S. Hofer et al. / Bioorg. Med. Chem. Lett. 23 (2013) 5397–5400
5399
afforded the target compounds 9–21 in high yields by reaction with
the appropriate substituted phenylsulfonylchlorides. Compound 22
was obtained in high yield by reacting 4-methyl benzoylchloride
with 2-piperazin-1-yl-benzimidazole.
All the compounds were characterized by IR and 1H NMR
spectroscopy. The purity of the compounds was determined by
elemental analyses and the data for C, H, N are within 0.4% of
the calculated values.
11b-HSD1 inhibitory activities of the synthesized compounds
were first evaluated at 20 lM in a cell-free assay (Table 1). At this
concentration, compounds 16, 18 and 21 did not reduce 11b-HSD1
activity, and compounds 9, 17, 19 and 22 showed 50% or less
inhibition at 20 lM. Among compounds10–15, bearing a 1-methyl
benzimidazol-2-yl moiety, 14 and 15 turned out to be weak inhib-
itors of the enzyme, whereas compounds 10–13 were potent inhib-
itors with IC50 of 0.7 0.2–6.4 0.6 lM. The nature of the
substituent on the phenyl ring within this class of compounds
revealed that the 4-methyl functionality (11) was the most favour-
able (IC50 = 0.7 0.2
(12, IC50 = 4.8 1.4
l
M). Exchanging the methyl for a tert-butyl
lM) or methoxy (13, IC50 = 4.2 2.2 M) in
l
the para-position results in a reduction of activity by a factor of
ꢀ5, both of these compounds were found to be equipotent.
Compared to the mono-methoxy compound 13, a dimethoxy
substitution (compound 14) led to further reduction in 11b-HSD1
inhibition by a factor ꢀ4 (Table 1). Compounds bearing an acidic
(16 and 17) or highly polar (18 and 19) functionalities showed
no or very low inhibitory effect at 20 lM. Another important
observation from the biological data is that the replacement of
the 1-methyl benzimidazol-2-yl by the isosteric benzimidazol-2-
yl (9) or a benzothiazol-2-yl (21) ring, albeit they share the same
substituent (methyl), results in compounds with weak or no inhib-
itory effect on 11b-HSD1 activity. Replacement by the benzoxazol-
2-yl (20) resulted in a poorly soluble derivative that was not
further investigated. Compound 22, which bears
a carbonyl
functionality instead of a sulfone showed weak inhibitory activity.
Assessment of 11b-HSD2 activity (performed as described
earlier11) revealed that none of the compounds analysed in the
present study inhibited this enzyme at a concentration of 20 lM.
The biological evaluation of the newly synthesized compounds
revealed that electron donating substituents (e.g., methyl, methoxy
or tert-butyl groups) in the benzene moiety are tolerated by the
enzyme, whereas acidic or electron withdrawing groups are unfa-
vourable for the inhibitory activity. To rationalize the biological
data, we predicted binding modes for all compounds that were
tested in the 11b-HSD1 assay using molecular docking. For the
docking, an X-ray crystal structure, where the co-crystallized
ligand belongs to the same chemical scaffold as our compounds
was chosen15 (PDB-code 3czr). The docking studies were per-
formed using GOLD,16,17 that uses a genetic algorithm for predict-
ing the binding orientations. The binding site was defined as a 7 Å
sphere, centered by the hydroxyl-oxygen of Tyr-177. To ensure the
optimal ligand positioning, a hydrogen bond constraint between
the ligand and backbone nitrogen of Ala-172 was set. GoldScore
was used as a scoring function. The program was allowed to termi-
nate the docking run in case the best ranked solutions were within
1 RMSD of each other, otherwise ten binding modes were pre-
dicted. With these settings, the program reproduced the binding
orientation of the cocrystallized ligand, thus validating the docking
settings. All of the active compounds 9–15 and 20 in their best-
ranked poses aligned well with the cocrystallized ligand. In view
of the predicted binding orientation, the hydrophobic substituents
of the compounds occupied a hydrophobic pocket located next to
the cofactor (Fig. 2A). Nonpolar substituents formed hydrophobic
interactions with the enzyme (Fig. 2B and C). Therefore, the inac-
tivity of the compounds with acidic or electron withdrawing
substituents could be caused by a lack of favourable interaction
Figure 2. The predicted binding orientations of the active compounds (9–15 and
20, A), ligand–protein interactions of compounds 11 (B) and 12 (C), and binding
orientations of the inactive compounds (D). The active compounds (grey) align with
the cocrystallized ligand (magenta) and occupy a hydrophobic pocket formed by
Ile-121, Thr-124, Thr-222, Ala-223 and Ala-226 (depicted in ball and stick style).
Inactive compounds 21 and 22 do not align with the cocrystallized ligand and stick
out of the binding pocket. Other inactive compounds have polar substituents in the
hydrophobic pocket, suggesting unfavourable interactions. The catalytic triad Ser-
170–Tyr-183–Lys-187 and the cofactor are highlighted in stick style. In the 2D-
representations, the ligand–protein interactions are colour-coded as following:
hydrophobic interaction—yellow, hydrogen bond acceptor—red arrow.
O
O2N
N
N
S
O
Figure
3. Compound
45
(2R-1-(4-tert-butylphenyl)sulfonyl-2-methyl-4-
(4-nitrophenyl)piperazine).15
partners in this area. In addition, the inactive compounds with
nonpolar substituents (compounds 21 and 22) have a different
binding orientation (Fig. 2D), and stick out of the binding pocket.
Previously, Sun et al.,15 have described synthesis- and
SAR-studies focusing on a similar 11b-HSD1 inhibitor scaffold.
Hofer, Sandra
