Enantiomerically Pure 2-Methyltetrahydro-3-benzazepin-1-ols Selectively Blocking GluN2B Subunit Containing N-Methyl-d-aspartate Receptors

Article abstract of DOI:10.1021/acs.jmedchem.5b00897

A chiral pool synthesis was developed to obtain all four stereoisomeric 2-methyl-3-(4-phenylbutyl)tetrahydro-3-benzazepin-1-ols 21, 31, and 32 in a seven- to eight-step sequence. The phenols 32 reveal slightly higher GluN2B affinity than the methyl ethers 21. The GluN2B affinity increases in the order (1R,2S) < (1S,2S) < (1S,2R) < (1R,2R). The stereoisomeric phenols (R,R)-32 and (S,R)-32 show the highest GluN2B affinity and the highest cytoprotective activity. Both compounds represent GluN2B selective allosteric NMDA receptor antagonists. Docking of the 3-benzazepin-1-ols into the ifenprodil binding site of the crystallized GluN1b/GluN2B N-terminal domains led to free binding energies, which correlate nicely with the experimentally determined GluN2B affinities. The similar GluN2B affinity of the stereoisomeric phenols (S,S)-32, (R,R)-32, and (S,R)-32 is explained by different binding modes of the 3-benzazepine scaffold. The benzyl ethers 31 reveal unexpectedly high GluN2B affinity but do not show cytoprotective effects. The additional benzyl moiety of 31 binds into a previously unrecognized lipophilic subpocket.

Full text of DOI:10.1021/acs.jmedchem.5b00897

Article
pubs.acs.org/jmc
Enantiomerically Pure 2‑Methyltetrahydro-3-benzazepin-1-ols
Selectively Blocking GluN2B Subunit Containing
N‑Methyl‑D‑aspartate Receptors
Bastian Tewes,† Bastian Frehland,† Dirk Schepmann,† Dina Robaa,‡ Tanaporn Uengwetwanit,‡
Friedemann Gaube,§ Thomas Winckler,§ Wolfgang Sippl,‡ and Bernhard Wunsch*,†,∥
̈
†Institut fur Pharmazeutische und Medizinische Chemie der Universitat Munster, Corrensstraße 48, D-48149 Munster, Germany
̈
̈
̈
̈
‡Institut fur Pharmazie, Martin-Luther-Universitat Halle-Wittenberg, Wolfgang-Langenbeck-Strasse 4, D-06120 Halle/Saale, Germany
̈
̈
§Institut fur Pharmazie der Friedrich-Schiller-Universitat Jena, Semmelweisstraße 10, D-07743 Jena, Germany
∥Cells-in-Motion Cluster of Excellence (EXC 1003-CiM), Universitat Munster, 48149 Munster, Germany
̈
̈
̈
̈
̈
S
* Supporting Information
ABSTRACT: A chiral pool synthesis was developed to obtain
all four stereoisomeric 2-methyl-3-(4-phenylbutyl)tetrahydro-
3-benzazepin-1-ols 21, 31, and 32 in a seven- to eight-step
sequence. The phenols 32 reveal slightly higher GluN2B
affinity than the methyl ethers 21. The GluN2B affinity
increases in the order (1R,2S) < (1S,2S) < (1S,2R) < (1R,2R).
The stereoisomeric phenols (R,R)-32 and (S,R)-32 show the
highest GluN2B affinity and the highest cytoprotective activity.
Both compounds represent GluN2B selective allosteric
NMDA receptor antagonists. Docking of the 3-benzazepin-1-
ols into the ifenprodil binding site of the crystallized GluN1b/
GluN2B N-terminal domains led to free binding energies, which correlate nicely with the experimentally determined GluN2B
affinities. The similar GluN2B affinity of the stereoisomeric phenols (S,S)-32, (R,R)-32, and (S,R)-32 is explained by different
binding modes of the 3-benzazepine scaffold. The benzyl ethers 31 reveal unexpectedly high GluN2B affinity but do not show
cytoprotective effects. The additional benzyl moiety of 31 binds into a previously unrecognized lipophilic subpocket.
block has to be removed by predepolarization of the
surrounding neuronal membrane, e.g., by activation of AMPA
or kainate receptors adjacent to the NMDA receptor. (2)
Activation of the NMDA receptor requires two agonists, (S)-
glutamate and the so-called coagonist glycine. (3) The NMDA
receptor controls the influx of Ca2+ ions. The physiological
concentration of Ca2+ ions is important for the development of
neurons, but an increased intracellular Ca2+ ion concentration is
associated with damage of neurons (excitotoxicity). (4) In
addition to the agonists (S)-glutamate and glycine, several other
small molecules are able to modulate the opening state of the
NMDA receptor by interacting with particular binding sites.
The receptor contains binding sites for phencyclidine (PCP)
within the ion channel, Zn2+ ions, protons, NO, polyamines,
and ifenprodil (1) at the amino terminal part of the receptor
protein.1,2
INTRODUCTION
■
The most important excitatory neurotransmitter in the central
nervous system is the amino acid (S)-glutamate. This
neurotransmitter mediates its effects via activation of
metabotropic (mGlu) and ionotropic glutamate receptors
(iGlu). According to their sequence homology, their signal
transduction pathway, and their ligand binding profile, the G-
protein-coupled metabotropic glutamate receptors are sub-
divided into three groups: group I contains the receptor
subtypes mGlu1 and mGlu5; group II contains the subtypes
mGlu2 and mGlu3, and the residual metabotropic glutamate
receptors mGlu4, mGlu6, mGlu7, and mGlu8 belong to group
III metabotropic glutamate receptors. In the field of ionotropic
glutamate receptors three subtypes have been identified, which
were termed according to their prototypical ligands: N-methyl-
D-aspartate gave name to the NMDA receptor; 2-amino-3-(5-
hydroxy-2-methylisoxazolyl)propionic acid is the prototypical
ligand for the AMPA receptor; the natural product kainic acid
isolated from Japanese red algae was the first compound that
activated kainate receptors.1,2
The NMDA receptor is a heterotetrameric protein. Cloning
of the different NMDA receptor subunits led to seven proteins,
which are classified into three types of subunits: GluN1 subunit
with eight splice variants termed GluN1a−h, four GluN2 and
The NMDA receptor is characterized by four particular
features: (1) at normal membrane potential the receptor is
blocked by Mg2+ ions. In order to open the channel, the Mg2+
Received: June 11, 2015
© XXXX American Chemical Society
A
DOI: 10.1021/acs.jmedchem.5b00897
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
two GluN3 subunits encoded by different genes and termed
GluN2A−D and GluN3A,B, respectively. A functional NMDA
receptor protein contains at least one GluN1 subunit bearing
the glycine binding site and one GluN2 subunit bearing the
(S)-glutamate binding site.3−5
development of 2 was discontinued, although it showed an
excellent side effect profile.14 However, the promising neuro-
protective effects of these compounds in animal assays (e.g.,
model of Parkinson’s disease)15 stimulated the further
development of GluN2B antagonists.
Herein we focus on novel ligands for the ifenprodil binding
site of the NMDA receptor. The ifenprodil binding site is
special because it is located at the interface between the N-
terminal domains of the GluN2B and GluN1 subunits.6 The N-
terminal domains of GluN2A and GluN2B subunits, which
have a clamshell-like structure, are responsible for the correct
subunit assembly.6,7 Whereas the GluN1 subunits are expressed
ubiquitously throughout the central nervous system, the
expression of the GluN2 subunits depends on the age and
the region of the human brain. For example, the GluN2B
subunit is widely expressed in the prenatal brain, but in the
adult brain it is found predominantly in the cortex, striatum,
and hippocampus. The amount of the GluN2B subunit in the
adult cerebellum is rather low, while GluN2C subunits are
expressed predominantly in this region of the brain.8 Ligands
modulating the opening state of NMDA receptors by
interaction with the ifenprodil binding site selectively address
GluN2B containing NMDA receptors but do not influence the
activity of the NMDA receptors containing GluN2A, GluN2C,
and GluN2D subunits, which leads to an optimized side effect
profile.9
Binding of 1 and its analogs to the NMDA receptor takes
place in a special binding site, which is termed ifenprodil
binding site. GluN1 and GluN2B N-terminal domains form
heterodimers, where residues lining the dimer interface
construct the binding pocket for the phenylethanolamine
derivatives. As can be seen from the crystal structures of the
heterodimer of the N-terminal domains of GluN1b and
GluN2B subunits in complex with 1 or its analog 3 (PDB
codes 3QEL and 3QEM; respectively), key polar and
hydrophobic interactions participate in the stabilization of the
antagonists in their binding pocket.6 These interactions are
discussed in detail in the section Molecular Modeling Studies.
Recently we have reported on the synthesis and pharmaco-
logical evaluation of tetrahydro-3-benzazepines 5−7, which
show high affinity to the ifenprodil binding site of GluN2B
containing NMDA receptors. The phenol 5 results from a
formal rearrangement of 1, i.e., disconnection of the C-3/C-4-
bond of the piperidine ring and reconnection of the free
methylene moiety to the benzene ring (Figure 2). Introduction
Compounds blocking GluN2B containing NMDA receptors
(GluN2B antagonists) by interacting with the ifenprodil
binding site have a high neuroprotective potential, which can
be exploited for the treatment of traumatic brain injury, stroke
(cerebral ischemia), neuropathic pain,10 and Parkinson’s
disease. Moreover, GluN2B antagonists could be useful for
the therapy of migraine, depression, and alcohol withdrawal
symptoms.1,2,8
In the 1990s it was shown that 1 (ifenprodil), which was
originally developed as α1 receptor antagonist, inhibits
selectively GluN2B containing NMDA receptors (IC50 = 13.3
nM) (Figure 1).11 The low selectivity of 1 addressing
Figure 2. Development of the methylated 3-benzazepinols 8.
of the flexible phenylethylamine part of 1 into the 3-
benzazepine scaffold of 5 retains high GluN2B affinity (Ki(5)
= 14 nM) but increased the receptor selectivity. The methyl
ether 6 represents an even more affine and selective GluN2B
antagonist (Ki(6) = 5.4 nM) but showed reduced antagonistic
effects. The GluN2B affinity of the benzyl ether 7 was
considerably reduced (Ki(7) = 187 nM).16,17
A comparison of the 3-benzazepines 5−7 with the lead
compound 1 shows that the methyl moiety in β-position to the
OH moiety of 1 is missing. In order to investigate the role of
this methyl moiety on the GluN2B affinity, we planned to
synthesize and pharmacologically evaluate 3-benzazepines 8
with an additional methyl moiety in β-position to the OH
group. Since the 3-benzazepines of type 8 contain two centers
of chirality, a stereoselective synthesis of all four stereoisomers
was envisaged.
Figure 1. Prominent GluN2B antagonists.
additionally some types of Ca2+ channels, 5-HT1A, 5-HT2, α1,
σ1, and σ2 receptors led to side effects (e.g., psychotomimetic
effects, memory deficit, hypertension). Moreover, the low
bioavailability of 1 was due to fast metabolic degradation.12
Nevertheless 1 has served as a lead compound for the
development of potent, selective, and bioavailable GluN2B
containing NMDA receptor antagonists. In Figure 1 the
prominent GluN2B antagonists traxoprodil (2), Ro 25-6981
(3), and besonprodil (4) derived from 1 are shown.13 2 was
studied in phase II clinical studies for the treatment of
traumatic brain injury. Because of lack of efficacy, the
SYNTHESIS
■
For the synthesis of 3-benzazepines 8, a chiral pool synthesis
was envisaged using the enantiomerically pure proteinogenic
amino acid (S)-alanine ((S)-11) as the starting material
(Scheme 1). Esterification of (S)-11 with methanol and
B
DOI: 10.1021/acs.jmedchem.5b00897
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
a
Scheme 1
a
Reagents and reaction conditions: (a) (1) n-BuLi, THF, −78 °C, 2 h; (2) ethylene sulfate, THF, −78 °C → rt, 16 h; (3) H2SO4, H2O, 47 h, Δ,
62%.21 (b) (1) CH3OH, SOCl2, rt, 2 h, 95%; (2) TosCl, CH2Cl2, Et3N, rt, 20 h, 73%. (c) (S)-12, THF, 0 °C, then 10, PPh3, DIAD, 0 °C, 1 h, rt, 16
h, 91%. (d) LiOH·H2O, THF/H2O (7:3), rt, 24 h, 87%. (e) Trifluoroacetic acid anhydride ((CF3CO)2O), SnCl4, CH2Cl2, −30 °C, 22 h. (f)
LiBH4, THF, −90 °C, 24 h. The enantiomers (S,R)-18 and (R,R)-18 were prepared in the same manner starting with (R)-configured sulfonamide
(R)-1222 prepared from (R)-configured alanine (R)-11.
subsequent reaction with tosyl chloride led to the N-tosylated
methyl ester (S)-12,18,19 which should be alkylated with the
phenylethanol derivative 10 in a Mitsunobu reaction.20
Phenylethanol 10 was prepared very efficiently from
bromobenzene 9 in a one-pot, three-step procedure comprising
bromine−lithium exchange, reaction of the resulting aryllithium
intermediate with cyclic ethylene sulfate, and hydrolysis of the
sulfuric acid hemiester intermediate.21 The Mitsunobu reaction
of the secondary sulfonamide (S)-12 with the alcohol 10 using
PPh3 and diisopropyl azodicarboxylate (DIAD) provided the
tertiary sulfonamide (S)-13 in 71% yield. The preparation of
the enantiomer (R)-13 was supported by microwave irradiation
giving the (R)-configured sulfonamide (R)-13 in 67% yield.
Alkylation of the sulfonamide (S)-12 with the corresponding
phenylethyl iodide or phenylethyl tosylate provided the tertiary
sulfonamide (S)-13 only in very low yields (0−18%). In order
to avoid racemization, the hydrolysis of the ester (S)-13 was
performed very carefully using LiOH at room temperature to
afford the acid (S)-14 in 87% yield.
The intramolecular Friedel−Crafts acylation of the carboxylic
acid (S)-14 was performed with P4O10 in CH2Cl2 analogously
to the cyclization of the nonmethylated analog.16 Unexpectedly,
the desired compounds (S)-15 and (S)-16 with a seven-
membered ring were formed in only 4% yield, while the
corresponding isoquinolines 17 predominated (19% yield).
Several reaction conditions (see Table SI-1 in Supporting
Information) were investigated to improve the yield of this key
step, and finally it was found that transformation of the
carboxylic acid (S)-14 with trifluoroacetic acid anhydride
(CF3CO)2O into a mixed anhydride and subsequent cyclization
with SnCl4 provided 60% of a 1:1 mixture of the 3-
benzazepines (S)-15 and (S)-16 and only small amounts of
isoquinolines 17.
As the regioisomeric 3-benzazepines (S)-15 and (S)-16 could
not be separated by flash chromatography, the mixture was
reduced with LiBH4 at −90 °C and the resulting alcohols were
separated and isolated in 25% ((R,S)-18), 19% ((S,S)-18), and
51% ((S)-19) yields, respectively.
The final reaction steps comprised removal of the tosyl
protective group of (R,S)-18 and (S,S)-18 with magnesium in
methanol and alkylation of the secondary amines (R,S)-20 and
(S,S)-20 with 1-chloro-4-phenylbutane to yield the diastereo-
meric 3-benzazepines (R,S)-21 and (S,S)-21 (Scheme 2). The
corresponding enantiomers (S,R)-21 and (R,R)-21 were
prepared in the same manner starting with (R)-alanine ((R)-
11).
The enantiomeric purity of the final compounds 21 was
analyzed by chiral HPLC (Chiralpak AD). It was shown that
about 25% racemization had occurred during the synthesis
(Table SI-2). We assume that the intramolecular Friedel−
Crafts acylation of (S)-14 and (R)-14 is responsible for the
racemization. In order to test enantiomerically pure com-
pounds, the 3-benzazepines 21 were purified by preparative
chiral HPLC. Finally the four stereoisomers (R,S)-21, (S,S)-21,
(S,R)-21, and (R,R)-21 were isolated with an enantiomeric
purity greater than 99.1:0.9 (Table SI-2, Figure SI-1A−D).
The relative configuration of the 3-benzazepines 18, 20, and
21 could not be assigned unequivocally by analysis of the
C
DOI: 10.1021/acs.jmedchem.5b00897
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
a
respectively, the triisopropylsilyl protected phenol (S)-24d gave
only 35% yield indicating the instability of the phenyl silyl
ether.
Scheme 2
The intramolecular Friedel−Crafts acylation of the benzyl
ether (S)-24b using (CF3CO)2O and SnCl4 as for the
cyclization of the methyl ether (S)-14 led to a mixture of the
cyclic ketones (S)-25b and (S)-26b in 20% yield (Scheme 3).
Additionally the O-debenzylated 3-benzazepines (S)-25a and
(S)-26a (10% yield) and the O-debenzylated isoquinoline 27a
(7% yield) were isolated. Obviously a fast debenzylation had
occurred under the Lewis acid catalyzed Friedel−Crafts
acylation conditions. In order to increase the yield of the
benzyl protected 3-benzazepinone (S)-25b, the crude reaction
mixture was treated with benzyl bromide and K2CO3 to convert
the phenol (S)-25a into the benzyl ether (S)-25b. After several
optimization experiments including the subsequent benzylation
of the free phenol (S)-26a, a 95:5 mixture of regioisomeric 3-
benzazepinones (S)-25b and (S)-26b was isolated in 41% yield.
The higher regioselectivity of the cyclization of the benzyl ether
(S)-24b compared to the cyclization of the methyl ether (S)-14
is explained by shielding the o-position by the larger benzyloxy
moiety.
Cyclization of the p-methoxybenzyl ether (S)-24c using the
optimized reaction conditions ((CF3CO)2O, SnCl4) led also
to considerable cleavage of the p-methoxybenzyl ether and
provided, after treatment of the crude reaction product with p-
methoxybenzyl bromide and K2CO3, a 30:70-mixture of the
regioisomeric 3-benzazepinones (S)-25c and (S)-26c in 25%
yield (Scheme 3). After reaction of the silyl ether (S)-24d with
(CF3CO)2O and SnCl4, the silyl ether (S)-25d and the
phenol (S)-26a were isolated in 18% and 7% yields,
respectively. In this case only the silyl ether in o-position to
the ketone ((S)-26d) was cleaved, whereas the corresponding
regioisomeric silyl ether (S)-25d remained intact. Finally the
phenol (S)-24a, which was obtained by hydrogenolytic cleavage
of the benzyl ether (S)-24b, was cyclized with (CF3CO)2O
and SnCl4 to yield 9% of the desired 3-benzazepinone (S)-25a
and 11% of the ring-contracted isoquinoline 27a; the
corresponding 9-hydroxy-3-benzazepinone (S)-26a could not
be detected.
a
Reagents and reaction conditions: (a) Mg, CH3OH, Δ, 16 h. (b) 1-
Chloro-4-phenylbutane, CH3CN, Bu4NI, K2CO3, Δ, 72 h. The
enantiomers (S,R)-21 and (R,R)-21 were prepared in the same
manner starting with (S,R)-18 and (R,R)-18.
coupling constants in the NMR spectra because of undefined
dihedral angles in the flexible seven-membered ring system.
Therefore, X-ray crystal structures of (S,R)-21 (42% ee) and
(R,R)-21 (52% ee) were recorded.23 The X-ray crystal structure
(Figure SI-4) clearly shows trans-configuration of the OH
moiety and the methyl group, proving unlike-configuration of
(S,R)-21 (42% ee). On the other hand cis-orientation of these
substituents is displayed in Figure SI-5, proving like-
configuration of (R,R)-21 (52% ee).
In addition to the four diastereomeric methyl ethers 21, the
corresponding phenols 32 were prepared and tested. For this
purpose the methyl ether of the ketone (S)-15 should be
cleaved in analogy to the cleavage of the corresponding
nonmethylated analog.17 However, all attempts using AlCl3,
BBr3, or LiBH(sec-Bu)3 failed to give the corresponding phenol.
Therefore, the synthesis of the 3-benzazepine scaffold was
repeated with different phenol protective groups.
Thus, the phenolic OH moiety of 3-/2-hydroxyethyl)phenol
(22a) was converted selectively into the benzyl ether 22b, the
p-methoxybenzyl ether 22c, and the triisopropylsilyl ether 22d
(Scheme 3). Mitsunobu reaction of the different ethers 22b−d
with the alanine derived sulfonamide (S)-12 afforded the
coupling products (S)-23b−d in 71−78% yields, which were
hydrolyzed with LiOH to obtain the carboxylic acids 24b−d.
Whereas the carboxylic acids (S)-24b and (S)-24c with benzyl
protective groups were obtained in 73% and 96% yields,
Since the cyclization of the benzyl ether (S)-24b gave the
highest yield of the 3-benzazepinones (41%) and the best
regioselectivity (95:5), the 95:5 mixture (S)-25b/(S)-26b was
a
Scheme 3
a
Reagents and reaction conditions: (a) (S)-12, DIAD, Ph3P, THF, 0 °C → rt, 18 h. (b) LiOH·H2O, THF/H2O (7:3), rt, 16 h. (c) (24b) SnCl4,
trifluoroacetic acid anhydride, CH2Cl2, −15 °C, 23 h, then benzyl bromide, acetone, K2CO3, Δ, 6 h. (d) (24c) SnCl4, trifluoroacetic acid anhydride,
CH2Cl2, −15 °C, 23 h, then 4-methoxybenzyl bromide, acetone, K2CO3, Δ, 6 h. (e) (24d) SnCl4, trifluoroacetic acid anhydride, CH2Cl2, −15 °C, 23
h. The enantiomers (R)-25b and (R)-26b were prepared in the same manner starting with (R)-12.
D
DOI: 10.1021/acs.jmedchem.5b00897
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
further processed to access the final GluN2B antagonists.
Reduction with NaBH4 afforded the diastereomeric alcohols
(R,S)-28 and (S,S)-28 as well as the regioisomeric alcohol 29,
which were separated and purified by flash chromatography
(Scheme 4). Removal of the tosyl group from (R,S)-28 and
RECEPTOR AFFINITY
■
Affinity toward GluN2B Containing NMDA Receptors.
The affinity toward GluN2B containing NMDA receptors of
the stereoisomeric 2-methyl-3-benzazepin-1-ols 21, 31, and 32
was determined in a competitive receptor binding assay, which
was recently developed in our group.25 Membrane preparations
obtained by ultrasonic irradiation of L(tk−) cells stably
expressing recombinant human GluN1a and GluN2B subunits
forming functional NMDA receptors were used as receptor
material, and tritium labeled [3H]1 served as radioligand. The
high amount of GluN2B containing NMDA receptors renders
this system selective, despite the limited selectivity of the
radioligand. After the adherent growing cells reached
approximately 90% confluency, dexamethasone was added to
the growth medium to induce the synthesis of NMDA
receptors. During this period, cell death was inhibited by
addition of the NMDA receptor antagonist ketamine
interacting with the phencyclidine binding site within the
channel pore.25,26
a
Scheme 4
The affinity toward GluN2B containing NMDA receptors is
summarized in Table 1. The GluN2B affinity of the phenols 32
is generally slightly higher than the GluN2B affinity of the
methyl ethers 21, but they show the same relationship between
the stereochemistry and the GluN2B affinity. The affinity
increases in the following order: (1R,2S) < (1S,2S) < (1S,2R) <
(1R,2R). In both series the (R,R)-configured stereoisomers
show the highest GluN2B affinity of 17 nM ((R,R)-32) and 41
nM ((R,R)-21). Whereas the GluN2B affinity of the (R,S)-
configured stereoisomers (R,S)-21 and (R,S)-32 is considerably
reduced, the GluN2B affinity of the other three stereoisomers is
in a similar range, respectively.
The lead compounds 1 and 2 have (S)-configuration at both
centers of chirality. The corresponding (S,S)-configured
stereoisomers (S,S)-21 and (S,S)-32 display slightly reduced
GluN2B affinity compared with the (R,R)- and (S,R)-
configured stereoisomers. However, taking the SEM values
into account, the differences between the (S,S)-, (R,R)-, and
(S,R)-configured stereoisomers are very small. Nevertheless the
low stereoselectivity in the interaction with GluN2B receptors
was unexpected.
Moreover the high GluN2B affinity of the stereoisomeric
benzyl ethers 31, originally prepared as synthetic intermediates,
was very surprising. In general the GluN2B affinity of the
benzyl ethers 31 exceeds the GluN2B affinity of the methyl
ethers 21 and the phenols 32 (exception (S,R)-32). The most
potent stereoisomer (R,S)-31 reveals a Ki value of 9.3 nM,
which is even lower than the Ki values of the reference
compounds 1 and eliprodil.
In all three series of 3-benzazepines, 21, 31, and 32, the
additional methyl moiety has a considerable influence on the
GluN2B affinity. In cases of the phenols 32 and the methyl
ether 21 the additional methyl moiety leads to a slight
reduction of the GluN2B affinity compared to 5 and 6. On the
contrary the GluN2B affinity of the benzyl ethers 31 was
dramatically increased upon introduction of the methyl moiety
in 2-position, which can be seen by comparing the Ki values of
7 and 31 stereoisomers.
a
Reagents and reaction conditions: (a) NaBH4, CH3OH, 20 h, rt. (b)
Mg, CH3OH, Δ, 36 h. (c) 1-Chloro-4-phenylbutane, CH3CN, K2CO3,
Bu4NI, Δ, 72 h. (d) H2, Pd/C, 1 bar, CH3OH, 1 h, rt. The
enantiomers (S,R)-31, (S,R)-32, (R,R)-31, and (R,R)-32 were
prepared in the same manner starting with (R)-25b and (R)-26b.
(S,S)-28 was performed with magnesium in methanol,
providing the secondary amines (R,S)-30 and (S,S)-30 in
65% and 60% yields, respectively, which were alkylated
subsequently with 1-chloro-4-phenylbutane to yield the tertiary
amines (R,S)-31 and (S,S)-31. The corresponding enantiomers
(S,R)-31 and (R,R)-31 were synthesized analogously starting
with (R)-alanine ((R)-11).
Analysis of the enantiomeric purity of the stereoisomeric
benzyl ethers 31 showed considerable racemization (Table SI-
3). Therefore, the phenylbutylamines 31 were purified by chiral
HPLC (Chiralpak AD) to obtain the stereoisomeric phenyl-
butyl substituted 3-benzazepines (R,S)-31, (S,S)-31, (S,R)-31,
and (R,R)-31 with high enantiomeric excess (98.4−100% ee)
(Table SI-3, Figure SI-2A−D). The configuration of the
stereoisomers 31 was derived from the configuration of the
analogous stereoisomeric methyl ethers 21. However, an X-ray
crystal structure analysis of the mixture of (S,R)-28/(R,S)-2824
confirmed unequivocally the trans-configuration and thus the
relative and absolute configuration of all four stereoisomers of
28 and 31 (Figure SI-6).
Finally the benzyl protective group of the stereoisomeric
benzyl ethers 31 was removed hydrogenolytically with H2 and
Pd/C. The enantiomers (S,S)-31 and (R,R)-31 were thus
converted into the phenols (S,S)-32, and (R,R)-32, respectively.
The enantiomerically pure phenols (R,S)-32 and (S,R)-32 were
obtained by hydrogenolytic cleavage of the racemic mixture
(S,R)-31/(R,S)-31 and subsequent separation of the enan-
tiomers by chiral preparative HPLC (Figure SI-3).
Receptor Selectivity. In addition to the affinity toward the
ifenprodil binding site of the NMDA receptor, the affinities
toward the phencyclidine binding site of the NMDA
receptor27,28 and the affinity toward σ1 and σ2 receptors29−31
were recorded in receptor binding studies (Table 1). In general
E
DOI: 10.1021/acs.jmedchem.5b00897
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Table 1. Affinities of 2-Methyl-3-benzazepin-1-ols to the Ifenprodil Binding Site of GluN2B Containing NMDA Receptors, the
a
PCP Binding Site of the NMDA Receptor and to σ1 and σ2 Receptors
Ki SEM (nM)
compd
R
configuration
GluN2B
PCP
35%
σ1
σ2
ΔH (kcal/mol) (MM/GBSA)
517
H
rac
14 1.5
5.4 0.4
187 72
67 16
41 0.7
47 9.4
224 22
28 3.1
14 3.5
38 7.8
9.3 0.4
31 13
17 5.7
26 11
180 52
10 0.7
13 2.5
194
182
293
>10 μM
554
616
CH3
Bn
rac
22%
39%
47%
28%
29%
33%
717
rac
>1 μM
25 2.7
19 2.4
48 9.0
6.1 0.9
826
b
b
b
b
(S,S)-21
(R,R)-21
(S,R)-21
(R,S)-21
(S,S)-31
(R,R)-31
(S,R)-31
(R,S)-31
(S,S)-32
(R,R)-32
(S,R)-32
(R,S)-32
1
CH3
CH3
CH3
CH3
Bn
1S,2S
1R,2R
1S,2R
1R,2S
1S,2S
1R,2R
1S,2R
1R,2S
1S,2S
1R,2R
1S,2R
1R,2S
4.5 0.5
3.2 0.8
6.3 2.0
42 5.7
196
−50.65 3.80
−54.37 3.30
−49.89 3.05
−45.79 4.20
−58.10 3.70
−60.45 3.40
−53.94 2.90
−64.12 2.99
−61.12 2.80
−53.46 3.19
−54.28 2.67
−49.62 2.12
b
0%
b
b
b
Bn
17%
47%
48%
13 2.0
46 13
31 5.2
326 36
86 2.5
665 96
56 7.7
125 24
138
Bn
154
Bn
44 18
129 26
148 11
66 6.9
129 26
98.3 34
H
1440
b
H
37%
39%
b
H
b
H
9%
eliprodil
dexoxadrol
haloperidol
di-o-tolylguanidine
32 7.4
6.3 1.6
89 29
78.1 2.3
57.5 18
a
b
The right column gives the binding energy ΔH (kcal/mol) calculated using the MM/GBSA method. For low-affinity compounds only the
inhibition of the radioligand binding at a test compound concentration of 1 μM is given.
Table 2. Inhibition of NMDA Receptor Mediated Excitotoxicity by 2-Methyl-3-benzazepin-1-ols 21, 31, and 32 Using Cell Lines
a
Stably Expressing GluN2B and GluN2A Subunit Containing NMDA Receptors, Respectively
b
c
compd
R
config Ki SEM (nM) GluN2B affinity excitotox (%) (GluN2B) IC50 SEM (nM) GluN2B activity excitotox (%) (GluN2A)
517
H
rac
14 1.5
5.4 0.4
187 72
67 16
41 0.7
47 9.4
224 22
28 3.1
14 3.5
38 7.8
9.3 0.4
31 13
17 5.7
26 11
180 52
18.6 9.8
360 6.5
>90
>90
616
CH3
Bn
rac
717
rac
(S,S)-21
(R,R)-21
(S,R)-21
(R,S)-21
(S,S)-31
(R,R)-31
(S,R)-31
(R,S)-31
(S,S)-32
(R,R)-32
(S,R)-32
(R,S)-32
memantine
CH3
CH3
CH3
CH3
Bn
1S,2S
1R,2R
1S,2R
1R,2S
1S,2S
1R,2R
1S,2R
1R,2S
1S,2S
1R,2R
1S,2R
1R,2S
−1.1 3.6
86
650 41
nd
132 36
114 10
114 16
106 2.7
177 89
366 186
102 18
116 18
46 10
2
12 3.4
4.4 5.7
50 9.0
118 3.3
41 12
5850 2200
119 15
nd
Bn
nd
Bn
nd
Bn
24 6.7
7.6 4.3
1.4 2.3
4.1 2.4
1.2 1.9
51.9 19.8
5600 1500
167 33
41 14
9.0 1.9
225 34
5323 1238
H
H
81
91 9.4
87
50.0 14.3
8
H
H
9
a
b
excitotox = excitotoxicity; inhib = inhibition; nd = not determined. Cytotoxicity after addition of 10 μM test compound, (S)-glutamate, and glycine
to GluN2B producing L13-E6 cells; IC50 values were determined only when the excitotoxicity in the screening was below 40% of controls.
c
Cytotoxicity after addition of 10 μM test compound, (S)-glutamate, and glycine to GluN2A producing L12-G10 cells.
F
DOI: 10.1021/acs.jmedchem.5b00897
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 3. (A) Binding mode of (S,R)-32 (beige). GluN1b residues are colored in cyan and GluN2B in green. H-bonds are shown as black dashed
lines. (B) Binding mode of (R,R)-32 (beige). GluN1b residues are colored in cyan and GluN2B in green. H-bonds are shown as black dashed lines.
(C) Binding mode of (R,S)-31 (beige). GluN1b residues are colored in cyan and GluN2B in green. H-bonds are shown as black dashed lines. (D)
Binding mode of (S,S)-31 (beige). GluN1b residues are colored in cyan and GluN2B in green. H-bonds are shown as black dashed lines.
the affinity toward the phencyclidine binding site of all
compounds is rather low, indicating high selectivity of the 3-
benzazepines 21, 31, and 32 toward the ifenprodil binding site
over the phencyclidine binding site.
a 4-fold selectivity over the σ2 subtype. The next potent
GluN2B ligand (R,R)-31 reveals the same affinity toward the
ifenprodil binding site and the σ1 receptor but a 10-fold
preference over the σ2 receptor. The stereoisomer (S,R)-31
shows a similar binding profile at the three receptors as (R,R)-
31. The highest selectivity for GluN2B receptors over σ1 (7-
fold) and σ2 receptors (30-fold) was observed for the
stereoisomer (S,S)-31.
The phenols 32 reveal moderate σ1 and σ2 affinity leading at
least for the high affinity GluN2B ligands (S,S)-32, (R,R)-32,
and (S,R)-32 to considerable selectivity over the σ receptors.
However, some of the methyl ethers 21 represent very potent
σ1 and/or σ2 ligands as well. The σ1 affinity of (S,S)-21, (R,R)-
21, and (S,R)-21 is in the low nanomolar range indicating even
a preference for the σ1 receptor over the GluN2B receptor.
Additionally the least potent GluN2B stereoisomer (R,S)-21
(Ki(σ2) = 6.1 nM) represents a potent and selective σ2 ligand.
In the case of the benzyl ethers 31 the stereochemistry has a
strong impact on receptor selectivity. The most potent GluN2B
ligand (R,S)-31 has a 3-fold selectivity over the σ1 receptor and
FUNCTIONAL ACTIVITY
■
The functional activity of the stereoisomeric 3-benzazepines 21,
31, and 32 was determined in excitotoxicity assays. For this
purpose L(tk−) cell lines L13-E6 and L12-G10 stably
expressing GluN1a/GluN2B and GluN1a/GluN2A subunits,
respectively, were grown and stimulated by dexamethasone to
produce the corresponding NMDA receptor subunits. Thirty
G
DOI: 10.1021/acs.jmedchem.5b00897
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 4. Correlation between calculated pKi values (calc pKi) and experimentally determined pKi values (expt pKi) (21 blue points, 31 red points,
32 green points). The binding energy was calculated by MM/GBSA method, and the calc pKi values were obtained from the linear regression
equation (ΔH = −11.05 pKi + 27.317).
minutes after addition of different amounts of a test compound,
the excitotoxicity induced by (S)-glutamate and glycine was
recorded by determination of the released amount of lactate
dehydrogenase (LDH).32 Experiments using L13-E6 and L12-
G10 cell lines show the effects on GluN2B and GluN2A
containing NMDA receptors, respectively.
In Table 2 the effects of the stereoisomeric 3-benzaepines 21,
31, and 32 on the glutamate/glycine induced cytotoxicity
toward GluN2B and GluN2A producing cells are summarized
and correlated with the GluN2B affinity and stereochemistry.
Although the benzyl ethers 31 display high GluN2B affinity
(Ki = 9−38 nM), they are not able to inhibit the glutamate/
glycine induced excitotoxicity at the screening concentration of
10 μM. Only the highly affine ligand (R,S)-31 (Ki = 9.3 nM)
reduced the cytotoxicity (GluN2B cells) considerably (23% of
controls), but recording the exact dose−response curves
resulted in an IC50 value greater than 5 μM.
The stereoisomeric methyl ethers 21 reveal the lowest
GluN2B affinity (Ki = 41−224 nM) of this series of
compounds. With the exception of the (R,R)-configured
stereoisomer (R,R)-21, the cytotoxic effects of glutamate and
glycine were reduced in the screening at a concentration of 10
μM. Finally the highest cytoprotective activity was found for
the (R,S)-configured stereoisomer (R,S)-21 with an IC50 value
of 119 nM.
The phenols 32 represent the most promising group within
this series of GluN2B ligands. At a concentration of 10 μM all
four stereoisomers 32 displayed high cytoprotective effects. The
resulting IC50 values correlate nicely with the Ki values
describing binding at the GluN2B receptor. The (R,R)- and
(S,R)-configured ligands (R,R)-32 and (S,R)-32 reveal the
highest GluN2B affinity (Ki = 17 and 26 nM) and the highest
inhibition of excitotoxicity of GluN2B subunit producing cells
(IC50 = 41 and 9 nM). For the judgment of the affinity data of
(R,R)-32 and (S,R)-32 the high SEM values have to be taken
into account. The strong protection of GluN2B subunit
producing cells by (S,R)-32 (IC50 = 9.0 nM) indicates strong
allosteric NMDA receptor antagonism.
GluN2B selective NMDA receptor antagonists, which do not
inhibit GluN2A containing NMDA receptors.
MOLECULAR MODELING STUDIES
■
In order to analyze the binding mode of the tetrahydro-3-
benzazepines 21, 31, and 32 and study the effect of their
respective configuration on the binding to NMDA receptors,
the different stereoisomers were docked into the binding
pocket formed by the dimer of the N-terminal domains of
GluN1b and GluN2B subunits of the NMDA receptor. The
obtained binding modes were further assessed by molecular
dynamic (MD) studies coupled with MM/GBSA binding
energy calculations.
Our docking studies showed that all the herein reported
tetrahydro-3-benzazepine derivatives adopt a similar binding
mode as found with 1 (Figure SI-7). Generally, the phenyl
group attached at the end of the butyl linker is embedded in the
same hydrophobic region as the benzyl group of 1 and is
stabilized by a cluster of hydrophobic amino acid residues in
GluN1 (Tyr109 and Ala75) and GluN2B (Pro78, Ile111, and
Phe114). The aromatic ring of the benzazepine moiety overlaps
with the corresponding phenolic group in 1, where it undergoes
similar hydrophobic interactions with Leu135 (GluN1),
Phe176 and Pro177 (GluN2B). Similar to the piperidine-N
of 1, the protonated benzazepine-N is able to form an H-bond
with the amide-oxygen of Gln110 side chain (GluN2B),
whereas the hydroxy group establishes H-bond interaction
with the amide-NH of Gln110 (GluN2B) or with the
carboxylate group of Glu106 (GluN2B).
The phenols 32 form H-bond interaction with Glu236
(GluN2B) like 1 (Figure 3A,B), which is lost with the methyl
and benzyl ethers 21 and 31. However, the additional benzyl
moiety in 31 protrudes into an adjacent subpocket which
explains the unexpected high affinity. In the cases of (R,S)-31
(Figure 3C) and (R,R)-31 the benzyl group is surrounded by
hydrophobic residues, namely, Leu135 (GluN1b), Tyr175
(GluN2B), and Met207 (GluN2B), whereas the benzyl group
of (S,S)-31 (Figure 3D) and (S,R)-31 interacts with Ile335
(GluN1b), Met207 (GluN2B) with the possibility of under-
going a cation−π interaction with Arg115 in GluN1b.
At a test compound concentration of 10 μM the stereo-
isomeric 3-benzazepines 21, 31, and 32 could not inhibit the
glutamate/glycine induced cytotoxic effects on GluN2A subunit
containing cells. Obviously the 3-benzazepines represent
The effect of the configuration at 1- and 2-position seems to
be mainly reflected in the positioning of both substituents in
the binding pocket. The 1-hydroxy and 2-methyl substituents in
H
DOI: 10.1021/acs.jmedchem.5b00897
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
the (1S,2R)- and (1S,2S)-configured stereoisomers show an
overlap with the hydroxy and methyl groups in 1 (Figure 3A,D
and Figure SI-7). There, the hydroxy group establishes an H-
bond with the amide-NH of Gln110. In the (1R,2R)- and
(1R,2S)-configured stereoisomers, the benzazepine scaffold is
flipped by 180° and both substituents are sitting in opposite
direction to the substituents in 1, moving closer to the GluN2B
chain (Figure 3B and Figure 3C). This allows the 1-hydroxy
group to form an additional H-bond with the side chain of
Glu106 (GluN2B) in addition to the H-bond formed with the
amide-NH of Gln110. The postulated binding modes of
selected representatives for each configuration are depicted in
Figure 3 and the binding modes of the other derivatives are
shown in Figure SI-8.
All the suggested binding modes were stable during the 10 ns
MD simulation as shown by the rmsd plots (Figure SI9A−L).
They showed favorable binding energies (Table 1), calculated
by the MM/GBSA method, and the established H-bonds were
maintained throughout the MD simulation. (Figure SI-9)
A significant correlation between the estimated binding
energies (MM/GBSA) and the observed affinities for GluN2B
containing NMDA receptors was obtained (Table 1, Figure 4).
Even when the correlation is not perfect, some useful
observations can be made. Most importantly, the (R,S)-
configured phenol and methyl ether (R,S)-21 and (R,S)-32
show the least favorable binding energies among all tested
compounds, which is in line with their reduced affinities for the
GluN2B containing NMDA receptors. (R,S)-31, which
displayed the highest affinity among the benzyl ethers 31 and
exhibited higher potency than all the herein tested compounds,
showed the most favorable binding energy.
configured phenols (R,R)-32 and (S,R)-32 with high GluN2B
affinity (Ki = 17 nM and Ki = 26 nM) and strong cytoprotective
effects (IC50 = 41 nM and IC50 = 9 nM). However, the benzyl
ethers 21 did not show cytoprotective effects in the functional
assays despite their high GluN2B binding affinity (Ki = 9−38
nM). It was shown that the benzyl ethers 21 are able to open a
lipophilic subpocket close to the ifenprodil binding site
resulting in high GluN2B affinity. However, these interactions
are not able to induce the conformational changes of the
receptor protein, which are required to close the ion channel.
EXPERIMENTAL SECTION
■
Chemistry. General. Thin layer chromatography (tlc): silica gel
60 F254 plates (Merck). Flash chromatography (fc): silica gel 60, 40−
64 μm (Merck); parentheses include diameter of the column, eluent,
fraction size, Rf value. Melting point: melting point apparatus SMP 3
(Stuart Scientific), uncorrected. Optical rotation: polarimeter 341
(PerkinElmer), wavelength 589 nm, path length (l), 1 dm; +20 °C;
unit [deg mL dm−1 g−1] is omitted, concentration [mg/mL] and
1
solvent given in parentheses. H NMR (400 MHz): Unity Mercury
Plus 400 spectrometer (Varian); δ in ppm related to tetramethylsilane;
coupling constants are given with 0.5 Hz resolution. The purity of all
described compounds including the test compounds is greater than
95%. For compound 10 and the four stereoisomers of 20 an elemental
analysis was performed. The purity of all other compounds was
analyzed by HPLC.
Methyl (S)-2-{N-[2-(3-Methoxyphenyl)ethyl]-N-(4-tosyl)-
amino}propionate ((S)-13). Under N2, sulfonamide (S)-12 (7.74
g, 50 mmol) was dissolved in absolute THF (600 mL) and cooled to 0
°C. Then phenylethanol 10 (13.1 g, 50 mmol), Ph3P (40.3 g, 0.15
mol), and DIAD (29.7 mL, 0.15 mol) were added dropwise. The
mixture was stirred at 0 °C for 1 h and at rt for 16 h. Then the solution
was diluted with n-hexane (700 mL) and the precipitated Ph3PO
was removed by filtration. The solvent was evaporated in vacuum and
the residue was purified by fc (8 cm, n-hexane/ethyl acetate 8:2, 65
mL, Rf = 0.35). Colorless oil, yield (13.9 g, 71%). [α]D −53.7 (c 1.15,
CONCLUSION
■
Three sets of stereoisomeric 2-methyltetrahydro-3-benzazepin-
1-ols were prepared, evaluated in receptor binding studies and
functional activity assays, and finally docked into the ifenprodil
binding site of the cocrystallized GluN1b/GluN2B dimer. The
binding modes of the differently substituted and configured
compounds were analyzed and evaluated by MM/GBSA
method. According to this analysis, the phenols 32 form a
similar H-bond with Glu236 (GluN2B) as the phenol of 1. A
flip of 180° of the 3-benzazepine scaffold explains the similar
GluN2B affinity of the stereoisomeric phenols (S,S)-32, (R,R)-
32, and (S,R)-32 and methyl ethers of (S,S)-21, (R,R)-21, and
(S,R)-21. The reversed binding modes are responsible for the
low influence of the ligand configuration on the GluN2B
binding affinity. This result will be taken into account during
the design and development of novel GluN2B selective NMDA
antagonists.
1
CH3OH). C20H25NO5S (391.3). H NMR (CDCl3): δ (ppm) = 1.36
(d, J = 7.3 Hz, 3H, N-CH-CH3), 2.39 (s, 3H, Ph-CH3), 2.81−2.91 (m,
1H, Ph-CH2-CH2-N), 2.96−3.05 (m, 1H, Ph-CH2-CH2-N), 3.21−3.31
(m, 1H, Ph-CH2-CH2-N), 3.43−3.49 (m, 1H, Ph-CH2-CH2-N), 3.47
(s, 3H, CO2-CH3), 3.78 (s, 3H, OCH3), 4.64 (q, J = 7.3 Hz, 1H, N-
CH-CH3), 6.71−6.77 (m, 3H, 2-H phenyl, 4-H phenyl, 6-H phenyl),
7.19 (t, J = 7.7 Hz, 1H, 5-H phenyl), 7.25 (d, J = 8.6 Hz, 2H, 3-H tosyl
and 5-H tosyl), 7.69 (d, J = 8.4 Hz, 2H, 2-H tosyl and 6-H tosyl).
(S)-2-{N-[2-(3-Methoxyphenyl)ethyl]-N-(4-tosyl)amino}-
propionic Acid ((S)-14). (S)-13 (1.54 g, 3.95 mmol) was dissolved in
a THF/H2O mixture (7:3, 60 mL), and LiOH·H2O (1.26 g, 19.8
mmol) was added. The mixture was stirred for 24 h at rt. Then the
mixture was neutralized with 6 M HCl and the aqueous layer was
extracted with diethyl ether (4 × 30 mL). The combined organic layers
were dried (Na2SO4), the solvent was evaporated in vacuum, and the
residue was purified by fc (2 cm, n-hexane/ethyl acetate 2:8, 30 mL, Rf
= 0.35). Colorless oil, yield 1.63 g (87%). [α]D −22.0 (c 1.06,
1
CH3OH). C19H23NO5S (377.4). H NMR (CDCl3): δ (ppm) = 1.36
The receptor binding studies showed unexpectedly high
GluN2B affinity of the benzyl ethers 31, which even exceeds
the GluN2B affinity of the phenols 32. Analysis of the docking
experiments resulted in an additional hydrophobic subpocket
formed by lipophilic residues of different amino acids of both
subunits (e.g., Leu135 (GluN1b), Ile335 (GluN1b), Tyr175
(GluN2B), Met207 (GluN2B)). This so far undetected
lipophilic subpocket of the GluN1b/GluN2B dimer will be
further exploited by modification of the lipophilic benzyl
residue of the ligands with the aim to find potent and selective
modulators of the NMDA receptor.
(d, J = 7.2 Hz, 3H, N-CH-CH3), 2.39 (s, 3H, Ph-CH3), 2.80−2.87 (m,
1H, Ph-CH2-CH2-N), 2.93−3.00 (m, 1H, Ph-CH2-CH2-N), 3.25−3.33
(m, 1H, Ph-CH2-CH2-N), 3.39−3.47 (m, 1H, Ph-CH2-CH2-N), 3.75
(s, 3H, OCH3), 4.57 (q, J = 7.1 Hz, 1H, N-CH-CH3), 6.67−6.68 (m,
1H, 2-H phenyl), 6.71−6.74 (m, 2H, 4-H, 6-H phenyl), 7.16 (t, J = 7.9
Hz, 1H, 5-H phenyl), 7.23 (d, J = 8.0 Hz, 2H, 3-H tosyl and 5-H
tosyl), 7.69 (d, J = 8.3 Hz, 2H, 2-H tosyl and 6-H tosyl). A signal for
the CO2H proton is not visible in the spectrum.
(S)-7-Methoxy-2-methyl-3-(p-tosyl)-2,3,4,5-tetrahydro-3-
benzazepin-1-one ((S)-15), (S)-9-Methoxy-2-methyl-3-(p-tosyl)-
2,3,4,5-tetrahydro-3-benzazepin-1-one ((S)-16), 6-Methoxy-1-
methyl-2-(p-tosyl)-1,2,3,4-tetrahydroisoquinoline (17a), and 8-
Methoxy-1-methyl-2-(p-tosyl)-1,2,3,4-tetrahydroisoquinoline
(17b). Under N2 the acid (S)-14 (5.10 g, 14.0 mmol) was dissolved in
CH2Cl2 (300 mL) and the solution was cooled to −30 °C. Then
In functional assays the cytoprotective effects of the phenols
32 correlate nicely with their GluN2B affinity. The most
promising ligands of this series are the (R,R)- and (S,R)-
I
DOI: 10.1021/acs.jmedchem.5b00897
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
trifluoroacetic acid anhydride (9.56 mL, 68.0 mmol) was added and
the mixture was stirred for 1 h at −30 °C. Afterward, SnCl4 (6.55 mL,
56.0 mmol) was added slowly. After 22 h of stirring at −30 °C, the
solution was neutralized with 5 M NaOH at 0 °C and the aqueous
layer was extracted with CH2Cl2 (3 × 50 mL). The organic layer was
dried (Na2SO4), concentrated in vacuum, and the residue was purified
by fc (8 cm, n-hexane/ethyl acetate 7:3, 65 mL, Rf (17) = 0.68) and Rf
((S)-15 and (S)-16) = 0.36).
3H, 2-H phenyl, 4-H phenyl and 6-H phenyl), 7.26−7.29 (m, 2H, 3-H
phenyl and 5-H phenyl). A signal for the OH proton is not visible.
Methyl (S)-2-{N-[2-(3-Benzyloxyphenyl)ethyl]-N-(4-tosyl)-
amino}propionate ((S)-23b). Under N2 a solution of phenylethanol
22b (5.77 g, 25.4 mmol) in THF (600 mL) was cooled to 0 °C. Then
sulfonamide (S)-12 (7.36 g, 28.6 mmol), Ph3P (20.2 g, 76.1 mmol),
and DIAD (14.7 mL, 76.1 mmol) were added and the mixture was
stirred at 0 °C for 1 h and at rt for 16 h. The solution was diluted with
n-hexane (700 mL) and the precipitate Ph3PO was removed by
filtration. The solvent was evaporated in vacuum and the residue was
purified by fc (8 cm, n-hexane/ethyl acetate 8:2, 65 mL, Rf = 0.35).
Colorless oil, yield 13.9 g (71%). [α]D −23.4 (c 1.01, CH3OH).
17a and 17b (Rf = 0.68): Pale yellow solid, mp 96 °C, yield 0.79 g
(17%).
(S)-15 and (S)-16 (Rf = 0.36): Colorless solid, yield 2.94 g (59%).
Spectroscopic data of the mixtures (S)-15/(S)-16: C19H21NO4S
1
1
C26H29NO5S (467.3). H NMR (CDCl3): δ (ppm) = 1.36 (d, J = 7.3
(359.2). H NMR (CD3OD): δ (ppm) = 1.44 (d, J = 7.5 Hz, 3 ×
Hz, 3H, N-CH-CH3), 2.40 (s, 3H, Ph-CH3), 2.84−2.91 (m, 1H, Ph-
CH2-CH2-N), 2.99−3.06 (m, 1H, Ph-CH2-CH2-N), 3.22−3.30 (m,
1H, Ph-CH2-CH2-N), 3.41−3.46 (m, 1H, Ph-CH2-CH2-N), 3.48 (s,
3H, CO2CH3), 4.65 (q, J = 7.3 Hz, 1H, N-CH-CH3), 5.05 (s, 2H, O-
CH2-Ph), 6.78−6.84 (m, 3H, 2-H phenyl, 4-H phenyl and 6-H
phenyl), 7.20 (t, J = 7.5 Hz, 1H, 5-H phenyl), 7.26−7.45 (m, 7H, CH
aromatic), 7.70 (d, J = 8.3 Hz, 2H, 2-H tosyl and 6-H tosyl).
0.5H, N-CH-CH3), 1.46 (d, J = 7.2 Hz, 3 × 0.5H, N-CH-CH3), 2.34
(s, 3 × 0.5H, Ph-CH3), 2.35 (s, 3 × 0.5H, Ph-CH3), 2.72−2.84 (m, 3
× 0.5H, 5-H), 2.91−3.02 (m, 0.5H, 5-H), 3.45−3.52 (m, 2 × 0.5H, 2
× 4-H), 3.66 (s, 3 × 0.5H, OCH3), 3.81 (s, 3 × 0.5H, OCH3), 3.84−
3.92 (m, 0.5H, 4-H), 3.94−4.05 (m, 0.5H, 4-H), 4.71 (q, J = 7.1 Hz,
0.5H, 2-H), 4.77 (q, J = 7.8 Hz, 0.5H, 2-H), 6.54 (d, J = 2.5 Hz, 0.5H,
6-H●), 6.64 (d, J = 7.4 Hz, 0.5H, 6-H□), 6.69 (dd, J = 8.5/2.5 Hz,
0.5H, 8-H●), 6.70 (d, J = 8.4 Hz, 0.5H, 8-H□), 7.04 (br d, J = 8.0 Hz,
2 × 0.5H, 3-H tosyl and 5-H tosyl), 7.06 (d, J = 8.0 Hz, 2 × 0.5H, 3-H
tosyl and 5-H tosyl), 7.31 (d, J = 8.5 Hz, 0.5H, 9-H●), 7.21−7.25 (m,
0.5H, 7-H□), 7.30 (br d, J = 8.3 Hz, 2 × 0.5H, 2-H tosyl and 6-H
tosyl).7.35 (d, J = 8.3 Hz, 2 × 0.5H, 2-H tosyl and 6-H tosyl). The
ratio of regioisomers 15 (marked with ●)/16 (marked with □) is
50:50.
(S)-7-Benzyloxy-2-methyl-3-(4-tosyl)-2,3,4,5-tetrahydro-3-
benzazepin-1-one ((S)-25b) and (S)-9-Benzyloxy-2-methyl-3-
(4-tosyl)-2,3,4,5-tetrahydro-3-benzazepin-1-one ((S)-26b) and
6-Benzyloxy-1-methyl-2-(4-tosyl)-1,2,3,4-tetrahydroisoquino-
line (27b). Under N2 a solution of the acid (S)-24b (4.77 g, 10.5
mmol) in CH2Cl2 (295 mL) was cooled to −15 °C. Then
trifluoroacetic acid anhydride (5.56 mL, 39.4 mmol) was added and
the mixture was stirred for 30 min at −15 °C. After addition of SnCl4
(4.61 mL, 39.4 mmol), the mixture was stirred at −15 °C for 24 h.
Water (150 mL) was added slowly, and the temperature was increased
to rt. The aqueous layer was neutralized with 2 M NaOH, saturated
with NaCl and extracted with CH2Cl2 (3 × 75 mL). The combined
organic layers were dried (Na2SO4), concentrated in vacuum, and the
residue was dissolved in acetone (180 mL). K2CO3 (5.44 g, 39.4
mmol) and benzyl bromide (1.56 mL, 13.1 mmol) were added, and
the suspension was heated to reflux for 6 h. The mixture was filtered,
concentrated in vacuum, and the residue was purified by fc (8 cm, n-
hexane/ethyl acetate 7:3, 65 mL, Rf (27b) = 0.66 and Rf ((S)-25b and
(S)-26b) = 0.34). (S)-25b and (S)-26b (Rf = 0.34): Colorless solid,
yield 1.88 g (41%). 27b (Rf = 0.66): Pale yellow resin, yield 855.8 mg
(20%). C25H25NO4S (435.6). 1H NMR (CDCl3): δ (ppm) = 1.43 (d, J
= 7.5 Hz, 3 × 0.95H, N-CH-CH3●), 1.59 (d, J = 7.6 Hz, 3 × 0.05H, N-
CH-CH3□), 2.30 (s, 3 × 0.95H, Ph-CH3●), 2.36 (s, 3 × 0.05H, Ph-
CH3□), 2.79−2.89 (m, 1H, 4-H), 2.95−3.07 (m, 1H, 5-H), 3.48 (ddd,
J = 13.0/6.7/1.8 Hz, 0.95H, 5-H●), 3.61 (ddd, J = 13.3/6.7/1.9 Hz,
0.05H, 5-H□), 3.99−4.08 (m, 1H, 4-H), 4.76 (q, J = 7.5 Hz, 0.95H, 2-
H●), 4.92 (q, J = 7.6 Hz, 0.05H, 2-H□), 5.06 (s, 2 × 0.95H, O-CH2-
Ph●), 5.08 (s, 2 × 0.05H, O-CH2-Ph□), 6.60 (d, J = 7.8 Hz, 0.05H, 6-
H□), 6.64 (d, J = 2.4 Hz, 0.95H, 6-H●), 6.71 (d, J = 8.5 Hz, 0.05H, 8-
H□), 6.76 (dd, J = 8.5/2.5 Hz, 0.95H, 8-H●), 7.00 (d, J = 8.7 Hz, 2 ×
0.05H, 3-H tosyl and 5-H tosyl□), 7.02 (d, J = 7.9 Hz, 2 × 0.95H, 3-H
tosyl● and 5-H tosyl●), 7.13−7.24 (m, 4 × 0.05H, C-H aromatic□),
7.28 (d, J = 8.5 Hz, 0.95H, 9-H●), 7.32−41 (m, 6.85H, C-H aromatic).
The ratio of regioisomers 25b/26b is 95 (●) and 5 (□), respectively.
(1S,2S)-7-Benzyloxy-2-methyl-3-(4-tosyl)-2,3,4,5-tetrahydro-
1H-3-benzazepin-1-ol ((S,S)-28) and (1R,2S)-7-Benzyloxy-2-
methyl-3-(4-tosyl)-2,3,4,5-tetrahydro-1H-3-benzazepin-1-ol
((R,S)-28). Under N2, NaBH4 (0.55 g, 14.3 mmol) was added to a 95:5
mixture of (S)-25b and (S)-26b (3.11 g, 7.16 mmol) in CH3OH (75
mL). After stirring for 20 h at rt, H2O (30 mL) was added. The
aqueous layer was extracted with CH2Cl2 (3 × 20 mL). The combined
organic layers were dried (Na2SO4), concentrated in vacuum, and the
residue was purified by fc (8 cm, n-hexane/ethyl acetate7:3, 65 mL, Rf
((S,S)-28) = 0.44, Rf ((R,S)-28) = 0.29). The potential product 29 was
not isolated. (S,S)-28 (Rf = 0.44): Colorless solid, mp 144 °C, yield
837.1 mg (27%). [α]D −0.21 (c 0.80, CHCl3, 8.3% ee). (R,S)-28 (Rf =
0.29): Colorless solid, mp 151 °C, yield 1.18 g (38%). [α]D +13.3 (c
1.01, CHCl3, 2.1% ee).
(1R,2S)-7-Methoxy-2-methyl-2,3,4,5-tetrahydro-1H-3-benza-
zepin-1-ol ((R,S)-20). Sulfonamide (R,S)-18 (0.72 g, 2.17 mmol) was
dissolved in CH3OH (85 mL), and Mg (1.13 g, 47.2 mmol) was
added. The reaction mixture was heated to reflux for 18 h. 2 M HCl
was added, residual Mg was filtered off, the pH value was adjusted to
pH 9−10 by addition of 2 M NaOH, and the aqueous layer was
extracted with CH2Cl2 (5 × 40 mL). The organic layer was dried
(Na2SO4), concentrated in vacuum, and the residue was purified by fc
(4 cm, ethyl acetate/CH3OH 95:5 + 2% N,N-dimethylethanamine, 30
mL, Rf = 0.07). Pale yellow oil, yield 254 mg (57%). [α]D +0.13 (c
1
0.99, CH3OH, 34% ee). C12H17NO2 (207.3). H NMR (CDCl3): δ
(ppm) = 0.93 (d, J = 6.9 Hz, 3H, N-CH-CH3), 2.42 (br s, 1H, NH),
2.54 (ddd, J = 14.9/5.4/1.6 Hz, 1H, 5-H), 2.91 (ddd, J = 13.4/5.4/2.9
Hz, 1H, 4-H), 3.00−3.06 (m, 1H, 5-H), 3.27−3.36 (m, 2H, 2-H and 4-
H), 3.76 (s, 3H, OCH3), 4.36 (d, J = 5.4 Hz, 1H, 1-H), 6.61 (br s, 1H,
6-H), 6.62 (dd, J = 8.8/2.7 Hz, 1H, 8-H), 7.04 (d, J = 8.8 Hz, 1H, 9-
H). A signal for the OH proton is not visible.
(1R,2S)-7-Methoxy-2-methyl-3-(4-phenylbutyl)-2,3,4,5-tetra-
hydro-1H-3-benzazepin-1-ol ((R,S)-21) and (1S,2R)-7-Methoxy-
2-methyl-3-(4-phenylbutyl)-2,3,4,5-tetrahydro-1H-3-benzaze-
pin-1-ol ((S,R)-21). A mixture of the secondary amine (R,S)-20 (70.8
mg, 0.34 mmol), 1-chloro-4-phenylbutane (111.4 μL, 0.68 mmol),
TBAI (251.2 mg, 0.68 mmol), K2CO3 (188.0 mg, 1.36 mmol), and
CH3CN (10 mL) was heated to reflux for 72 h. The mixture was
filtered, and the filtrate was concentrated in vacuum. The residue was
purified by fc (2 cm, n-hexane/ethyl acetate 95:5 and 1% N,N-
dimethylethanamine, 10 mL, Rf = 0.10) to afford a mixture of the
enantiomers (R,S)-21 and (S,R)-21, in the ratio of 67:33. Colorless
solid, mp 113 °C, yield 55.6 mg (48%). Purity (HPLC): 98.1%, tR =
18.7 min. The enantiomers (R,S)-21 and (S,R)-21 were separated on
an analytical chiral HPLC (column, Chiralpak AD (Chiral
Technologies Europe), solvent n-hexane/isopropanol (9:1), tR
((R,S)-21) = 11.5 min, tR ((S,R)-21) = 14.4 min). (R,S)-21: Colorless
oil. [α]D +31.8 (c 0.53, CH3OH, 99.8% ee). (S,R)-21: Colorless resin.
[α]D −32.4 (c 0.52, CH3OH, 99.4% ee). Spectroscopic data of (R,S)-
21 and (S,R)-21: C22H29NO2 (339.5). 1H NMR (CDCl3): δ (ppm) =
0.64 (d, J = 6.8 Hz, 3H, N-CH-CH3), 1.52−1.62 (m, 2H, N-CH2-CH2-
CH2-CH2-Ph), 1.64−1.73 (m, 2H, N-CH2-CH2-CH2-CH2-Ph), 2.49
(ddd, J = 10.9/4.9/1.5 Hz, 1H, 5-H), 2.52 (t, J = 7.1 Hz, 2H, N-CH2-
CH2-CH2-CH2-Ph), 2.61−2.71 (m, 4H, N-CH2-CH2-CH2-CH2-Ph
and 4-H), 3.23−3.30 (m, 2H, 2-H and 5-H), 3.76 (s, 3H, OCH3), 4.31
(d, J = 5.9 Hz, 1H, 1-H), 6.60 (d, J = 2.5 Hz, 1H, 6-H), 6.62 (dd, J =
8.0/2.7 Hz, 1H, 8-H), 6.99 (d, J = 8.0 Hz, 1H, 9-H), 7.15−7.19 (m,
(S,S)-28: C25H27NO4S (437.3). 1H NMR (CDCl3): δ (ppm) = 0.57
(d, J = 6.8 Hz, 3H, N-CH-CH3), 1.64 (br s, 1H, OH), 2.38 (s, 3H, Ph-
CH3), 2.68 (dd, J = 14.5/5.5 Hz, 1H, 5-H), 2.80−2.96 (m, 2H, 4-H
J
DOI: 10.1021/acs.jmedchem.5b00897
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
and 5-H), 4.02−4.08 (m, 1H, 4-H), 4.35 (ddd, J = 6.9/2.4/0.9 Hz, 1H,
2-H), 5.00 (s, 2H, O-CH2-Ph), 5.00−5.05 (m, 1H, 1-H), 6.68 (d, J =
2.6 Hz, 1H, 6-H), 6.82 (dd, J = 8.6/2.6 Hz, 1H, 8-H), 7.25 (d, J = 7.9
Hz, 2H, 3-H and 5-H tosyl), 7.27−7.31 (m, 1H, 4-H benzyl), 7.33−
7.40 (m, 5H, 2-H, 3-H, 5-H, 6-H benzyl and 9-H), 7.68 (d, J = 8.3 Hz,
2H, 2-H and 6-H tosyl).
(1S,2S)-2-Methyl-3-(4-phenylbutyl)-2,3,4,5-tetrahydro-1H-
benzazepine-1,7-diol ((S,S)-32). Pd/C (10%, 50.0 mg) was added
to a solution of the enantiomerically pure benzyl ether (S,S)-31 (60.1
mg, 0.14 mmol) in absolute CH3OH (6 mL), and the suspension was
stirred under H2 atmosphere (1 bar) for 1 h at rt. The catalyst was
removed by filtration over Celite 535, the solvent was evaporated in
vacuum, and the residue was purified by fc (2 cm, n-hexane/ethyl
acetate 5:5 and 1% N,N-dimethylethanamine, 10 mL, Rf = 0.13).
Colorless solid, mp 89 °C, yield 44.9 mg (95%). [α]D −25.0 (c 0.5,
CH3OH, 98.8% ee). C21H27NO2 (325.5). 1H NMR (CDCl3): δ (ppm)
= 0.82 (d, J = 6.3 Hz, 3H, N-CH-CH3), 1.49−1.68 (m, 4H, N-CH2-
CH2-CH2-CH2-Ph), 2.58−2.77 (m, 7H, N-CH2-CH2-CH2-CH2-Ph, 4-
H and 1 × 5-H), 2.82−2.89 (m, 1H, 5-H), 3.01 (q, J = 6.3 Hz, 1H, 2-
H), 4.88 (br s, 1H, 1-H), 6.54 (d, J = 2.4 Hz, 1H, 6-H), 6.63 (dd, J =
8.2/2.5 Hz, 1H, 8-H), 7.16−7.19 (m, 4H, 9-H, 2-H phenyl, 4-H
phenyl and 6-H phenyl), 7.26−7.29 (m, 2H, 3-H phenyl and 5-H
phenyl). A signal for the OH proton is not visible.
Receptor Binding Studies. Cell Culture and Preparation of
Membrane Homogenates for the GluN2B Assay. The assay for
the GluN2B affinity has been reported in refs 25 and 26. Briefly, in the
assay mouse L(tk−) cells stably transfected with the dexamethasone
inducible eukaryotic expression vectors pMSG NR1a, pMSG NR2B in
a 1:5 ratio were used. The transformed L(tk−) cells were grown in
modified Earle’s medium (MEM) containing 10% of standardized fetal
calf serum (FCS, Biochrom AG, Berlin, Germany). The expression of
the NMDA receptor at the cell surface was induced after the cell
density of the adherent growing cells had reached approximately 90%
of confluency. For the induction, the original growth medium was
replaced by growth medium containing 4 μM dexamethasone and 4
μM ketamine (final concentration). After 24 h the cells were harvested
by scraping and pelleted (10 min, 5000g, Hettich Rotina 35R
centrifuge, Tuttlingen, Germany).
For the binding assay, the cell pellet was resuspended in phosphate
buffer saline (PBS) buffer and the number of cells was determined
using an improved Neubauer’s counting chamber (VWR, Darmstadt,
Germany). Subsequently, the cells were lysed by sonication (4 °C, 6 ×
10 s cycles with breaks of 10 s; device, Soniprep 150, MSE, London,
U.K.). The resulting cell fragments were centrifuged with a high
performance cool centrifuge (20 000g, 4 °C, Sorvall RC-5 plus,
Thermo Scientific). The supernatant was discarded and the pellet
resuspended in a defined volume of PBS yielding cell fragments of
approximately 500 000 cells/mL. The suspension of membrane
homogenates was sonicated again (4 °C, 2 × 10 s cycles with a
break of 10 min) and stored at −80 °C.
Performing the GluN2B Binding Assay. The competitive
binding assay was performed with the radioligand [3H]1 (60 Ci/
mmol, PerkinElmer) using standard 96-well-multiplates (Diagonal,
Muenster, Germany). The thawed cell membrane preparation (about
20 μg of protein) was incubated with six different concentrations of
test compounds, 5 nM [3H]1, and Tris/EDTA buffer (5 mM/1 mM,
pH 7.5) in a total volume of 200 μL for 120 min at 37 °C. The
incubation was terminated by rapid filtration through the presoaked
filtermats by using the cell harvester. After washing each well five times
with 300 μL of water, the filtermats were dried at 95 °C. Subsequently,
the solid scintillator was placed on the filtermat and melted at 95 °C.
After 5 min, the solid scintillator was allowed to solidify at rt. The
bound radioactivity trapped on the filters was counted in the
scintillation analyzer. The nonspecific binding was determined with
10 μM unlabeled 1. The Kd value of 1 is 7.6 nM.
(R,S)-28: C25H27NO4S (437.3). 1H NMR (CDCl3): δ (ppm) = 0.69
(d, J = 6.7 Hz, 3H, N-CH-CH3), 1.59 (br s, 1H, OH), 2.39 (s, 3H, Ph-
CH3), 2.57−2.62 (m, 1H, 5-H), 3.01 (ddd, J = 14.3/5.7/1.1 Hz, 1H, 4-
H), 3.43−3.50 (m, 1H, 5-H), 4.06 (ddd, J = 14.5/5.7/2.4 Hz, 1H, 4-
H), 4.44−4.51 (m, 2H, 1-H and 2-H), 5.00 (s, 2H, O-CH2-Ph), 6.68−
6.71 (m, 2H, 6-H and 8-H), 7.01 (d, J = 8.7 Hz, 1H, 9-H), 7.25 (d, J =
7.9 Hz, 2H, 3-H tosyl and 5-H tosyl), 7.27−7.32 (m, 1H, 4-H benzyl),
7.33−7.40 (m, 4H, 2-H benzyl, 3-H benzyl, 5-H benzyl and 6-H
benzyl), 7.68 (d, J = 8.3 Hz, 2H, 2-H tosyl and 6-H tosyl). 13C NMR
(CDCl3): δ (ppm) = 13.7 (1C, N-CH-CH3), 21.7 (1C, Ph-CH3), 37.3
(1C, C-5), 41.4 (1C, C-4), 54.4 (1C, C-2), 70.1 (1C, O-CH2-Ph), 78.5
(1C, C-1), 111.7 (1C, C-8), 118.0 (1C, C-6), 127.1 (2C, C-2 tosyl and
C-6 tosyl), 127.7 (2C, C-2 benzyl and C-6 benzyl), 128.2 (1C, C-4
benzyl), 128.8 (2C, C-3 benzyl and C-5 benzyl), 129.9 (2C, C-3 tosyl
and C-5 tosyl), 130.8 (1C, C-9a), 132.4 (1C, C-9), 137.0 (1C, C-1
benzyl), 138.1 (1C, C-1 tosyl), 141.2 (1C, C-5a), 143.4 (1C, C-4
tosyl), 158.8 (1C, C-7).
(1S,2S)-7-Benzyloxy-2-methyl-2,3,4,5-tetrahydro-1H-3-ben-
zazepin-1-ol ((S,S)-30). Mg (1.37 g, 56.9 mmol) was added to a
solution of (S,S)-28 (1.13 g, 2.58 mmol) in CH3OH (130 mL). The
reaction mixture was heated under reflux for 36 h. 2 M HCl (pH 3)
was added, and the mixture was filtered. The solution was adjusted
with 2 M NaOH to pH 10, the water layer was extracted with CH2Cl2
(5 × 50 mL), the organic layer was dried (Na2SO4), concentrated in
vacuum, and the residue was purified by fc (6 cm, ethyl acetate/
CH3OH 95:5 + 2% N,N-dimethylethanamine, 30 mL, Rf = 0.23).
Colorless solid, mp 175 °C, yield 438.1 mg (60%). [α]D −0.90 (c 0.77,
1
CH3OH, 8.3% ee). C18H21NO2 (283.4). H NMR (CDCl3): δ (ppm)
= 0.96 (d, J = 6.9 Hz, 3H, N-CH-CH3), 2.55 (ddd, J = 14.7/5.3/1.3
Hz, 1H, 5-H), 2.90 (ddd, J = 13.4/5.4/2.8 Hz, 1H, 4-H), 3.03−3.09
(m, 1H, 4-H), 3.27−3.34 (m, 2H, 2-H and 5-H), 4.37 (d, J = 5.4 Hz,
1H, 1-H), 5.04 (s, 2H, O-CH2-Ph), 6.70−6.73 (m, 2H, 6-H and 8-H),
7.06 (d, J = 7.8 Hz, 1H, 9-H), 7.30−7.35 (m, 1H, 4-H benzyl), 7.37−
7.44 (m, 4H, 2-H benzyl, 3-H benzyl, 5-H benzyl and 6-H benzyl).
Signals for the OH-and NH-protons are not visible.
(1S,2S)-7-Benzyloxy-2-methyl-3-(4-phenylbutyl)-2,3,4,5-tet-
rahydro-1H-3-benzazepin-1-ol ((S,S)-31) and (1R,2R)-7-Benzyl-
oxy-2-methyl-3-(4-phenylbutyl)-2,3,4,5-tetrahydro-1H-3-ben-
zazepin-1-ol ((R,R)-31). A mixture of the secondary amine (S,S)-30
(122.2 mg, 0.43 mmol), 1-chloro-4-phenylbutane (108.0 μL, 0.66
mmol), Bu4NI (241.0 mg, 0.65 mmol), K2CO3 (301.0 mg, 2.18
mmol), and CH3CN (20 mL) was heated to reflux for 72 h. The
mixture was filtered, concentrated in vacuum, and the residue was
purified by fc (3 cm, n-hexane/ethyl acetate 7:3 and 1% N,N-
dimethylethanamine, 10 mL, Rf = 0.22). Colorless solid, mp 101 °C,
yield 154.1 mg (86%), mixture of enantiomers (S,S)-31 and (R,R)-31
in the ratio 54:46. The enantiomer (R,R)-30 was transformed in the
same manner. The enantiomers (S,S)-31 and (R,R)-31 were separated
on an analytical chiral HPLC (column Chiralpak IB, Chiral
Technologies Europe, solvent n-hexane/ethanol/isopropanol 96:3:1,
tR ((S,S)-31) = 48.6 min, tR ((R,R)-31) = 58.6 min).
(S,S)-31 (tR = 48.6 min): Colorless resin, [α]D +1.33 (c 1.33,
CH2Cl2, 98.8% ee).
(R,R)-31 (tR = 58.6 min): Colorless resin, [α]D −2.00 (c 0.40,
CH2Cl2, 98.4% ee).
Affinity toward the PCP Binding Site of the NMDA Receptor.
The affinity toward the PCP binding site of the NMDA receptor was
determined in receptor binding studies as described in refs 27 and 28
using [3H](+)-MK-801 as radioligand.
Affinity toward σ1 and σ2 Receptors. The affinity toward σ1 and
σ2 receptors was recorded in receptor binding studies as described in
refs 29−31 using [3H](+)-pentazocine and [3H]di-o-tolylguanidine as
radioligands, respectively.
Spectroscopic data of (S,S)-31 and (R,R)-31: C28H33NO2 (415.6).
1H NMR (CDCl3): δ (ppm) = 0.83 (d, J = 6.5 Hz, 3H, N-CH-CH3),
1.49−1.59 (m, 2H, N-CH2-CH2-CH2-CH2-Ph), 1.61−1.69 (m, 2H, N-
CH2-CH2-CH2-CH2-Ph), 2.63−2.91 (m, 8H, N-CH2-CH2-CH2-CH2-
Ph, 4-H and 5-H), 3.00 (q, J = 6.7 Hz, 1H, 2-H), 4.88 (br s, 1H, 1-H),
5.04 (s, 2H, O-CH2-Ph), 6.72 (d, J = 2.6 Hz, 1H, 6-H), 6.79 (dd, J =
8.3/2.6 Hz, 1H, 8-H), 7.17−7.20 (m, 3H, C-H aromatic), 7.25−7.44
(m, 8H, C-H aromatic). A signal for the OH proton is not visible.
Functional Activity. Glutamate induced cytotoxicity was
determined as described recently.32 Briefly, L(tk−) cells stably
expressing either NR1-1a/NR2A (L12-G10 cells) or NR1-1a/NR2B
K
DOI: 10.1021/acs.jmedchem.5b00897
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
(L13-E6 cells) were preincubated with compounds (concentration
range 1 nM to 10 μM, dissolved in DMSO) for 30 min. Then a
mixture of (S)-glutamate and glycine (10 μM each) was added and the
cells were further incubated for 4 h. Excitotoxicity was determined by
detection of LDH release from the cells into the culture supernatants.
In this assay 0% excitotoxicity is defined as LDH release in the
presence of (S)-glutamate/glycine (10 μM each) and 100 μM
ketamine, whereas 100% excitotoxicity is defined as LDH release
after addition of (S)-glutamate/glycine in the absence of ketamine.
Experiments were repeated at least three times with six replicates per
test concentration. IC50 values were calculated using GraphPad Prism,
version 5.0.
Computational Methods. General. The crystal structure of
NMDA-GluN1b/GluN2B dimer in complex with 1 (3QEL)6 was
taken from the Protein Data Bank. All water molecules were removed
except one water molecule (HOH415) located in the binding pocket.
The crystal structure was then protonated using MOE modeling
package (Chemical Computing Group)33 and subsequently energy-
minimized by applying Amber99 force field and the GB/SA
continuum solvent model. A tethering constant of 100 kcal/(mol Å)
was applied on the backbone during the minimization process.
Docking. MOE modeling software was used to generate the
molecular structures of all compounds at the physiological pH. Initial
ligand conformations were obtained by energy minimization using the
MMFF94x force field implemented in MOE. The prepared structures
were docked into ifenprodil binding pocket using the GOLD docking
program (version 5.1);34 a binding pocket of 6 Å around the
cocrystallized ligand 1 was assigned, and the water molecule
(HOH415) was kept in the binding pocket and set to toggle.
Goldscore was chosen as the fitness function, and a maximum number
of 10 GA runs for each compound was allowed. By use of this setup,
the cocrystallized inhibitors could be correctly docked (rmsd below 1
Å) into the NMDA-GluN1b/GluN2B binding pocket.
trajectories were used for energy calculations, GBOBC model was
selected, and the entropy contributions were not included.
ASSOCIATED CONTENT
* Supporting Information
■
S
Tables showing the optimization of the intramolecular Friedel−
Crafts acylation, enantiomeric purity data, H-bond analyses,
figures displaying chiral HPLC, docking of ifenprodil and all
test compounds, rmsd plots of MD simulations, synthetic
methods, physical, spectroscopic and purity data of all
compounds, details of receptor binding studies, and a csv file
containing molecular formula strings. The Supporting
Information is available free of charge on the ACS Publications
website at DOI: 10.1021/acs.jmedchem.5b00897.
AUTHOR INFORMATION
Corresponding Author
*Phone: +49-251-8333311. Fax: +49-251-8332144. E-mail:
wuensch@uni-muenster.de.
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We are grateful to Prof. Dr. D. Steinhilber, Department of
Pharmacy, University of Frankfurt, for donating us the L(tk−)
cells stably expressing GluN1a/GluN2A and GluN1a/GluN2B
receptor proteins. F.G. and T.W. acknowledge the expert
technical assistance by Heidemarie Graf in performing the
excitotoxicity assays. This work was funded by the Deutsche
Forschungsgemeinschaft (Grant GRK 1143), which is gratefully
acknowledged.
Molecular Dynamics. The sander module in AMBER 1235 was
applied for the energy minimization steps while pmemd was utilized
for the MD simulation. Particle-mesh-Ewald (PME) method was used
to treat electrostatic interactions, and a cutoff radius of 8 Å was set for
nonbonded interactions. Preparation of the input files was
accomplished as follows: The general Amber force field (GAFF)
with AM1-BCC charges and Amber03 force field, as implemented in
AMBER 12, were applied to assign atom types to the ligand and the
protein, respectively. The inhibitor−protein complexes were solvated
in an octahedral box of TIP3P water with 13 Å between the complex
surface and the box boundary, and counterions were added to
neutralize the system.
The solvated complexes were then subjected to two steps of energy
minimization: a 3000 steps minimization while keeping the protein
fixed (2000 cycles of steepest descent followed by 1000 cycles of
conjugate gradient), followed by 2000 cycles of minimization of the
whole complex structure using 1000 steps steepest descent and
subsequent 1000 steps conjugate gradient. The system was gradually
heated to 300 K over 100 ps and then a 10 ns MD simulation was
performed in the NPT ensemble (pressure of 1.0 bar and temperature
of 300 K). During MD all covalent bonds involving H atoms were
constrained using the SHAKE procedure, and the time step was set to
2 fs.
The root-mean-square deviation of the protein’s backbone atoms
and of the studied inhibitors was calculated using the first frame in the
MD simulation as reference.
The hydrogen bonds formed between the inhibitors and
neighboring protein residues in the binding pocket were analyzed
using the program ptraj from AMBER 12 after extracting 1000
snapshots from the trajectories of the complex over the 10 ns MD
simulation. The distance cutoff was set to 3.5 Å, and no angle cutoff
was applied.
MM/GBSA Binding Energy Calculation. Estimation of the
binding energy was accomplished using the molecular mechanics
generalized Born surface area (MM/GBSA) method implemented in
AMBER 12. A total of 500 frames obtained from the complex’
ABBREVIATIONS USED
■
NMDA, N-methyl-D-aspartate; AMPA, 2-amino-3-(3-hydroxy-
5-methylisoxazol-4-yl)propionic acid; PCP, (1-
phenylcyclohexyl)piperidine; DTG, di-o-tolylguanidine; mGlu,
metabotropic glutamate receptor; iGlu, ionotropic glutamate
recetptor; DIAD, diisopropyl azodicarboxylate; MD, molecular
dynamics; MM/GBSA, molecular mechanics generalized Born
surface area; rmsd, root-mean-square deviation; SEM, standard
error of the mean; MEM, modified Earle’s medium; FCS, fetal
calf serum; PBS, phosphate buffer saline; PME, particle mesh
Ewald
REFERENCES
■
(1) Brauner-Osborne, H.; Egebjerg, J.; Nielsen, E. O.; Madsen, U.;
̈
Krogsgaard-Larsen, P. Ligands for glutamate receptors: design, and
therapeutic potential. J. Med. Chem. 2000, 43, 2609−2645.
(2) Kew, J. N. C.; Kemp, J. A. Ionotropic and metabotropic glutamate
receptor structure and pharmacology. Psychopharmacology 2005, 179,
4−29.
(3) Paoletti, P.; Neyton, J. NMDA recptor subunits: function and
pharmacology. Curr. Opin. Pharmacol. 2007, 7, 39−47.
(4) Lodge, D. The history of the pharmacology and cloning of
ionotropic glutamate receptors and the development of idiosyncratic
nomenclature. Neuropharmacology 2009, 56, 6−21.
(5) Furukawa, H.; Singh, S. K.; Mancusso, R.; Gouaux, E. Subunit
arrangement and function in NMDA receptors. Nature 2005, 438,
185−192.
(6) Karakas, E.; Simorowski, N.; Furukawa, H. Subunit arragement
and phenylethanolamine binding in GluN1/GluN2B NMDA
receptors. Nature 2011, 475, 249−253.
(7) Mony, L.; Kew, J. N. C.; Gunthorpe, M. J.; Paoletti, P. Allosteric
modulators of NR2B-containing NMDA receptors: molecular
L
DOI: 10.1021/acs.jmedchem.5b00897
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
mechanisms and therapeutic potential. Br. J. Pharmacol. 2009, 157,
1301−1317.
(26) Benner, A.; Bonifazi, A.; Shirataki, C.; Temme, L.; Schepmann,
D.; Quaglia, W.; Shoji, O.; Watanabe, Y.; Daniliuc, C.; Wunsch, B.
̈
(8) Paoletti, P.; Bellone, C.; Zhou, Q. NMDA receptor subunit
diversity: impact on receptor properties, synaptic plasticity and disease.
Nat. Rev. Neurosci. 2013, 14, 383−400.
GluN2B-selective N-methyl-D-aspartate (NMDA) receptor antago-
nists derived from 3-benzazepines: synthesis and pharmacological
evaluation of benzo[7]annulen-7-amines. ChemMedChem 2014, 9,
741−751.
(9) Layton, M. E.; Kelly, M. J., III; Rodzinak, K. J. Recent advances in
the development of NR2B subtype-selective NMDA receptor
antagonists. Curr. Top. Med. Chem. 2006, 6, 697−709.
(10) Wu, L.-J.; Zhou, M. Targeting the NMDA receptor subunit
NR2B for the treatment of neuropathic pain. Neurotherapeutics 2009,
6, 693−702.
(27) Kohler, J.; Bergander, K.; Fabian, J.; Schepmann, D.; Wunsch, B.
̈
̈
Enantiomerically pure 1,3-dioxanes as highly selective NMDA and σ1
receptor ligands. J. Med. Chem. 2012, 55, 8953−8957.
(28) Banerjee, A.; Schepmann, D.; Kohler, J.; Wurthwein, E.-U.;
̈
̈
Wunsch, B. Synthesis and SAR studies of chiral non-racemic
̈
dexoxadrol analogues as uncompetitive NMDA receptor antagonists.
Bioorg. Med. Chem. 2010, 18, 7855−7867.
(11) Williams, K. Ifenprodil, a novel NMDA receptor antagonist: site
and mechanism of action. Curr. Drug Targets 2001, 2, 285−298.
(29) Meyer, C.; Neue, B.; Schepmann, D.; Yanagisawa, S.;
(12) Falck, E.; Begrow, F.; Verspohl, E.; Wunsch, B. Metabolism
̈
Yamaguchi, J.; Wurthwein, E.-U.; Itami, K.; Wunsch, B. Improvement
̈
̈
studies of ifenprodil, a potent GluN2B receptor antagonist. J. Pharm.
Biomed. Anal. 2014, 88, 96−105.
of σ1 receptor affinity by late-stage C-H-bond arylation of spirocyclic
lactones. Bioorg. Med. Chem. 2013, 21, 1844−1856.
́
(13) Borza, I.; Domany, G. NR2B selective NMDA antagonists: the
(30) Miyata, K.; Schepmann, D.; Wunsch, B. Synthesis and σ
̈
evolution of the ifenprodil-type pharmacophore. Curr. Top. Med.
Chem. 2006, 6, 687−695.
receptor affinity of regioisomeric spirocyclic furopyridines. Eur. J. Med.
Chem. 2014, 83, 709−716.
(14) Yurkewicz, L.; Weaver, J.; Bullock, M. R.; Marshall, L. F. The
effect of the selective NMDA receptor antagonist traxoprodil in the
treatment of traumatic brain injury. J. Neurotrauma 2005, 22 (12),
1428−1443.
(31) Hasebein, P.; Frehland, B.; Lehmkuhl, K.; Frohlich, R.;
̈
Schepmann, D.; Wunsch, B. Synthesis and pharmacological evaluation
̈
of like- and unlike-configured tetrahydro-2-benzazepines with α-
substituted benzyl moiety in 5-position. Org. Biomol. Chem. 2014,
12, 5407−5426.
(15) Wessell, R. H.; Ahmed, S. M.; Menniti, F. S.; Dunbar, G. L.;
Chase, T. N.; Oh, J. D. NR2B selective NMDA receptor antagonist
CP-101,606 prevents levodopa-induced motor response alterations in
hemi-parkinsonian rats. Neuropharmacology 2004, 47, 184−194.
(16) Tewes, B.; Frehland, B.; Schepmann, D.; Schmidtke, K.-U.;
(32) Rook, Y.; Schmidtke, K.-U.; Gaube, F.; Schepmann, D.;
Wunsch, B.; Heilmann, J.; Lehmann, J.; Winckler, T. Bivalent beta-
̈
carbolines as potential multitarget anti-Alzheimer agents. J. Med. Chem.
2010, 53, 3611−3617.
Winckler, T.; Wunsch, B. Design, synthesis, and biological evaluation
̈
(33) Molecular Operating Environment (MOE), version 2012.10;
Chemical Computing Group Inc. (1010 Sherbooke St. West, Suite No.
910, Montreal, QC, Canada, H3A 2R7), 2012.
of 3-benzazepin-1-ols as NR2B-selective NMDA receptor antagonists.
ChemMedChem 2010, 5, 687−695.
(17) Tewes, B.; Frehland, B.; Schepmann, D.; Schmidtke, K.-U.;
(34) Verdonk, M. L.; Cole, J. C.; Hartshorn, M. J.; Murray, C. W.;
Taylor, R. D. Improved protein-ligand docking using GOLD. Proteins:
Struct., Funct., Genet. 2003, 52, 609−623.
(35) Case, D. A.; Darden, T. A.; Cheatham, T. E., III; Simmerling, C.
L.; Wang, J.; Duke, R. E.; Luo, R.; Walker, R. C.; Zhang, W.; Merz, K.
M.; Roberts, B.; Hayik, S.; Roitberg, A.; Seabra, G.; Swails, J.; Goetz, A.
Winckler, T.; Wunsch, B. Conformationally constrained NR2B
̈
selective NMDA receptor antagonists derived from ifenprodil:
Synthesis and biological evaluation of tetrahydro-3-benzazepine-1,7-
diols. Bioorg. Med. Chem. 2010, 18, 8005−8015.
(18) Ordonez, M.; De La Cruz-Cordero, R.; Fernandez-Zertuche,
M.; Munoz-Hernandez, M. A.; Garcia-Barradas, O. Diastereoselective
reduction of dimethyl γ-[(N-p-toluenesulfonyl)amino]-β-ketophosph-
onates derived from amino acids. Tetrahedron: Asymmetry 2004, 15,
3035−3043.
́
W.; Kolossvary, I.; Wong, K. F.; Paesani, F.; Vanicek, J.; Wolf, R. M.;
Liu, J.; Wu, X.; Brozell, S. R.; Steinbrecher, T.; Gohlke, H.; Cai, Q.; Ye,
X.; Wang, J.; Hsieh, M.-J.; Cui, G.; Roe, D. R.; Mathews, D. H.; Seetin,
M. G.; Salomon-Ferrer, R.; Sagui, C.; Babin, V.; Luchko, T.; Gusarov,
S.; Kovalenko, A.; Kollman, P. A. AMBER 12; University of California:
San Francisco, CA, 2012.
(19) Chen, B. C.; Skoumbourdis, A. P.; Guo, P.; Bednarz, M. S.;
Kocy, O. R.; Sundeen, J. E.; Vite, G. D. A facile method for the
transformation of N-(tert-butoxycarbonyl) α-amino acids to N-
unprotected α-amino methyl esters. J. Org. Chem. 1999, 64, 9294−
9296.
(20) Swamy, C. C. K.; Kumar, N. N. B.; Balaraman, E.; Kumar, K. V.
P. P. Mitsunobu and related reactions: advances and applications.
Chem. Rev. 2009, 109, 2551−2651.
(21) Schlager, T.; Oberdorf, C.; Tewes, B.; Wunsch, B. Novel, one-
̈
̈
pot procedure for the synthesis of 2-arylethanol derivatives. Synthesis
2008, 2008, 1793−1797.
(22) Duhamel, L.; Fouquay, S.; Plaquevent, J.-C. Ligand exchange in
asymmetric reactions of lithium enolates: application to the
deracemization of α-aminoacids. Tetrahedron Lett. 1986, 27, 4975−
4978.
(23) Tewes, B.; Frehland, B.; Frohlich, R.; Wunsch, B. Novel
̈
̈
GluN2B selective NMDA receptor antagonists: relative configuration
of 7-methoxy-2-methyl-tetrahydro-3-benzazepin-1-ols. Acta Crystal-
logr., Sect. E, submitted 2015.
(24) Tewes, B.; Frehland, B.; Frohlich, R.; Wunsch, B. Relative
̈
̈
configuration of 7-benzyloxy-2-methyl-3-tosyltetrahydro-3-benzazepin-
1-ol for the elucidation of the relative configuration of potent allosteric
GluN2B selective NMDA receptor antagonists. Acta Crystallogr., Sect.
E, submitted 2015.
(25) Schepmann, D.; Frehland, B.; Lehmkuhl, K.; Tewes, B.;
Wunsch, B. Development of a selective competitive receptor binding
̈
assay for the determination of the affinity to NR2B containing NMDA
receptors. J. Pharm. Biomed. Anal. 2010, 53, 603−608.
M
DOI: 10.1021/acs.jmedchem.5b00897
J. Med. Chem. XXXX, XXX, XXX−XXX