Application of Protein–Ligand NOE Matching to the Rapid Evaluation of Fragment Binding Poses - Fragment-Based Drug Discovery - page 104

For LFA-1/ 2 , the pose with the minimum COST value has an RMSD of 0.91 Å to the
target pose. The target pose itself ranks 73 out of 1501 poses. LFA-1 was not experimentally
assigned for this complex, so statistics regarding correctly assigned 1 H 13 C groups are not
available for this case. While NOEmatching was successful overall for LFA-1/ 2 , the results
were not as good as those obtained with mFABP/ 1 . Although the lowest COST poses
cluster around the known target pose, there are few poses with relatively low COST that
are quite different from known poses. Discriminating among substantially different poses
with similar low COST, in the absence of a known pose, is an issue that can be addressed,
in part, by developments discussed later in this chapter.
Although the exact composition of compound 3 (Scheme 5.1) cannot be revealed at
this time, it is more favorable than compound 2 for NOE matching. Compound 3 has a
core that contains two methyl groups, both of which give rise to protein-ligand NOEs.
For LFA-1/ 3 , the target pose was derived from a well-resolved experimental ensemble,
generated as described (Supporting Information for ref. 15) from NMR and X-ray data.
A total of 350 trial poses were generated for LFA-1/ 3 , with minimum and maximum
RMSDs to the trial pose of 0.18 and 7.63 Å, respectively. A single protein conformation
was used in this case. The experimental 3D X-filtered NOESY spectrum ( τ m =

100 ms)
for this complex yielded 74 peaks, which were clustered into 44 protein 1 H 13 C groups.
Results for LFA-1/ 3 (SDM = 0.25) are shown in Figure 5.5 ( r = 0.973 between COST

and RMSD).

(a) (b)

7000

6500

6000

5500

5000

4500

4000

3500

3000

0 1 2 3 4 5 6 7 8

RMSD TO TARGET

Figure 5.5 (A) COST pose versus the RMSD (Å) to the target pose for LFA-1/ 3 with SDM 0.25

and with predicted protein chemical shifts set to the corresponding BMRB average values.
(B) Superposition of the target pose and the minimum cost pose (dark gray) from (A), showing
only the non-proprietary moieties.

=

The pose with the minimum COST value has an RMSD of 0.41 Å to the target pose.
The target pose itself ranks 11 out of 351 poses. As with mFABP/ 1 , a majority of the
experimentally identified 1 H 13 C groups are correctly assigned for low COST poses for
LFA-1/ 3 .