Chemistry Reference
In-Depth Information
(clusters 2 and 7) are similar to each other, exhibiting a translation and small tilt in the
binding pocket. The members of the higher COST cluster (cluster 1) show a much larger
translation and are rotated by
60° in the binding pocket with respect to the poses in
clusters 2 and 7.
5.7 Consequences of Protein Structure Selection on NOE Matching
As mentioned, protein conformational variability presents a significant challenge for NOE
matching and also for all other computational docking/scoring procedures. Although not
attempting to address this issue thoroughly in this chapter, we demonstrate the effects
of using a limited number of different protein conformations for the Bcl-x L / 7 complex.
In the previous Bcl-x L / 7 example, we had used the protein coordinates from the NMR
structure of Bcl-x L taken from the Bcl-x L / 7 complex as our starting structure. Since this
structure itself was modeled primarily by using assigned protein-ligand NOEs [ 14 ] in con-
junction with a (mostly fixed) starting structure, the possibility of a significantly different
protein conformation with 7 bound cannot be excluded. Because numerous structures
have been determined for Bcl-x L , both in the absence and presence of bound ligands,
we had the opportunity to address the consequences of initial protein structure selection
on the NOE matching outcome. We ran several more test cases, using protein coordin-
ates generated by different methods. These included apo Bcl-x L structures [ 33 ] determined
by either NMR (PDB entry 1LXL) or X-ray crystallography (PDB entry 1MAZ) and the
structure of a 16-residue BAK peptide-Bcl-x L complex determined by NMR (PDB entry
1BXL). [ 34 ]
In the first test case, we used the Bcl-x L coordinates from the apo NMR structure (1LXL).
The RMSD of backbone atoms involved in secondary structure of 1LXL is 2.03 Å from
1YSG. Trial binding poses of 7 were generated with Poser using a binding site box set
to ligand binding site plus an additional 1 Å in all coordinate axes and a 5° rotational
sampling. Over 728 000 poses were evaluated, with only 184 poses being retained by
Poser . The low number of poses retained by Poser reflects both the compressed binding
site typically found in NMR structures due to the force fields used during refinement and
real structural differences between apo and bound structures that partially occlude the
binding site. NOE matching was applied using our 57 experimental intermolecular NOEs
and with the predicted protein resonance assignments set to the BMRB average values. The
COST values from NOE matching ranged from 1745 to 4562, as shown in Figure 5.15A.
Clustering of the poses based on RMSD and a similarity of 0.8 Å resulted in 18 clusters, of
which four were singletons. Comparison with the poses in each of the clusters to the pose
for 7 found in the Bcl-x L / 7 complex is complicated by the structural differences between
the apo and bound protein structures. However, if one superimposes the protein backbone
atoms involved in secondary structure and then calculates the RMSD for resultant positions
of 7 in each of the structures, the pose with the lowest COST has an RMSD of 3.32 Å with
the target pose. For the ensemble of Poser trial poses, RMSDs to the target pose ranged
from 3.16 to 7.50 Å. As mentioned above and shown in Figure 5.15B, the addition of 7
causes some residues in the binding pocket of apo Bcl-x L to adopt different conformations
in bound Bcl-x L . For example, F97, Y101 and R139 rearrange upon complex formation
with the largest movement occurring with Y101.
Search WWH ::




Custom Search