Chemistry Reference
In-Depth Information
not be a critical concern. As discussed above, if structure-based drug design by X-ray
crystallography is going to be used to drive the chemistry after the fragment screening,
large amounts of target protein will be necessary in any case. With a target protein that
is readily produced in large quantities, the question of whether the NMR fragment screen
will require 1 or 3 mg of target protein may not be a burning issue. When the availability
of target protein is scarce, recycling of protein should be considered. After the fragment
screen (or part of the fragment screen), the samples are collected, pooled and subjected
to exhaustive dialysis in order to remove the fragments and DMSO. A prerequisite is
high stability of the target protein. In this context, the lower fragment concentrations used
with the transverse relaxation filter technique is advantageous compared with STD and
WaterLOGSY.
4.5.4 Competition Experiments
The purpose of a competition experiment is to obtain information on the binding site of
fragment hits and to rank them with respect to affinity. If a protein-detected method has
been used, this information is already at hand. The same applies if displacement of a spy
molecule has been monitored in the primary fragment screen. Otherwise, either a previ-
ously known high affinity binder with known binding site (e.g. staurosporine for many
protein kinases) or a weakly binding spy molecule (e.g. ATP for protein kinases)
[
112
]
can
be used. The fragment hits binding to the same binding site, can be ranked with respect
to binding affinity (see Figure 4.3). If the
K
d
of the known binder, here denoted I, is
known (e.g. from an isothermal titration calorimetry measurement), then it is possible to
estimate the
K
d
for the primary hits, denoted L, by assessing the displacement caused
by the addition of the known binder, by using the relation
[
6, 113
]
K
d
(L)
=
[L]
K
I
/(
I
50
-
K
I
),
where
K
I
is the dissociation constant of the known binder used for competition and
I
50
is
the concentration of the known binder that causes the bound population of L to decrease
to half of what it was in the absence of the known binder. The concentration [L] can
be approximated to the total ligand concentration in cases where L has been added to
the sample in large molar excess compared with the target. Usually, however, the abso-
lute values of primary hit dissociation constants are not very important, but rather the
relative affinities.
The competition experiments also serve to find out whether the primary hits bind specific-
ally or nonspecifically. Fragments for which it is not possible to compete out by an excess
of a known high-affinity binder bind either specifically to another site or nonspecifically
to many hydrophobic patches on the protein surface. Alternatively, nondenaturing ESI-MS
may have to be applied to find out whether fragments that are not displaced by a known
binder bind specifically to another site or if they are nonspecific binders. If there are frag-
ment hits that bind specifically to nonoverlapping binding sites, it is worthwhile to test if it
is possible to observe interligand NOEs, especially if it has not been possible to determine
any structures of fragment-target complexes. The observed interligand NOEs show that the
fragments bind simultaneously at adjacent sites and the observed NOEs should give hints
on how to link the two fragments or expand one of them.
[
79, 80
]
Fast exchanging ligands with
overlapping binding sites can also exhibit interligand NOEs that are mediated via protons
in the target protein.
[
114
]
It is possible to exploit these interligand NOEs to determine the
relative orientation of the two competing ligands in the binding pocket.
