Chemistry Reference
In-Depth Information
primary screen does not have to be very large in numbers: 500-1000 chemically diverse
fragments are in general sufficient. Such relatively small libraries will not need the infra-
structure necessary for large HTS compound collections that are two to three orders of
magnitude larger in number. Fragment libraries will also be easier to keep at high quality
standards with respect to purity, not only due to the smaller number of compounds but
also since fragments contain fewer features that could cause stability problems. Rather
than having a large fragment library, it is more important to have access to a large num-
ber of analogues to the fragments in the primary fragment screening library. Analogues are
preferably already present in-house or it should be possible to synthesize them quickly from
fragments with chemical handles for straightforward expansion using one-step chemistries.
Analogues could also be ordered from commercial suppliers, but that will introduce a lag
time waiting for the compounds to arrive. As the number of performed fragment screening
campaigns increases for a given fragment library, the number of fragment analogues that
are immediately available will accumulate.
The literature on fragment library design is limited following the publications on the
SHAPES strategy from Vertex.
[
61 67
]
Fragment library design is essential since the physical
properties of fragments remain largely the same when incorporated into larger drug-like
molecules.
[
64
]
The size of the fragments in the library should be relatively small in order to
have a high probability of binding to the target, but at the same time complex enough so
that the probability that the fragment binds to the target in one way only is high.
[
7, 68
]
These
requirements would be fulfilled by fragments containing approximately 10-20 heavy atoms.
Further, it is very important that the fragments are highly soluble (preferably
>
0.5 mM
in buffer solution) since one main purpose of the primary screening is to identify binding
fragments that have a high likelihood of producing structures of the fragment-target com-
plex. Another point to consider is that compounds with a low hydrogen bonding capability
are less likely to be active.
[
66
]
In order to increase throughput and decrease protein con-
sumption, the fragments are mixed into cocktails. For ligand-detected NMR experiments,
the cocktails usually contain 6-10 fragments and with a maximum of chemical diversity
within each cocktail. In order to avoid any chemical reactions taking place within the cock-
tail, separation of acids from bases and electrophiles from nucleophiles is important. For
protein-detected NMR experiments, the number of compounds in a cocktail can be much
higher and the diversity within a cocktail should be small, in order to minimize the number
of cocktails necessary to deconvolute.
Typical restrictions applied to fragments at iNovacia include that they should contain
9-20 heavy atoms and have Clog
P <
2.All fragments are available as powder andmost stock
solutions of individual fragments are 0.5M in deuterated DMSO. This permits high frag-
ment concentrations in co-crystallization and soaking attempts without excessive DMSO
concentration. The main part of the iNovacia fragment library is a diversity set where
the fragments have many analogues among in-house scaffolds in addition to many larger
compounds from the in-house compound collection where the fragments are present as sub-
structures. A large fraction of the fragments have masked synthetic handles to allow quick
expansion with one-step chemistries. For example, diverse fragments containing an amino
group were purchased and converted into acetamides using parallel one-step chemistry. The
acetamides are entered into the fragment library and if one of them is a hit, the acetamide
group can easily be exchanged into a larger amide by the use of the corresponding amino
fragment. Another example is the purchase of aryl bromides and boronic acids from which
