Chemistry Reference
In-Depth Information
Poser . Due to the simplicity of the fragment, only one low-energy ligand conformation was
used as input. Trial binding poses were generated using Poser allowing a 10º rotational
sampling. The 'posing box' was defined as the compound binding site plus an additional
3 Å in all coordinate axes. In total, over 191 million poses were generated and evaluated
by Poser , with 9556 poses being retained. The NOE matching protocol was run using
BMRB-predicted chemical shifts. The experimental NOE list for NOE matching on Bcl-
x L / 7 was derived from a 3D 13 C-edited NOESY spectrum ( τ m
100 ms) recorded on
a sample consisting of 3.0 mM 7 -1.2 mM 13 C/ 15 N Bcl-x L . The data were recorded on a
Varian INOVA spectrometer at 303 K equipped with a Cold Probe. We identified a total
of 57 NOEs. (We note that the published NOE list, which was unavailable for our use,
contained 75 NOEs. [ 12 ] ) The 57 experimental NOEs were clustered into 30 protein 1 H 13 C
groups for NOE matching.
The results obtained from applying NOE matching to Bcl-x L / 7 are shown in Figure 5.11.
The pose with the minimum COST value has an RMSD of 0.71 Å to the target pose;
this pose had the lowest RMSD of all 9556 with respect to the target pose. As shown in
Figure 5.11B, the pose with the lowest COST overlays nicely with the target pose.
=
(a)
(b)
6000
5000
4000
3000
2000
1000
0
0
2
4 6
RMSD to Target
8
10
Figure 5.11 (A) COST versus the RMSD (Å) to the target pose for Bcl-x L / 7 . The predicted
protein chemical shifts were set to the corresponding BMRB average values. Cross peaks
from the experimental 3D X-filtered NOESY spectrum were used as input for NOE matching.
(B) Superposition of target pose and the minimum cost pose (dark gray) from (A).
5.6 Pose Analysis in the Absence of a Target Structure
In all of the cases examined above, the results of NOEmatching were evaluated with respect
to the target pose, that is, trial poses were compared directly with a known structure. Poses
with the lowest COST were bound similarly to the target pose (low RMSD), whereas poses
with higher COST tended to be bound more dissimilarly to the target pose (higher RMSD).
Although this is important to demonstrate in test cases, one will not be able to rely on such
a measure in practice (as the target pose will be unknown). An alternative way to look at
Search WWH ::




Custom Search