 H
LO-NMR), and more than 200 mg (PO-NMR) are required. For
PO-NMR, labeling with a stable isotope (usually
13
C or
15
N) is
typically required. This protein should be as homogeneous and
pure as possible and importantly should be stable for the duration
of the NMR screen (often several days).
2.4 Tool Compounds
If possible, a potent competitor molecule should be identified,
which binds to the site of interest. Although not absolutely
required, observation of displacement of weak binding fragments
by a potent competitor increases the confidence level of the fragment screen. This competitor is typically a compound known from
the literature, although proteins, peptides, or other tight binding
molecules can also be used effectively. The K D for this competitor
molecule should be μM or better.
If available, a low-affinity molecule with a K D of 100 μM or
greater can be used to identify conditions where weak binding can
be reliably identified. This molecule can be a fragment of a larger
known ligand, or another molecule such as a substrate or small
peptide.
2.5 Analysis
Software
Several software tools are available which assist in the analysis of
NMR binding experiments. The most widely used are FragmentBased Screening (Bruker, https://www.bruker.com), MNova Binding (Mestrelab, http://mestrelab.com), and ACD/Labs (https://
www.acdlabs.com). Each of these can assist in the organization and
analysis of the extensive data generated during an NMR FBS campaign, with visualization and guided spectral interpretation being
common features. However, it is also possible to analyze the NMR
data acquired manually using freely available NMR tools, although
this requires careful and consistent use of a database in order to
track the results from the screen.
254
Ben J. Davis