26
S. Ulmschneider et al. / Bioorg. Med. Chem. Lett. 16 (2006) 25–30
Figure 1. Superimposition of compounds: (a) alignment of inhibitors from 17 to 39 and 66 to 81 (see Table 1); (b) skeletal structure of the
pharmacophore model: overlay of the unsubstituted 2-(3-pyridyl) naphthalene (66 yellow) and the unsubstituted E- and Z-3-pyridylmethylene-indane
(17 and 18; red and green); the pharmacophore (magenta) is built by connecting the ring centroids (C1, C2, and C2b) and the aromatic nitrogen N;
(c) sideview of the pharmacophore.
of 1.0 has been used throughout and the optimizations
have been terminated at a gradient of 0.001 kcal/mol.
After defining the ring centroids for each compound of
the three classes (C1, C2, and C2b, Fig. 1b), the com-
pounds 17–39—with the exception of the non-inhibi-
tors—were first superimposed on compounds 20 and
21 with Sybyl by a three-point-fit strategy using the very
same atoms for E-(red) and Z-(green) 3-pyridyl substi-
tuted tetrahydronaphthalenes and -indanes (Fig. 1a).15
The steric features of active and non-active CYP11B2-
inhibitors were further explored by the Ôsteric inclusion
area (SIA)Õ and the Ôsteric exclusion area (SEA),Õ respec-
tively (Figs. 2a and b). The SIA is mainly located in the
region of substituents R3, R4, and R8 (see Table 2).
Larger groups as formic acid methyl ester can only be
introduced in position R4 (72), whereas position R8 is
of limited size. The SEA is located at the non-aromatic
ring and the exocyclic double bond of the Z-isomer
(substituents R1 and R2, Table 2). Additionally, large
substituents as benzyloxy- or 3-pyridyl-groups in the re-
gion of R4 and R5 are not suitable for proper inhibitor
binding (32 and 79). The SEA is also located close to
R9, since methyl-groups lead to a strong decrease of
activity (34).
The naphthalene inhibitors (66–69, 72–74, and 76–78;
yellow, Fig. 1a) were aligned with both E- and Z-(3-pyr-
idyl)methylene-indanes and -tetrahydronaphthalenes,
respectively.16,17 These naphthalenes as well as the E-
and Z-imidazolyl substituted tetrahydronaphthalene
and -indane inhibitors (not shown) fitted very well into
the pharmacophore (Fig. 1b). The superimposition of
non-inhibitors—using the same procedure as applied
for the inhibitors—revealed that the corresponding sub-
stituents did not match with the model (Table 2).
To validate the model, a compound was generated as
a hybrid structure of the used E- and Z-isomers as
well as the naphthalene compounds, the acenaphthene
derivative (A). Compound A was synthesized in four
steps: nitration of acenaphthene18 and subsequently
hydrogenation19 leading to a mixture of two isomers:
3- and 5-aminoacenaphthene. In the following Sand-
meyer reaction, the bromo derivatives were formed
and used for Suzuki coupling with 3-pyridine boronic
acid. The resulting mixture was subsequently chro-
matographed and the isomers A and B were isolated
(Scheme 1).
The pharmacophore model (Fig. 1b) was built on four
pharmacophore points, chosen on the basis of the global
superimposition (Fig. 1):
• N: heterocyclic nitrogen (all inhibitors).
• C1: ring centroid (E-(3-pyridyl)methylene-tetrahy-
dronaphthalenes and -indanes and naphthalene
derivatives).
• C2: ring centroid (E- and Z-(3-pyridyl)-methylene-
indanes and -tetrahydronaphthalenes).
• C2b: ring centroid (Z-(3-pyridyl)methylene-tetrahy-
dronaphthalenes and -indanes and naphthalene
derivatives).
Both isomers were tested for activity in V79 cells,12
expressing human CYP11B2.6 The IC50 values of the
compounds are given in Table 4.
As expected compound A exhibited strong inhibitory
activity (IC50 = 10 nM), thus confirming the validity of
the pharmacophore model. Surprisingly the isomer B
was also very potent. This can be explained by the align-
ment of the acenaphthene derivatives A and B (Fig. 3).
The compounds display a very similar shape; they only
differ in the position of the non-aromatic cyclopenten-
e-ring.
Each compound fits with three of these points and has
nitrogen N in common. The lone pair of the heterocyclic
nitrogen of all compounds has to point in almost the
same direction (Fig. 1a) for complexing the heme iron.
The three ring centroids (C1, C2, and C2b) are located
in one plane, P0, forming a planar three-ring-system
(ace-naphthene or dihydro-phenalene), while the aro-
matic nitrogen N is located slightly above plane P0
(see geometric parameters, Table 3). Actually, the whole
pharmacophore seems to be quite planar.
After having discovered the acenaphthene isomers
as highly potent lead compounds, we wanted to
know about their selectivity toward CYP11B1 (Table 4).