3.1 Characterization
of the Fragment
Library
Before using a fragment library for screening, it is important to
characterize the fragments thoroughly—both in order to verify the
compound identity and also to reduce the risk of artifacts associated
with poor compound behavior (such as instability or insolubility) in
the assay. This characterization can be performed in a number of
ways, but the focus here will be on NMR characterization (since the
library is to be screened by NMR). Additionally, the library should
be quality controlled periodically (typically every 1–2 years) in
order to identify degradation or precipitation of the stock
solutions.
3.1.1 Sample
Preparation
Stock solutions of all compounds intended for inclusion in the
library should be made up at 200 mM in d6-DMSO. This high
concentration enables cocktails of compounds to be prepared while
keeping the total concentration of d6-DMSO at levels that are
tolerated by most proteins (see Subheading 3.2). Samples of each
compound should then be made in buffered aqueous solution;
typically, a concentration of 500 μM compound is used in QC
buffer (such as 20 mM sodium phosphate (pH 7.5), 2% final
concentration of d6-DMSO, 10% D 2 O, 50 μM DSS). A large
number of samples will typically be required in order to characterize
the entire fragment library (see Note 1).
3.1.2 Data Acquisition
1D
1 H NMR spectra should be acquired on each aqueous sample,
using a robust solvent suppression method (such as excitation
sculpting [39]) in order to remove the solvent signal (see Note
2). If the compound contains
19
F nuclei, a
19
F 1D NMR spectrum
should be acquired at this point. Additionally, it is useful to acquire
a water-LOGSY spectrum [26] on this sample in order to identify
compounds that readily self-associate [34].
Since fragment screening by NMR typically takes 1–2 days of
acquisition time, it is important to confirm that the fragments are
stable under “typical” aqueous conditions for this length of time.
This is simply done by repeating the data acquisition after 24 h
using the same samples. Examination of the two sets of spectra
rapidly reveals samples that are not stable under these conditions.
3.1.3 Data Analysis
For each compound, the 1D
1 H NMR spectrum should be analyzed in order to confirm that the spectrum is consistent with the
chemical structure. If the spectrum is not clearly consistent, it may
be necessary to acquire a 1D
1 H spectrum of the compound stock
in DMSO and/or to acquire LCMS data in order to check the
identity of the compound (even commercially supplied compounds
are not always what they are supposed to be). Additionally, the 1D
1 H NMR spectrum acquired after 24 h should be analyzed in order
to identify compounds that are not stable for extended periods in
aqueous solution. Compounds that fail either of these tests should
not be included in the fragment library.
256
Ben J. Davis
of the Fragment
Library
Before using a fragment library for screening, it is important to
characterize the fragments thoroughly—both in order to verify the
compound identity and also to reduce the risk of artifacts associated
with poor compound behavior (such as instability or insolubility) in
the assay. This characterization can be performed in a number of
ways, but the focus here will be on NMR characterization (since the
library is to be screened by NMR). Additionally, the library should
be quality controlled periodically (typically every 1–2 years) in
order to identify degradation or precipitation of the stock
solutions.
3.1.1 Sample
Preparation
Stock solutions of all compounds intended for inclusion in the
library should be made up at 200 mM in d6-DMSO. This high
concentration enables cocktails of compounds to be prepared while
keeping the total concentration of d6-DMSO at levels that are
tolerated by most proteins (see Subheading 3.2). Samples of each
compound should then be made in buffered aqueous solution;
typically, a concentration of 500 μM compound is used in QC
buffer (such as 20 mM sodium phosphate (pH 7.5), 2% final
concentration of d6-DMSO, 10% D 2 O, 50 μM DSS). A large
number of samples will typically be required in order to characterize
the entire fragment library (see Note 1).
3.1.2 Data Acquisition
1D
1 H NMR spectra should be acquired on each aqueous sample,
using a robust solvent suppression method (such as excitation
sculpting [39]) in order to remove the solvent signal (see Note
2). If the compound contains
19
F nuclei, a
19
F 1D NMR spectrum
should be acquired at this point. Additionally, it is useful to acquire
a water-LOGSY spectrum [26] on this sample in order to identify
compounds that readily self-associate [34].
Since fragment screening by NMR typically takes 1–2 days of
acquisition time, it is important to confirm that the fragments are
stable under “typical” aqueous conditions for this length of time.
This is simply done by repeating the data acquisition after 24 h
using the same samples. Examination of the two sets of spectra
rapidly reveals samples that are not stable under these conditions.
3.1.3 Data Analysis
For each compound, the 1D
1 H NMR spectrum should be analyzed in order to confirm that the spectrum is consistent with the
chemical structure. If the spectrum is not clearly consistent, it may
be necessary to acquire a 1D
1 H spectrum of the compound stock
in DMSO and/or to acquire LCMS data in order to check the
identity of the compound (even commercially supplied compounds
are not always what they are supposed to be). Additionally, the 1D
1 H NMR spectrum acquired after 24 h should be analyzed in order
to identify compounds that are not stable for extended periods in
aqueous solution. Compounds that fail either of these tests should
not be included in the fragment library.
256
Ben J. Davis