C. Neudorfer et al. / Bioorg. Med. Chem. Lett. xxx (2014) xxx–xxx
5
pyrazole H-5 (6.88 ppm) and anthranyl H-8 (7.73 ppm) is detected,
the appropriate effect is considerably smaller between pyrazole
H-5 and anthranyl H-1 (8.47 ppm). Moreover, a distinct NOE
between the acetyl-H protons and anthranyl H-8 is evident, in
contrast, the interaction between acetyl protons and anthranyl
H-1 is very small. The relatively large chemical shift difference
between accordant anthranyl protons H-1 (8.47 ppm) and H-8
(7.73 ppm) can be explained by the magnetic anisotropy effect of
the acetyl C@O group which obviously influences H-1 considerably
more than H-8 in the preferred rotameric status at hand. In the
NOESY spectrum of 17 the signals due to anthranyl H-1 and H-8
exhibit
a characteristic cross peak resulting from chemical
exchange (different phase property compared to the cross peaks
originating from NOEs), thus giving a hint to a slow rotation of
the anthranyl unit compared to the NMR timescale.
In order to estimate the binding affinity of the synthesized
compounds, they were docked into the crystal structure of the
human MAO-B in complex with FAD and noncovalently bound
p-nitro-benzylamine (PDB code 2C70).
Compounds 16–20 and 28–31, as well as reference compounds
34 and 35, being the most active substances in Mishra and
Sasmal,13 were prepared for a physiological pH using LigPrep
(LigPrep, version 2.5, Schrödinger, LLC, New York, NY, 2012).
As the acetyl- or thiocarbamoyl substituents on the pyrazoline
moiety had to be accommodated within a relatively narrow bind-
ing pocket, side chains of the residues in the hydrophobic pocket
(I198 and Y435) were allowed to take possible rotamer orienta-
tions defined within the GOLD docking software.22 A maximum
of ten poses per ligand was allowed, primarily ranked by the
default scoring function, ChemPLP.23 In order to reduce strain ener-
gies caused by the placement algorithm, the complexes were
energy minimized using a stepwise relaxation protocol within
the MOE software package (Molecular Operating Environment
(MOE) version 2013.0801, Chemical Computing Group, Montreal,
Canada). Subsequently, the binding free energy of the minimized
complexes was estimated by external rescoring using X-Score,24
as the original scoring values of ChemPLP are per se dimensionless.
Ki values were further calculated from
DG values of the best scored
pose per ligand, using Eq. 1.
ꢀ
ꢁ
D
R ꢂ T
G
Ki ¼ exp
R ¼ 8:3144621 J=mol ꢂ K; T ¼ 298:15 Kelvin
ð1Þ
Analysis of the docking results indicates that the novel com-
pounds are able to be accommodated in a similar way as described
for the reference compounds 34 and 35. As expected, introduction
of bulky groups on the central heteroaromatic linker enforced a
slightly different orientation of the synthesized compounds.
Figure 3. MAO-B binding pocket in complex with reference compounds and
exemplary docking poses of methyl- and fluoroethyl ester derivatives. Key residues
are highlighted in orange. (A) Reference compounds 34 (cyan) and 35 (green) in
superposition. (B) Docking poses of 34 (cyan) and 19 (purple). (C) Docked
complexes of 34 (cyan) and 29 (purple).
Figure
3 shows the most prominent interaction among the
observed poses, a pi stacking interaction between Y435 and ring
A or C of the anthracene moiety. Simultaneously, the phenolic
hydroxy group of the aromatic side chain was frequently observed
to stabilize the substituted heteroaromatic linker (acetyl or
thiocarbamoyl moiety). Potentially less favorable side chain
orientations caused by the bulky substituents hence could be
compensated. In analogy to the nitro group of compound 34, the
carbonyl oxygen of the distal ester function among the investi-
gated molecules sporadically acted as hydrogen bond acceptor
for the side chain of Y326. Due to the absence of an acceptor atom
in the corresponding position, reference compound 35 is unable to
form this particular interaction, which could be an explanation for
the inferior experimentally determined Ki value. In general, the
calculated Ki values were about an order of magnitude larger than
the literature values of 34 and 35. However, as demonstrated in
Table 1, it can be assumed that the activity of compounds 16–20
and 28–31 is in the range of the reference compounds (34 and
35). As the significant activity difference between 34 and 35 was
at least indicated by the calculated
DG values, the energy estimates
could as well be used for prioritizing new members of this com-
pound class prior to in vitro testing (Fig. 4).
Taken together, thirteen new target compounds have been pre-
pared within the scope of this work, which aimed at the develop-
ment of new, selective and reversible PET tracer precursors and
reference compounds for the imaging of the MAO-B system via
PET. Substitution of the pyrazoline ring with an additional acetyl
moiety at position 1 of the pyrazoline ring has been proven