Systematic variation of the benzoylhydrazine moiety of the GluN2A selective NMDA receptor antagonist TCN-201

Article abstract of DOI:10.1016/j.ejmech.2018.09.006

GluN2A containing N-methyl-D-aspartate receptors (NMDARs) are important ion channels in the central nervous system and highly involved in several different neurophysiological but also neuropathophysiological processes. However, current understanding of th

Full text of DOI:10.1016/j.ejmech.2018.09.006

European Journal of Medicinal Chemistry 158 (2018) 259e269
Contents lists available at ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Research paper
Systematic variation of the benzoylhydrazine moiety of the GluN2A
selective NMDA receptor antagonist TCN-201
Julian A. Schreiber 1, Sebastian L. Müller a , Stefanie E. Westph a€ linger ,
a
,
,
1
a
a
b
b, **
Dirk Schepmann , Nathalie Strutz-Seebohm , Guiscard Seebohm
,
a, c, *
Bernhard Wünsch
a
Institute of Pharmaceutical and Medicinal Chemistry, University of Münster, Corrensstr. 48, Münster, D-48149, Germany
Myocellular Electrophysiology and Molecular Biology, Institute for Genetics of Heart Diseases (IfGH), Department of Cardiovascular Medicine, University
b
Hospital Münster, Münster, D-48149, Germany
c
Cells-in-Motion Cluster of Excellence (EXC 1003 e CiM), University Münster, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 22 May 2018
Received in revised form
GluN2A containing N-methyl-D-aspartate receptors (NMDARs) are important ion channels in the central
nervous system and highly involved in several different neurophysiological but also neuro-
pathophysiological processes. However, current understanding of the contribution of GluN2A containing
NMDARs in these processes is incomplete. Therefore, highly selective compounds are required to further
investigate these ion channels. In 2010, TCN-201 (2), one of the first selective negative allosteric mod-
ulators was reported. While the binding site of 2 and the influence of the substitution pattern of the
benzenesulfonamide part has been reported recently, detailed structure-activity-relationships of the
diacylhydrazine part and the linked phenyl moiety are still missing. In order to examine the critical
interactions between these moieties and the binding site, several TCN-201 analogs with modified
diacylhydrazine part were synthesized. The negative allosteric effect was recorded by two-electrode
voltage clamp (TEVC) experiments using GluN1a/GluN2A expressing Xenopus laevis oocytes. Our data
23 August 2018
Accepted 3 September 2018
Available online 6 September 2018
Keywords:
NMDA receptor
GluN2A selective antagonists
TCN-201
Synthesis
Electrophysiology
Two-electrode voltage clamp
led to the conclusion, that the terminal phenyl moiety is involved in a cation-p-interaction with the
guanidinium moiety of Arg755 of the GluN1a subunit, which plays a crucial role for high activity.
Additionally, structure optimization by replacing the phenyl moiety with a thiophen-2-yl (10c), indol-2-
yl (10g) or indol-3-yl (10h) moiety significantly increased the activity of 2 by the factor 2.5. At a test
compound concentration of 200 nM, the negative allosteric effect of the most potent ligands 10c, 10h and
1
7 was significantly influenced by the glycine concentration. Although glycine dependency is higher than
those of the lead compound 4, 10c and 17 showed significantly higher negative allosteric effects than 4 at
glycine concentrations from 1 M up to 10 M. The potent GluN2A-NMDA receptor inhibitors 10c, 10h
and 17 did not influence the ion current of GluN2B-NMDA receptors.
m
m
©
2018 Elsevier Masson SAS. All rights reserved.
1
. Introduction
important excitatory ionotropic glutamate receptors in the central
nervous system. To trigger channel opening and conduction,
binding of the two native ligands (S)-glutamate and glycine
together with slight depolarization of the surrounding membrane
Ionotropic glutamate receptors are involved in many different
neurophysiological processes, such as learning, fast excitatory
signal transmission and long-term synaptic plasticity [1e4]. The N-
methyl-D-aspartate receptor (NMDAR) belongs to the most
2
þ
is needed. Without membrane depolarization, Mg -ions block the
pore and prevent ion flux through the selectivity filter pore of the
channel [5,6]. Opening of the NMDAR associated ion channel leads
þ
þ
2þ
to an efflux of K -ions and influx of Na - and Ca -ions resulting in
*
Corresponding author. Institute of Pharmaceutical and Medicinal Chemistry,
depolarization of postsynaptic membranes and initialization of
secondary mechanisms mediated by Ca -ions [7,8]. In contrast,
University of Münster, Corrensstr. 48, Münster, D-48149, Germany.
2þ
** Corresponding author.
overactivation of NMDARs can lead to accumulation of intracellular
E-mail address: wuensch@uni-muenster.de (B. Wünsch).
Both authors contributed equally.
2
þ
1
Ca -ions with subsequent excitotoxicity and cell death [9].
https://doi.org/10.1016/j.ejmech.2018.09.006
223-5234/© 2018 Elsevier Masson SAS. All rights reserved.
0

260
J.A. Schreiber et al. / European Journal of Medicinal Chemistry 158 (2018) 259e269
The high functional diversity of NMDARs and their involvement
modulators (NAMs) in 2010: compound 6 (1) and TCN-201 (2) [27]
(Fig. 2). A mutagenesis study identified the binding site of 2 at the
interface of the LBDs of GluN1 and GluN2A subunits (Figs. 1 and 3).
The residues forming the binding pocket contributing to the
binding of 2 are located at the late S2 segment of GluN2A and
GluN1 subunits. Especially, the residues Leu780 and Val783 of the
GluN2A subunit and Arg755 of the GluN1 subunit were identified
as crucial residues for inhibitory activity of 2 [28]. Recently, these
results for binding site localization were confirmed by co-
crystallization of the GluN1-GluN2A LBDs with 1 and 2 [17,29]
(Figs. 1 and 3).
in many different processes in the central nervous system are based
on various combinations of the seven different subunits (GluN1,
GluN2A-D, GluN3A-B) in different heterotetrameric complexes
[
10,11]. The majority of functional NMDARs contain two mandatory
GluN1 subunits, which harbor the binding site for glycine and can
be expressed in 8 different splice variants (GluN1a-h) [12].
Furthermore, functional NMDARs contain at least one of the four
GluN2 subunits comprising the binding site for (S)-glutamate [13].
The two different GluN3 subunits, which also have a binding site for
glycine, are often part of triheteromeric receptors with two GluN1
and one GluN2 subunit [14]. Additionally, triheteromeric receptors
with two different GluN2 subunits have been described [15].
All subunits GluN1-3 share a similar architecture, which can be
divided into four domains [18] (Fig. 1). The N-terminal clamshell-
like amino-terminal domain (ATD) modulates channel gating
directly by interacting with the ligand binding domain (LBD). This
domain is formed by the first 400 amino acids [19]. The LBD also
showing a clamshell-like structure can be subdivided into the up-
per S1 and the lower S2 region, which together form the glycine
binding site (GluN1, GluN3) or the (S)-glutamate binding site
While Val783 of the GluN2A subunit and Phe754 of the GluN1
subunit are responsible for the inhibition mechanism and selec-
tivity, Arg755 and Tyr535 of the GluN1 subunit are involved in
binding of 2 by cation-p- and p-p-interactions, respectively (Fig. 3).
In order to optimize the moderate activity, the high glycine
dependency and the insufficient solubility of 1 and 2, lead com-
pound 1 was further developed into MPX-004 (3) and MPX-007 (4)
[30]. While structural modifications reduced glycine dependency
and increased solubility, activity of 3 and 4 is only slightly increased
compared to the activity of lead compound 2 at glycine concen-
(
GluN2A-D), respectively [20]. The ion channel pore is formed by
trations below 10 mM [17].
the transmembrane domain (TMD) containing three helices M1,
M3, M4 and the reentry loop M2 and shows close structural simi-
larity to potassium channels [21]. The intracellular carboxy-
terminal domain (CTD) influences protein stability, conductivity,
trafficking and regulation of the NMDARs [22].
The functional properties of the different NMDARs are mostly
depending on the expressed GluN2 subunit [10,16]. Moreover, the
four GluN2 subunits exert distinct expression patterns in particular
brain regions and cell types [23]. Both facts result in differences in
their association with several neurophysiological and neuro-
pathophysiological processes and relevance for different diseases.
The involvement of GluN2A subunit containing NMDARs in anxiety
disorders, chronic schizophrenia, depression, seizure disorders and
cerebral ischemia was shown in different studies. However, the
precise role of GluN2A-NMDARs in these diseases remains to be
elucidated [24e26].
To identify the central interactions of these compounds with
specific residues located in the binding pocket additional infor-
mation about structure activity relationships of GluN2A selective
NAMs are required. Recently, we reported on the structural modi-
fications of the benzenesulfonamide part of TCN-201 (2). It was
found that a Br-atom in 3-position (5) resulted in a 2.5 fold
increased ion channel inhibition compared to 2 [31]. However, in
crystal structures and mutagenesis studies the aromatic system of
the diacylhydrazine part of the compound was identified to
participate in binding and maybe also in activity of 2 by a cation-
p
interaction with Arg755 (GluN1a) [17,28]. To analyze the influence
of this interaction on activity in more detail derivatives of TCN-201
(2) with a modified benzoylhydrazine moiety were designed.
2. Synthesis
New possibilities for targeting selectively GluN2A-NMDARs and
investigate their contribution in these diseases arose by the dis-
covery of two highly selective GluN2A-NMDAR negative allosteric
In order to get fast access to analogs of 2 with various acyl
moieties at the hydrazide part the concept of late stage diversifi-
cation was followed. For this purpose, the hydrazide 9 was selected
Fig. 1. Schematic depiction of triheteromeric GluN1 (yellow)/GluN2A (red)/GluN2B (blue) NMDAR (left, PDB ID 5UOW), single GluN2A subunit (right, PDB ID 5UOW) and their
subdivision into ATD, LBD and TMD [16]. Overlay of triheteromeric NMDAR (PDB ID 5UOW) with crystal structure of 2 (PDB ID 5I56) reveals binding site localization at the LBD
interface of GluN1 and GluN2A subunits with binding contribution of residues from the S2 segment [17]. (For interpretation of the references to color in this figure legend, the
reader is referred to the Web version of this article.)

J.A. Schreiber et al. / European Journal of Medicinal Chemistry 158 (2018) 259e269
261
Fig. 2. Development of potent GluN2A subunit selective NAMs.
Fig. 3. X-ray crystal structure of GluN1-GluN2A-LBD co-crystallized with 2 (PDB ID 5I56) [17]. Amino acids important for the interaction with the ligand are shown. The GluN1
subunit is displayed in yellow and the GluN2A subunit in red. On the right, the glycine binding pocket together with glycine can be identified. (For interpretation of the references to
color in this figure legend, the reader is referred to the Web version of this article.)
as key intermediate. The hydrazide 9 was prepared in three steps as
previously reported. Reaction of 4-(aminomethyl)benzoic acid (6)
with sulfonyl chloride provided the sulfonamide 7, which was
diacylhydrazine moiety, its flexibility is reduced and its H-bond
donor and acceptor properties are altered. Further diversity was
introduced by synthesis of the sulfonamide 12 by sulfonylation of
hydrazide 9 and benzyl derivative 13 bearing only one carbonyl
moiety by acylation of benzylhydrazine with the carboxylic acid 7
®
coupled with Boc-protected hydrazine in the presence of COMU .
Cleavage of the Boc-protective group of 8 afforded the mono-
acylhydrazide 9 (Scheme 1) [31], which was acylated with diverse
acid chlorides, acids and anhydrides to give a diverse set of diac-
ylhydrazines 10a-k.
®
and COMU .
In the electrophysiological experiments, the thiophene deriva-
tive 10c exhibited the highest inhibition of the GluN2A-NMDA re-
In order to investigate the influence of cation-
p
-interactions
ceptor associated ion channel. In a recent study, the 3-
between the positively charged side chain of Arg755 of the GluN1a
splice variant with the aryl system R of the ligands, particular
attention was paid on the size and electron density of the (hetero)
aromatic ring system. Furthermore, acidic (e.g. 10i) and basic
functional groups (e.g. 10a) were introduced and the benzoyl
moiety of 2 was replaced by a phenylacetyl (10k) and morpholi-
nocarbonyl moiety (10j). The substituents R of compounds 10 are
defined in Scheme 1.
bromophenyylsulfonyl moiety was identified as optimal substitu-
ent on the benzylamino group [31]. Therefore, the hybrid molecule
17 was designed combining the best structural elements on both
sides, i.e. the thiophen-2-ylcarbonyl moiety at the hydrazide and
the 3-bromophenylsulfonyl moiety on the left-hand side. Thus, 4-
(aminomethyl)benzoic acid (6) was sulfonated with 3-
2 4 7
bromophenylsulfonyl chloride in Na B O solution to yield the
sulfonamide 14 (Scheme 3). Coupling of 14 with Boc-protected
hydrazine provided the carbamate 15, which was deprotected
with HCl in 1,4-dioxane. The resulting free hydrazine 16 was
coupled with thiophene-2-carboxylic acid to yield the hybrid
antagonist 17.
As a further modification of the 1,2-diacylhydrazine structural
element non-classical bioisosteric 1,3,4-oxadiazole was planned
(
Scheme 2) [32]. For this purpose, diacylhydrazine 2 was treated
with SOCl to form the 1,3,4-oxadiazole 11 in 92% yield. Although
the 1,3,4-oxadiazole ring seems to be similar to the 1,2-
2

262
J.A. Schreiber et al. / European Journal of Medicinal Chemistry 158 (2018) 259e269
ꢀ
Scheme 1. Reagents and reaction conditions: (a) 3-Cl-4-F-PhSO
31]. (c) 1. TFA, CH Cl , rt, 1 h; 2. Et N, MeOH, rt, 5 min, 81% [31]. (d) R-COCl pyridine, CH
6 h.
2
Cl, Na
2
B
4
O
7
, H
2
O, rt, 4 h, 98% [31]. (b) 1. SOCl
2
, 75 C, 1 h; 2. tert-Butyl carbazate, pyridine, CH
2
Cl
2
, reflux, 2 h, 90%
®
[
1
2
2
3
2
Cl
2
, reflux 3e16 h (e) R-CO
2
H, COMU , DIPEA, THF, rt, 16 h. (f) Succinic anhydride, THF, rt,
®
Scheme 2. Reagents and reaction conditions: (a) SOCl
6%.
2 2 2 2
, reflux, 1 h, 92%. (b) PhSO Cl, pyridine, CH Cl , reflux 3 h, 85%. (c) Benzylhydrazine hydrochloride, COMU , DIPEA, THF, rt, 3 d,
4

J.A. Schreiber et al. / European Journal of Medicinal Chemistry 158 (2018) 259e269
263
®
2 4 7 2
Scheme 3. Reagents and reaction conditions: (a) 3-Bromophenylsulfonyl chloride, Na B O , H O, rt, 16 h, 99%. (b) tert-Butyl carbazate, COMU , DIPEA, THF, rt, 16 h, 81%. (c) HCl, 1,4-
®
dioxane, rt, 16 h, 97%. (d) Thiophene-2-carboxylic acid, COMU , DIPEA, THF, rt, 16 h, 67%.
3
. Pharmacological activity (Results and discussion)
Table 1
Normalized Inhibition Inorm of test compounds. Inhibition at 500 nM compound was
normalized relative to the inhibition of 2 at 500 nM tested within the same series of
experiments. n reflects the number of independent oocytes.
3.1. Negative allosteric modulation of novel ligands
compound
n
normalized inhibition
SEM
Pharmacological activity was evaluated by two-electrode
voltage clamp experiments using Xenopus laevis oocytes express-
ing GluN1a/GluN2A NMDARs. Channel activation in each experi-
I
norm
2
22
5
5
5
5
5
5
5
5
5
5
5
3
5
5
1,00
0,92
1,10
1,35
1,20
0,44
0,58
1,28
1,33
0,30
0,52
0,64
0,31
0,58
0,73
0,02
0,09
0,02
0,02
0,09
0,11
0,04
0,06
0,03
0,04
0,04
0,02
0,11
0,02
0,06
1
1
1
0a
0b
0c
ment was evoked by 10 mM (S)-glutamate and 10 mM glycine in
2
þ
2þ
Mg -free Ba -Ringer solution. To identify potent compounds for
further characterization a screening was performed at a concen-
tration of 500 nM using 3 to 5 different oocytes for each compound.
The inhibition was normalized related to the inhibition of TCN-201
10d
10e
1
1
1
1
0f
0g
0h
0i
(
2) at a concentration of 500 nM (inhibition ¼ 1.0, black bar in
Fig. 4) in the same series of experiments (Fig. 4 and Table 1).
Screening of compounds led to clear relationships between the
structure and the negative allosteric modulation of NMDARs con-
taining the GluN2A subunit. Replacement of the terminal phenyl
moiety of the diacylhydrazine part with electron rich five-
membered heterocycles such as furan-2-yl (10b), thiophen-2-yl
10j
1
1
1
1
0k
1
2
3
(10c) or pyrrol-2-yl (10d) led to higher potency compared to the
lead compound 2. Whereas the activity of the furan and pyrrole
derivatives 10b and 10d is not significantly increased, the potency
of the thiophene derivative 10c is significantly higher (p < 0.001)
than the potency of 2. The order of activity correlates with the
aromaticity of the three different heterocycles increasing from
furan (10b) (least aromatic, least active) to thiophene (10c) (most
aromatic, most active) [33]. However, differences between the ac-
tivity of 10b, 10c and 10d are not significant. Introduction of the
electron deficient six-membered pyrimidin-4-yl substituent (10a)
resulted in decreased activity (not significant) compared to the
phenyl compound 2, which is, however, significantly decreased
compared to the activity of the thiophene and pyrrole derivatives
10c and 10d (p < 0.05e0.001). These observations are in good
agreement with the cation- -interaction hypothesis between these
p
compounds and Arg755 and also with the observation of previous
mutagenesis studies [28].
Enlargement of the monocyclic to a bicyclic aromatic system by
introduction of an indol-2-yl (10g; p < 0.01) or indol-3-yl (10h;
p < 0.001) substituent led to significantly increased antagonistic
activity compared to 2 with a small preference for 10h bearing the
indol-3-yl moiety. In contrast, naphthalen-2-yl (10e) and 1-
benzothiophen-2-yl (10f) residues were less tolerated in GluN2A-
NMDAR inhibition and reduced the activity of 10e and 10f signifi-
cantly compared to the activity of the lead compound 2 and the
indole derivatives 10g and 10h. This result can be interpreted as a
Fig. 4. Normalized Inhibition Inorm of test compounds with heterocyclic aromatic
moieties (red), enhanced aromatic ring systems (green), non-aromatic moieties (yel-
low) and modified diacylhydrazine part (blue). Inhibition at a test compound con-
centration of 500 nM normalized to the inhibition of 500 nM TCN-201 (2, black,
n ¼ 22); n ¼ 3e5, ±SEM, One-Way ANOVA, post hoc mean comparison Tukey Test
compared to 2, ns p > 0.05, *p < 0.05, **p < 0.01, ***p < 0.001. (For interpretation of the
references to color in this figure legend, the reader is referred to the Web version of
this article.)

264
J.A. Schreiber et al. / European Journal of Medicinal Chemistry 158 (2018) 259e269
size limitation of the receptor binding pocket around the terminal
aromatic system.
increased activity, the inhibitory activity was recorded at different
glycine concentrations. All test compounds 2, 4, 10c, 10h, and 17
were tested at a concentration of 200 nM and the inhibition of the
The hypothesis of cation-
p interaction involving the terminal
aromatic substituent at the diacylhydrazine part also correlates
well with the activity of compounds with non-aromatic sub-
stituents at this position. Replacement of the terminal aromatic
ring by a short alkyl chain bearing a carboxylic acid (10i) or a
morpholine ring (10j) reduced significantly the negative allosteric
modulation activity of the compounds (p < 0.001).
While replacement of the terminal phenyl ring of 2 by hetero-
cyclic aromatic moieties could increase the inhibitory activity, all
modifications of the diacylhydrazine part led to significant reduc-
tion of NAM activity. Removal of the terminal carbonyl group led to
a moderate, but significant reduction in activity of 13 compared to 2
ion currents were tested at glycine concentrations of 1
10 M and 30 M (Fig. 6).
While activity of 2 decreases dramatically at low glycine con-
centrations of 1 and 3 M, 10c, 10h, 17 and 4 showed an almost
linear development of the activity over the complete range of
glycine levels. To quantify the glycine dependency and activity of 2,
4 and the synthesized compounds the data were fitted in a linear
matter. The calculated intercept and slope of the linear fitting as
measure for the glycine dependency are shown in Table 3.
mM, 3 mM,
m
m
m
Due to the fact that increment of glycine concentration from
1 mM to 3 mM influences activity of 2 much more than further
(
p < 0.01). Obviously, the terminal carbonyl moiety, which is
increment of glycine concentration, linear dependency is not given
and calculated slope of linear fit is not comparable to the other
calculated slopes. The different slopes for 4, 10c, 10h and 17 show
nearly linear dependency for all tested glycine levels and also
indicate a clear difference in glycine dependency between these
four compounds. Whereas the calculated slopes for the synthesized
compounds 10c, 10h and 17 are almost identical (ꢁ54 to ꢁ57), the
decreased slope determined for the lead compound 4 (ꢁ41) in-
dicates lower glycine dependency. This result led to the conclusion
that the newly synthesized compounds display identical glycine
dependency, which is higher than the glycine dependency of 4, but
lower than those of 2 due to the dramatically loss of activity even at
low glycine concentrations.
missing in the lead compounds 3 and 4, contributes to the inhibi-
tory activity of this class of TCN-201 analogs. Introduction of a
methylene spacer between the carbonyl moiety and the terminal
phenyl ring (10k) or bioisosteric replacement of this carbonyl
moiety with a sulfonyl group (12) also led to significantly decreased
activity. The reduced activity may result from weakening the
cation-p-interactions between the terminal aromatic system of the
ligands with the guanidinium moiety of Arg755 of the GluN1a
subunit. Finally, the bioisosteric replacement of the complete
diacylhydrazine part by a 1,3,4-oxadiazole ring in 11 decreased
significantly the activity, which could be caused by reduced flexi-
bility or also by displacement of the terminal aromatic system away
from Arg755.
At glycine concentrations of 1, 3 and 10 mM the newly synthe-
In order to quantify exactly the antagonistic activity of the most
promising compounds 10c and 10h dose response curves were
generated. Additionally, the hybrid compound 17 was included in
these experiments combining the most promising structural ele-
ments in one molecule, i.e. the 3-bromophenyl group of 5 on the
left side and the thiophen-2-yl moiety of 10c on the right side. In
Table 2 the recorded IC50 values are summarized and compared to
previously recorded IC50 values of 2, 4 and 5.
sized test compounds 10c, 10h and 17 (200 nM) show significantly
(p < 0.001e0.05) higher GluN2A-NMDAR inhibitory activity than
the lead compounds 2 and 4 (200 nM). Only the inhibitory activity
of the indole derivative 10h is not significantly higher than the
inhibitory activity of 4 at a glycine concentration higher than 1 mM
(p > 0.05). Although the differences in activity between 10c, 10h
and 17 are not significant (p > 0.05), the order of activity correlates
nicely with the previously determined order in the screening.
The recorded inhibitory activity of 2 and 4 at a concentration of
200 nM confirms the recently published GluN2A activity of these
The new test compounds 10c, 10h and 17 show a 2.5-fold higher
GluN2A inhibitory activity than the lead compound 2 (Fig. 5,
Table 2). The corresponding IC50 values are in the range of
lead compounds [17]. At low glycine concentrations (1
show similar activity (p > 0.05). However, at higher glycine con-
centrations (3 M, 10 M) the activity of 4 is slightly increased
compared to the activity of 2 (p < 0.05). At the highest glycine
concentration recorded (30 M) the activity of 2 and 4 at 200 nM
are very similar (p > 0.05). Considering only the inhibitory activity
of 2 and 4 at glycine concentrations ꢂ3 M, reveals a higher glycine
dependency for TCN-201 (2) than for MPX-007 (4).
mM), 2 and 4
2
00 nM at a glycine level of 10 mM. Surprisingly, the IC50 values of
compounds 5, 10c and 17 are not significantly different, which in-
dicates that combination of optimal substitution patterns of both
terminal aromatic rings does not lead to additive effects.
m
m
m
m
3.2. Glycine dependency of NAM activity at GluN2A-NMDA
receptors
3.3. Selectivity over GluN2B-NMDA receptors
Recently published IC50 values of 2 and 4 in two-electrode
voltage clamp measurements are also in the range of 200 nM.
However, in these experiments the glycine concentration was only
In order to investigate, whether modifications of the substitu-
tion pattern influence the selectivity of 10c, 10h and 17 the inhib-
itory activity at GluN1a/GluN2B containing NMDARs was recorded
as well. Due to the moderate solubility, only concentrations of
3
m
M instead of 10
is reduced to an IC50 value of 512 nM at a glycine concentration of
M, comparable IC50 values at this glycine concentration are not
mM in our experiments [17]. Whereas activity of
2
10
m
10
mM of these test compounds together with 2 and ifenprodil
available for the lead compound 4. In order to analyze whether
structural variations increased the activity of the novel test com-
pounds or the reduced glycine sensitivity is responsible for the
(10
mM each) as GluN2B selective inhibitor were tested. Each
compound was tested at five different oocytes. The resulting data
presented in Fig. 7 display that selectivity towards GluN2B con-
taining NMDARs is not influenced by the modification of the sub-
stitution pattern (Fig. 7).
Table 2
IC50 values ± SEM (nM) of compounds 2, 5, 10c, 10h and 17 in presence of 10
glycine compared to previous published IC50 value of 4 in presence of 3 M glycine
17].
mM
m
4. Conclusion
[
Introduction of various aryl moieties confirm the contribution of
2
4
5
10c
10h
17
cation-p-interactions between the guanidinium moiety of Arg755
IC50 ± SD [nM] 512 ± 64 198 ± 21 204 ± 28 198 ± 23 216 ± 44 222 ± 16
within the GluN1a subunit and the terminal aryl moiety of the

J.A. Schreiber et al. / European Journal of Medicinal Chemistry 158 (2018) 259e269
265
Fig. 5. Dose response curves of 10c (left), 10h (middle) and 17 (right) (mean ± SEM; n ¼ 5 oocytes for each concentration).
Fig. 6. Glycine dependency of GluN2A-NMDAR inhibition. Test compounds were
employed in a concentration of 200 nM. The ion current inhibition (mean ± SEM; n ¼ 3
different oocytes for each glycine concentration) was recorded after stimulation of the
Fig. 7. Current inhibition at GluN1a e GluN2B NMDARs. All compounds were tested at
a concentration of 10
oocytes for each compound).
m
M at GluN1a/GluN2B NMDARs (mean ± SEM; n ¼ 4e5 different
oocytes with 10
mM (S)-glutamate and glycine concentrations of 1 mM, 3 mM, 10 mM and
30 m
M. Data were fitted linearly by equation y ¼ a þ b*x (a ¼ intercept, b ¼ slope).
The thiophene derivative 10c, the indole derivative 10h and the
hybrid ligand 17 show 2.5-fold higher activity at GluN2A containing
NMDARs than the lead compound 2, but retain selectivity over
GluN2B containing NMDARs. Unfortunately, combination of struc-
tural elements optimized for the left- and right-hand side in the
hybrid ligand 17 did not lead to an additive increase of GluN2A
activity suggesting that 10c and 17 are close to the optimal scaffold.
Compounds 10c, 10h and 17 (200 nM) showed significantly
Table 3
Linear fitting coefficients for glycine dependency of 2, 4, 10c, 10h and 17. Mean in-
hibitions of 3 different oocytes for each glycine concentration were fitted linearly to
reveal slope, intercept, standard error (SE) of calculated values and coefficient of
determination for the linear fit (COD, R-Square).
2
4
10c
10h
17
Intercept ±SE
slope ±SE
R-Square (COD)
47 ± 7
ꢁ30 ± 5
0.94
65 ± 3
ꢁ41 ± 2
0.99
93 ± 5
ꢁ57 ± 5
0.99
88 ± 2
ꢁ54 ± 1
0.99
95 ± 2
ꢁ57 ± 2
0.99
higher activity at glycine concentrations from 1
mM (10c, 10h, 17) up
to 10 M (10c, 17) than lead compounds 2 and 4. According to the
m
determined slope the glycine dependency of 10c, 10h and 17 is
higher than the glycine dependency of the lead compound 4.
ligands. Screening of the synthesized compounds led to decreasing
activity from electron rich heterocycles over electron deficient
heterocycles to non-aromatic substituents. In particular the very
low activity of compounds without a terminal aryl moiety un-
4.1. Experimental part
4.1.1. Chemistry, general
Moisture sensitive reactions were conducted under dry nitro-
gen. THF was dried with sodium/benzophenone and was freshly
distilled before use. Thin layer chromatography: Silica gel 60 F254
plates (Merck). Flash chromatography (fc): Silica gel 60, 40e43 mm
(Macherey-Nagel); parentheses include: diameter and length of the
derscores the importance of the cation-p-interaction for high ac-
tivity. These data are in accordance with previous mutagenesis
studies showing loss of activity for 2 at NMDARs with mutation
Arg755Ala within the GluN1a subunit [28]. However, the size of the
aromatic system seems to be limited to the size of an indole ring.
Whilst the results of this study clearly display the role of the ter-
minal aromatic system, the contribution of the diacylhydrazine
moiety to the GluN2A activity is not completely understood.
column, eluent, R value. In order to obtain high yields and due to
f
the poor solubility some compounds were adsorbed on silica gel by
addition of silica gel to a solution of the compound in an appro-
priate solvent, removal of the solvent in vacuo and giving the

266
J.A. Schreiber et al. / European Journal of Medicinal Chemistry 158 (2018) 259e269
mixture on top of the column. Melting point: Melting point system
MP50 (Mettler Toledo), uncorrected. NMR spectra were recorded
4.2.3. 3-Chloro-4-fluoro-N-[4-({2-[(naphthalen-2-yl)carbonyl]
hydrazino}carbonyl)benzyl]benzenesulfonamide (10e)
1
13
on Agilent DD2 spectrometers ( H NMR: 600 MHz, 400 MHz;
C
2-Naphthoyl chloride (94.0 mg, 0.49 mmol) suspended in
CH2Cl2 (2 mL) was added dropwise to a solution of 9 (176 mg,
0.49 mmol) in a mixture of CH2Cl2 (8 mL) and pyridine (2 mL). The
resulting solution was heated to reflux overnight. The reaction
mixture was cooled to rt and was extracted with 1 M NaOH
(3 ꢃ 15 mL). The combined aqueous layer was washed with CH2Cl2.
Afterwards, the aqueous layer was acidified with conc. HCl to adjust
pH 1 and the acidic solution was neutralized by addition of satu-
rated NaHCO3 solution. The formed precipitate was filtered,
washed with H2O and dissolved in acetone. After removal of the
solvent under reduced pressure, the formed solid was recrystallized
NMR: 151 MHz, 100 MHz); in ppm related to tetramethylsilane;
d
coupling constants are given with 0.5 Hz resolution; the assign-
ments of 13C and 1H NMR signals were supported by 2D NMR
13
techniques. For better comparability, assignments of aromatic
and H signals were done according to the following Figure.
C
1
Aromatic signals, which are part of the 3-chloro-4-fluoro moiety
were marked with A. Aromatic signals which are part of the 4-
(
aminomethyl)benzoate moiety were marked with B, whereas the
signals, which are part of the derivatization were marked with C. If
the compound did not include one or more of the three parts, the
marks for these parts are no longer required.
ꢀ
from acetone/H2O. Colorless solid, mp 233 C, yield 45.0 mg (18%).
C25H19ClFN3O4S (511.6). Rf ¼ 0.67 (ethyl acetate þ 1% NH3 solu-
tion (25%), detection: 254 nm). 1H NMR (400 MHz, DMSO-D6):
[ppm] ¼ 4.15 (d, J ¼ 6.1 Hz, 2H, NHCH2), 7.39 (d, J ¼ 8.0 Hz, 2H,
4
4
.2. Purity
2
-HB, 6-HB), 7.59e7.70 (m, 3H, 5-HA, 6-HC, 7-HC), 7.82 (ddd,
.2.1. The purity of all compounds was determined by HPLC analysis
HPLC (method 1): Merck Hitachi Equipment; UV detector: L-
J ¼ 8.9/4.7/2.4 Hz, 1H, 6-HA), 7.87 (d, J ¼ 8.1 Hz, 2H, 3-HB, 5-HB),
7.93 (dd, J ¼ 6.9/2.3 Hz, 1H, 2-HA), 7.97e8.10 (m, 4H, 3-HC, 4-HC, 5-
HC, 8-HC), 8.44 (t, J ¼ 6.3 Hz, 1H, SO2NHCH2), 8.56 (s, 1H, 1-HC),
10.53 (s, 1H, NHNHCOArC), 10.65 (s, 1H, ArBCONHNH). HPLC
(Method 1): tR ¼ 18.9 min, purity 99.9%.
7400; autosampler:L-7200; pump: L-7100; degasser: L-7614; col-
®
®
umn: LiChrospher 60 RP-select B (5
cartridge; flow rate: 1.0 mL/min; injection volume: 5.0
tion at O with 0.05% (v/v)
trifluoroacetic acid; solvent B: acetonitrile with 0.05% (v/v) tri-
fluoroacetic acid: gradient elution (% A): 0e4 min: 90.0%;
m
m); LiCroCART 250-4 mm
mL; detec-
l
¼ 210 nm; solvent A: demineralized H
2
4.2.4. 3-Chloro-4-fluoro-N-[4-({2-[(indol-2-yl)carbonyl]hydrazino}
carbonyl)benzyl]benzenesulfonamide (10g)
4
e29 min: gradient from 90% to 0%; 29e31 min: 0%; 31e31.5 min:
A mixture of indole-2-carboxylic acid (80.0 mg, 0.50 mmol) and
ꢀ
gradient from 0% to 90%; 31.5e40 min: 90%.
SOCl
2
(2 mL) was heated to 72 C under reduced pressure
HPLC (method 2): Thermo Fisher Scientific Equipment; UV de-
(950 mbar) and was stirred for 1 h. The excess of SOCl
removed under reduced pressure and the resulting acyl chloride
was dissolved in CH Cl (10 mL). 9 (176 mg, 0.49 mmol) dissolved in
a mixture of CH Cl (10 mL) and pyridine (2 mL) was added drop-
wise to the solution and the resulting solution was heated to reflux
for 5 h. After cooling to rt, it was extracted with 1 M NaOH
2
was
tector: VWD-3400RS; autosampler:ACC-3000T; pump: LPG-
®
3
400SD; degasser: DG-1210; column: LiChrospher 60 RP-select
2
2
®
B (5
injection volume: 5.0
demineralized H O with 0.05% (v/v) trifluoroacetic acid; solvent B:
mm); LiCroCART 250-4 mm cartridge; flow rate: 1.0 mL/min;
2
2
m
L; detection at
l
¼ 210 nm; solvent A:
2
acetonitrile with 0.05% (v/v) trifluoroacetic acid: gradient elution (%
A): 0e4 min: 90%; 4e29 min: gradient from 90% to 0%; 29e31 min:
(3 ꢃ 15 mL). The combined aqueous layer was washed with CH
2 2
Cl .
During this procedure a solid was formed, which was filtered and
later treated together with the aqueous layer. Afterwards, the
aqueous layer was acidified with conc. HCl to adjust pH 1 and the
0
%; 31e31.5 min: gradient from 0% to 90.0%; 31.5e40 min: 90%.
According to HPLC analysis the purity of all test compounds is
95%, if not reported otherwise. Selected synthetic procedures for
>
acidic solution was neutralized by addition of saturated NaHCO
solution. The formed precipitate was filtered off, washed with H
and dissolved in CH OH. After removal of the solvent under
3
different synthetic pathways are shown below. Complete synthetic
procedures can be found in Supporting Information.
2
O
3
reduced pressure, the formed solid was recrystallized from
ꢀ
acetone/H
2
O. Pale yellow solid, mp 285 C, yield 85.3 mg (35%).
4
.2.2. 3-Chloro-4-fluoro-N-[4-({2-[(pyrimidine-4-yl)carbonyl
hydrazino}carbonyl)benzyl]benzenesulfonamide (10a)
(176 mg, 0.49 mmol), pyrimidine-4-carboxylic acid (68.0 mg,
.55 mmol) and COMU (250 mg, 0.58 mmol) were suspended in
THF (10 mL). After addition of DIPEA (200 L, 1.15 mmol), the
d
]
C
23
H
18ClFN
4
O
H
4
S
(500.9).
NMR (600 MHz, DMSO-D
J ¼ 6.2 Hz, 2H, NHCH ), 7.07 (t, J ¼ 7.5 Hz, 1H, 5-H
J ¼ 7.6 Hz, 1H, 6-H ), 7.28 (d, J ¼ 2.2 Hz, 1H, 3-H ), 7.39 (d, J ¼ 7.9 Hz,
2H, 2-H , 6-H ), 7.63 (t, J ¼ 8.9 Hz, 1H,
), 7.46 (d, J ¼ 8.3 Hz, 1H, 7-H
), 7.66 (d, J ¼ 8.1 Hz, 1H, 4-H ), 7.81 (ddd, J ¼ 8.7/4.5/2.3 Hz, 1H,
), 7.87 (d, J ¼ 7.9 Hz, 2H, 3-H , 5-H ), 7.92 (dd, J ¼ 6.9/2.3 Hz,1H,
), 8.45 (t, J ¼ 6.3 Hz, 1H, SO NHCH ), 10.49 (s, 1H, Ar CONHNH),
), 11.72 (t, J ¼ 2.3 Hz, 1H, 1-H ). HPLC
¼ 19.9 min, purity 94.7%.
R
f
¼ 0.84 (ethyl acetate, detection:
):
[ppm] ¼ 4.15 (d,
), 7.22 (t,
1
254 nm).
6
d
9
2
C
0
C
C
m
B
B
C
resulting yellow solution was stirred at rt overnight. The solvent
was removed under reduced pressure and the resulting orange oil
was dissolved in 1 M NaOH. The solution was washed with ethyl
acetate (2 x) to remove byproducts. Afterwards, the aqueous layer
was acidified with conc. HCl to adjust pH 1 and the acidic solution
5-H
6-H
2-H
A
A
A
C
B
B
2
2
B
10.53 (s, 1H, NHNHCOAr
(Method 1): t
C
C
R
was neutralized by addition of saturated NaHCO
formed precipitate was filtered off, washed with H O and dissolved
in CH OH. The solvent was removed under reduced pressure. Light
brown solid, mp 209 C, yield 174 mg (76%). C19
3
solution. The
2
4.2.5. 4-[2-(4-{[(3-Chloro-4-fluorophenyl)sulfonamido]methyl}
benzoyl)hydrazinyl]-4-oxobutanoic acid (10i)
3
ꢀ
H
15ClFN
OH 4:1 þ 1% NH solution
25%), detection: 254 nm). H NMR (400 MHz, DMSO-D ):
ppm] ¼ 4.14 (s, 2H, NHCH ), 7.38 (d, J ¼ 7.9 Hz, 2H, 2-H , 6-H ), 7.63
t, J ¼ 8.9 Hz, 1H, 5-H ), 7.78e7.88 (m, 3H, 6-H , 3-H , 5-H ), 7.94
dd, J ¼ 6.8/2.3 Hz, 1H, 2-H ), 8.06 (dd, J ¼ 5.1/1.4 Hz, 1H, 5-H ), 8.45
br s, SO NHCH ), 9.39 (d, J ¼ 1.3 Hz,
), 9.12 (d, J ¼ 5.1 Hz, 1H, 6-H
H, 2-H ), 10.76 (br s, 2H, Ar CONHNHCOAr ). HPLC (Method 1):
¼ 16.3 min, purity 98.2%.
5
O
4
S
9 (176 mg, 0.49 mmol) and succinic anhydride (50.0 mg,
0.50 mmol) were dissolved in THF (10 mL) and stirred at rt over-
night. The solvent was removed under reduced pressure and the
resulting solid was dissolved in 2 M NaOH. The solution was
washed with CH2Cl2 (2 x) to remove byproducts. Afterwards, the
aqueous layer was acidified with conc. HCl to adjust pH 2 and the
formed precipitate was filtered off, washed with 0.1 M HCl and
dissolved in THF. The solvent was removed under reduced pressure.
(
(
[
(
(
(
463.9). R
f
¼ 0.56 (ethyl acetate/CH
3
3
1
6
d
2
B
B
A
A
B
B
A
C
2
2
C
1
C
B
C
ꢀ
t
R
Colorless solid, mp 180 C, yield 224 mg (>99%). C18H17ClFN3O6S

J.A. Schreiber et al. / European Journal of Medicinal Chemistry 158 (2018) 259e269
267
(
457.9). Rf ¼ 0.43 (ethyl acetate þ 1% acetic acid, detection:
2 2
dissolved in CH Cl . The solution was extracted three times with
2 M NaOH. Afterwards, the aqueous layer was acidified with conc.
HCl to adjust pH 1 and the acidic solution was neutralized by
2
0
4
54 nm). 1H NMR (400 MHz, DMSO-D6):
d
[ppm] ¼ 2.40 (s, 4 x
.1H, COCH2CH2CO2H*), 2.41e2.49 (m, 4 x 0.9H, COCH2CH2CO2H),
.11 (d, J ¼ 6.3 Hz, 2 x 0.9H, NHCH2), 4.13 (d, J ¼ 6.5 Hz, 2 x 0.1H,
3
addition of saturated NaHCO solution. The formed precipitate was
NHCH2*), 7.33 (d, J ¼ 8.2 Hz, 2 x 0.9H, 3-HB, 5-HB), 7.35 (d,
J ¼ 8.2 Hz, 2 x 0.1H, 3-HB*, 5-HB*), 7.59 (t, J ¼ 8.9 Hz, 0.1H, 5-HA*),
filtered off, washed with H O and dissolved in THF. After removal of
2
the solvent under reduced pressure, the resulting solid was
ꢀ
7
7
2
.61 (t, J ¼ 8.9 Hz, 0.9H, 5-HA), 7.75e7.82 (m, 3H, 6-HA, 2-HB, 6-HB),
recrystallized from CH
3
OH/H
2
O. Colorless solid, mp 189 C, yield
.86 (dd, J ¼ 6.9/2.4 Hz, 0.1H, 2-HA*), 7.91 (dd, J ¼ 6.8/2.2 Hz, 0.9H,
101 mg (46%). C21 19ClFN
H
3
O
3
S (447.9). R
f
¼ 0.78 (ethyl acetate,
):
[ppm] ¼ 3.97
), 4.09 (q, J ¼ 6.2 Hz, 2H, NHCH Ar ),
5.41 (q, J ¼ 5.8 Hz, 1H, NHNHCH Ar ), 7.24e7.30 (m, 3H, 2-H , 4-H
6-H ), 7.31e7.35 (m, 2H, 3-H , 5-H ), 7.35e7.39 (m, 2H, 2-H
7.60 (t, J ¼ 8.9 Hz, 1H, 5-H ), 7.68e7.72 (m, 2H, 3-H , 5-H
(ddd, J ¼ 8.7/4.5/2.3 Hz,1H, 6-H ), 7.88 (dd, J ¼ 6.9/2.3 Hz,1H, 2-H
.39 (t, J ¼ 6.3 Hz, 1H, SO NHCH
CONHNH). HPLC (Method 1): t
¼ 19.1 min, purity 93.3%.
1
-HA), 8.41 (t, J ¼ 6.3 Hz, 1H, SO2NHCH2), 9.20 (s, 0.1H,
detection: 254 nm). H NMR (600 MHz, DMSO-D
(d, J ¼ 5.3 Hz, 2H, NHNHCH Ar
6
d
NHNHCOCH2*), 9.90 (d, J ¼ 1.5 Hz, 0.9H, NHNHCOCH2), 10.27 (d,
2
C
2
B
J ¼ 1.5 Hz, 0.9H, ArBCONHNH), 10.54 (s, 0.1H, ArBCONHNH*), 12.13
2
C
C
C
,
(
s, 1H, CH2CO2H). Ratio of rotamers 90:10; signals of the minor
C
C
C
B
, 6-H
), 7.78
B
),
rotamer are marked with asterisks. HPLC (Method 1):
tR ¼ 15.8 min, purity 95.8%.
A
B
B
A
A
),
8
2
2
), 10.01 (d, J ¼ 5.8 Hz, 1H, Ar
B-
4
.2.6. 3-Chloro-4-fluoro-N-[4-(5-phenyl-1,3,4-oxadiazol-2-yl)
R
benzyl]benzenesulfonamide (11)
A mixture of 2 (116 mg, 0.25 mmol) and SOCl
2
(2.5 mL) was
4.2.9. 4-{[(3-Bromophenyl)sulfonamido]methyl}benzoic acid (14)
heated to reflux for 1 h. After complete conversion monitored by
TLC, the reaction mixture was poured on ice (50 mL). The resulting
4-(Aminomethyl)benzoic acid (6, 592 mg, 3.92 mmol) and
Na2B4O7 (4.00 g, 19.9 mmol) were suspended in H2O (70 mL)
solid was filtered off, washed with H
2
O and dissolved in THF. After
under ultra-sonication. 3-Bromobenzenesulfonyl chloride (563 mL,
removal of the solvent under reduced pressure, the formed solid
was further purified by flash column chromatography (ethyl ace-
tate/cyclohexane 1:3 (350 mL) then CH Cl /THF 4:1 (300 mL),
2 2
3.91 mmol) was added and the reaction mixture was treated with
ultra-sonication for 5 min. After vigorous stirring at rt overnight,
conc. HCl was added slowly to adjust pH 1 and the formed pre-
cipitate was filtered off, washed with H2O and dissolved in THF. The
solvent was removed under reduced pressure. Colorless solid, mp
Ø ¼ 3 cm, h ¼ 10 cm, V ¼ 20 mL, R
f
¼ 0.51 (ethyl acetate/cyclo-
ꢀ
hexane 1:1)). Colorless solid, mp 195 C, yield 102 mg (92%).
ꢀ
C
21
H
15ClFN
detection: 254 nm, blue fluorescence). H NMR (400 MHz, DMSO-
): ), 7.45e7.50 (m, 2H, 2-
[ppm] ¼ 4.18 (d, J ¼ 5.3 Hz, 2H, NHCH
, 6-H ), 7.59e7.70 (m, 4H, 5-H , 3-H , 4-H , 5-H ), 7.80 (ddd,
), 7.88 (dd, J ¼ 6.9/2.3 Hz, 1H, 2-H ),
), 8.11e8.17 (m, 2H, 2-H , 6-H ), 8.49 (t,
). HPLC (Method 1): t
¼ 22.3 min, purity
3
O
3
S (443.9). R
f
¼ 0.51 (ethyl acetate/cyclohexane 1:1,
229 C, yield 1.45 g (>99%). C14H12BrNO4S (370.2). Rf ¼ 0.51 (ethyl
1
acetate/cyclohexane 1:1 þ 1% acetic acid, detection: 254 nm). 1H
D
H
6
d
2
NMR (400 MHz, DMSO-D6):
d
[ppm] ¼ 4.12 (d, J ¼ 6.2 Hz, 2H,
B
B
A
C
C
C
NHCH2), 7.31e7.36 (m, 2H, 3-HB, 5-HB), 7.52 (t, J ¼ 7.9 Hz, 1H, 5-
HA), 7.77 (dt, J ¼ 8.0/1.3 Hz, 1H, 6-HA), 7.79e7.85 (m, 4H, 2-HA, 4-
HA, 2-HB, 6-HB), 8.42 (t, J ¼ 6.3 Hz, 1H, SO2NHCH2), 12.88 (br s,
1H, CO2H). HPLC (Method 1): tR ¼ 19.1 min, purity 99.6%.
J ¼ 8.7/4.5/2.3 Hz, 1H, 6-H
A
A
8
.01e8.06 (m, 2H, 3-H
B
, 5-H
B
C
C
J ¼ 5.6 Hz, 1H, SO
2
NHCH
2
R
9
5.8%.
4.2.10. tert-Butyl 2-(4-{[(3-bromophenyl)sulfonamido]methyl}
4.2.7. 3-Chloro-4-fluoro-N-(4-{[2-(phenylsulfonyl)hydrazino]
benzoyl)hydrazine-1-carboxylate (15)
carbonyl}benzyl)benzenesulfonamide (12)
(176 mg, 0.49 mmol) was dissolved in CH2Cl2 (8 mL) and
pyridine (2 mL). A solution of benzenesulfonyl chloride (65 L,
.51 mmol) in CH2Cl2 (2 mL) was added dropwise and the resulting
14 (1.43 g, 3.86 mmol), tert-butyl carbazate (511 mg, 3.87 mmol)
and COMU (1.82 g, 4.25 mmol) were suspended in THF (30 mL).
After addition of DIPEA (1.35 mL, 7.75 mmol), the resulting yellow
9
m
0
solution was stirred at rt overnight. After addition of H
the organic solvent was removed under reduced pressure and the
formed precipitate was filtered off, washed with H O and CH Cl
2
O (30 mL),
solution was heated to reflux for 3 h. The reaction mixture was
cooled to rt and was extracted with 1 M NaOH (3 ꢃ 15 mL). The
aqueous extracts showed a deep red color. The combined aqueous
layer was washed with CH2Cl2. Afterwards, the aqueous layer was
acidified with conc. HCl to adjust pH 1 and the acidic solution was
neutralized by addition of saturated NaHCO3 solution. The formed
precipitate was filtered off, washed with H2O and dissolved in
CH3OH. The solvent was removed under reduced pressure. Pale
2
2
2
and dissolved in THF. The solvent was removed under reduced
ꢀ
pressure. Colorless solid, mp 159 C, yield 1.51 g (81%).
C
19
H
22BrN
detection: 254 nm). H NMR (400 MHz, DMSO-D
(s, 9H, (H C) ), 7.30e7.38 (m, 2H, 3-
C), 4.10 (d, J ¼ 6.2 Hz, 2H, NHCH
, 5-H ), 7.71e7.81 (m, 3H, 6-H , 2-H
), 7.54 (t, J ¼ 7.9 Hz, 1H, 5-H
6-H ), 7.90 (t, J ¼ 1.8 Hz, 1H, 2-H
), 7.83 (dt, J ¼ 8.0/1.4 Hz, 1H, 4-H
NHCH ), 8.88 (br s, 1H, NHNHCO ), 10.14
CONHNH). HPLC (Method 1): t
¼ 20.1 min,
3
O
5
S (484.4). R
f
¼ 0.38 (ethyl acetate/cyclohexane 1:1,
1
6
):
d
[ppm] ¼ 1.43
3
3
2
H
B
B
A
A
B
,
ꢀ
yellow solid, mp 190 C, yield 208 mg (85%). C20H17ClFN3O5S2
B
A
A
),
(
2
497.9). Rf ¼ 0.47 (ethyl acetate/cyclohexane 1:1, detection:
54 nm). 1H NMR (400 MHz, DMSO-D6):
[ppm] ¼ 4.08 (s, 2H,
NHCH2), 7.28 (d, J ¼ 7.9 Hz, 2H, 2-HB, 6-HB), 7.52 (t, J ¼ 7.6 Hz, 2H,
-HC, 5-HC), 7.56e7.64 (m, 4H, 5-HA, 2-HC, 4-HC, 6-HC), 7.78 (ddd,
J ¼ 8.6/4.5/2.3 Hz, 1H, 6-HA), 7.83 (d, J ¼ 7.7 Hz, 2H, 3-HB, 5-HB),
8.39 (t, J ¼ 6.3 Hz, 1H, SO
2
2
2
d
(d, J ¼ 1.4 Hz, 1H, Ar
B
R
purity 99.5%.
3
4.2.11. 3-Bromo-N-[(4-hydrazinocarbonyl)benzyl]
benzenesulfonamide hydrochloride (16)
7
(
(
.89 (dd, J ¼ 6.9/2.3 Hz, 1H, 2-HA), 8.42 (br s, 1H, SO2NHCH2), 10.00
br s, 1H, NHNHSO2ArC), 10.61 (br s, 1H, ArBCONHNH). HPLC
Method 1): tR ¼ 20.1 min, purity 96.9%.
4 M HCl in 1,4-dioxane (1.55 mL, 6.20 mmol) was added to a
solution of 15 (1.43 g, 2.95 mmol) in 1,4-dioxane (12 mL). The
resulting reaction mixture was stirred at rt overnight. The formed
suspension was dissolved by addition of CH3OH and the solvent
was removed under reduced pressure. The solid residue was dis-
solved in CH3OH (20 mL) and the solution was slowly added to
CH2Cl2 (250 mL), leading to the formation of a precipitate. The solid
was filtered off, washed with CH2Cl2 and dissolved in CH3OH. The
solvent was removed under reduced pressure. Pale yellow solid, mp
4
.2.8. N-{4-[(2-Benzylhydrazino)carbonyl]benzyl}-3-chloro-4-
fluorobenzenesulfonamide (13)
(169 mg, 0.49 mmol), benzylhydrazine hydrochloride (80 mg,
.50 mmol) and COMU (250 mg, 0.58 mmol) were suspended in
THF (10 mL). After addition of DIPEA (300 L, 1.72 mmol), the
7
0
m
resulting yellow solution was stirred at rt for 3 d. The solvent was
removed under reduced pressure and the resulting residue was
ꢀ
218 C, yield 1.21 g (97%). C14H15ClBrN3O3S (420.7). Rf ¼ 0.21

268
J.A. Schreiber et al. / European Journal of Medicinal Chemistry 158 (2018) 259e269
(
ethyl acetate þ 1% NH3 solution (25%)), detection: 254 nm). 1H
for DMSO. Therefore, for the selectivity testing the agonist solution
was supplemented with 0.1% DMSO to avoid an inhibition by
NMR (400 MHz, DMSO-D6):
d
[ppm] ¼ 4.12 (d, J ¼ 6.3 Hz, 2H,
NHCH2), 7.37e7.42 (m, 2H, 2-HB, 6-HB), 7.53 (t, J ¼ 7.9 Hz, 1H, 5-
HA), 7.80 (ddd, J ¼ 7.9/1.9/1.2 Hz, 1H, 6-HA), 7.81e7.86 (m, 3H, 4-
HA, 3-HB, 5-HB), 7.88 (t, J ¼ 1.8 Hz, 1H, 2-HA), 8.50 (t, J ¼ 6.3 Hz,
DMSO. Recordings were performed at
of ꢁ70 mV. Recording pipettes were backfilled with 3 M KCl and
had resistances in a range of 0.5e1.5 M
a holding potential
U
.
1
H, SO2NHCH2), 10.28 (br s, 3H, CONHNH3), 11.52 (s, 1H, ArB-
CONHNH3). HPLC (Method 1): tR ¼ 15.6 min, purity 99.5%.
5.3. Data analysis and graphical illustration
4
.2.12. 3-Bromo-N-[4-({2-[(thiophen-2-yl)carbonyl]hydrazino}
Data were analyzed with Ana (Dr. Michael Pusch, Genova, Italy)
and OriginPro 2016 (OriginLab Corporation, Northampton, USA).
Inhibition was calculated by following equation:
carbonyl)benzyl]benzenesulfonamide (17)
6 (84.1 mg, 0.20 mmol), thiophene-2-carboxylic acid (25.6 mg,
.20 mmol) and COMU (94.2 mg, 0.22 mmol) were suspended in
THF (5 mL). After addition of DIPEA (100 L, 0.57 mmol), the
1
0
I
c
ꢁ I
m
h
Inhibtion ¼ 1 ꢁ I
ꢁ I
resulting yellow solution was stirred at rt overnight. After addition
of H2O (20 mL), the organic solvent was removed under reduced
pressure and the formed precipitate was filtered off, washed with
ethyl acetate and H2O and dissolved in THF. The solvent was
removed under reduced pressure. Colorless solid, mp 158 C, yield
6
a
h
I
h
is the holding current before adding agonist solution; I
a
is the
current after adding of agonist solution; I is the resulting current in
c
presence of compound solution. To compare the activity of com-
pounds, the average inhibition of the compound was normalized to
the average inhibition of 2 in the same experimental measurement.
ꢀ
6.0 mg (67%). C19H16BrN3O4S2 (494.4). Rf ¼ 0.82 (ethyl acetate,
detection: 254 nm). 1H NMR (600 MHz, DMSO-D6):
d
[ppm] ¼ 4.12
Inorm is determined as normalized Inhibition.
(
(
(
d, J ¼ 6.2 Hz, 2H, NHCH2), 7.22 (dd, J ¼ 5.0/3.7 Hz, 1H, 4-HC), 7.37
d, J ¼ 8.2 Hz, 2H, 2-HB, 6-HB), 7.54 (t, J ¼ 7.9 Hz, 1H, 5-HA), 7.80
ddd, J ¼ 7.9/1.7/0.9 Hz, 1H, 6-HA), 7.81e7.85 (m, 3H, 4-HA, 3-HB, 5-
Inhibition500 nM compound
I
norm
¼
Inhibition500 nM 3
HB), 7.86 (dd, J ¼ 5.0/1.1 Hz,1H, 5-HC), 7.89 (dd, J ¼ 3.7/1.2 Hz,1H, 3-
HC), 7.90 (t, J ¼ 1.9 Hz,1H, 2-HA), 8.42 (t, J ¼ 6.3 Hz,1H, SO2NHCH2),
Dose response curves of 10c, 10h and 17 were fitted to logistic
sigmoid equitation:
1
0.47 (s, 1H, ArBCONHNH), 10.51 (s, 1H, NHNHCOArC). HPLC
(
Method 2): tR ¼ 18.2 min, purity 98.5%.
A1 ꢁ A2
y ¼ 1 ꢀ ꢁ þ A2
p
5
. Electrophysiological evaluation
x
þ
x
0
5
.1. Molecular biology
A1 and A2 are defined as the minimal and maximal inhibition of
compound and were determined as A1 ¼0% and A2 ¼ 100%; p is
Preparation of the needed cRNAs and oocytes were done like
defined as slope of the curve; x is the concentration of half-
0
previously descripted [31]. Defoliculated oocytes were commer-
cially obtained from EcoCyte Bioscience (Castrop-Rauxel, Germany)
and were injected with 0.8 ng GluN1a cRNA and 0.8 ng GluN2A
cRNA using nanoliter injector 2000 (WPI, Berlin, Germany). For
selectivity testing 0.8 ng of GluN2A cRNA was replaced by 0.8 ng
GluN2B cRNA. cRNA was generated by in vitro transcription with
Ambion T7 Message Machine kit (Life Technologies, Darmstadt,
Germany) from linearized cDNA templates (wildtype rat, subcloned
into pSGEM vector, linearized with restriction enzyme PacI).
maximal inhibition and x is the tested concentration.
Inhibition data of 200 nM 4, 10c, 10h and 17 were fitted linear by
following equation:
y ¼ a þ b*x
a is defined as intercept, while b gives the slope of the linear fit.
Significance of Inorm and glycine dependency data was tested for all
screenings respectively all concentrations by One-way-ANOVA and
post hoc mean comparison Tukey Test. Complete data of One-way-
ANOVA for Inorm and glycine dependency are given in the sup-
porting information.
ꢀ
Injected oocytes were stored for 3e5 d at 18 C in Bath's solution
containing (mmol/L): 88 NaCl, 1 KCl, 0.4 CaCl
2 3 2
, 0.33 Ca(NO ) , 0.6
MgSO , 5 Tris-HCl, 2.4 NaHCO and supplemented with 80 mg/L
4
3
Figs. 1 and 3 were illustrated using YASARA structure 17.12.24
theophylline, 63 mg/L penicillin, 40 mg/L streptomycin and 100 mg/
L gentamycin.
[34]. and previous published cryo electron microscopy structure of
triheteromeric GluN1-GluN2A-GluN2B NMDAR (PDB ID 5UOW;
[
(
16]) and x-ray crystal structure of 2 bound to GluN1-GluN2A LBD
PDB ID 5I56; [17]). Alignment of both PDB files was generated
5.2. Two electrode voltage clamp (TEVC)
using MUSTANG algorithm, which is implemented in YASARA [35].
Functional activity was tested by two electrode voltage clamp
(
TEVC) in oocytes at room temperature like previously descripted.
2þ
For measurements oocytes were constantly perfused with Ba
-
Acknowledgements
Ringer containing (mmol/L): 10 HEPES, 90 NaCl, 1 KCl, 1.5 BaCl
was adjusted with 1 M NaOH to 7.4. Agonist solution containing
M glycine and 10 M (S)-glutamate was freshly prepared from
00 mM stock solutions in Ba -Ringer for each measurement.
2
. pH
This work was supported by the Deutsche For-
schungsgemeinschaft which is gratefully acknowledged. Moreover,
we are grateful to Cells-in-Motion (CiM) Cluster of Excellence for
supporting this project.
1
1
0
m
m
2
þ
Solutions of different test compounds were prepared from 10 mM
DMSO stock solutions by dilution with agonist solution to final
concentrations of 1 nMe10
pounds were diluted to final concentrations of 500 nM. For testing
solutions of 1 nM to 1 M the resulting DMSO concentration in the
test solution had no influence on the assay. At compound concen-
trations of 10 M, which caused a DMSO concentration of 0.1%, a
weak inhibition of GluN2B containing NMDARs could be detected
m
M. For the screening of activity, com-
Abbreviations
m
NMDAR N-methyl-D-aspartate receptor
CNS
ATD
LBD
central nervous system
amino terminal domain
ligand binding domain
m

J.A. Schreiber et al. / European Journal of Medicinal Chemistry 158 (2018) 259e269
269
TMD
CTD
transmembrane domain
carboxy terminal domain
proteins : physiological role of GluN2B subunits, J. Neurophysiol. 109 (2013)
391e1402, https://doi.org/10.1152/jn.00176.2012.
16] W. Lü, J. Du, A. Goehring, E. Gouaux, Cryo-EM structures of the triheteromeric
NMDA receptor and its allosteric modulation, Science 80 (2017) 355, https://
doi.org/10.1126/science.aal3729.
17] F. Yi, T.C. Mou, K.N. Dorsett, R.A. Volkmann, F.S. Menniti, S.R. Sprang,
K.B. Hansen, Structural basis for negative allosteric modulation of GluN2A-
containing NMDA receptors, Neuron 91 (2016) 1316e1329, https://doi.org/
10.1016/j.neuron.2016.08.014.
1
[
[
®
COMU ¼ (1-Cyano-1-ethoxycarbonylmethylenaminoxy)-
dimethylamino-morpholino-carbenium
hexafluorophosphate
DIPEA
NAM
SD
N,N-diisopropylethylamine
negative allosteric modulator
standard deviation
[
[
[
[
18] C.H. Lee, W. Lü, J.C. Michel, A. Goehring, J. Du, X. Song, E. Gouaux, NMDA
receptor structures reveal subunit arrangement and pore architecture, Nature
SEM
standard error of the mean
511 (2014) 191e197, https://doi.org/10.1038/nature13548.
19] E. Karakas, N. Simorowski, H. Furukawa, Structure of the zinc-bound amino-
terminal domain of the NMDA receptor NR2B subunit, EMBO J. 28 (2009)
Appendix A. Supplementary data
3910e3920, https://doi.org/10.1038/emboj.2009.338.
20] H. Furukawa, E. Gouaux, Mechanisms of activation, inhibition and specificity :
Supporting Information is available free of charge via the
Internet at http://pubs.acs.org and includes physical, spectroscopic
and purity data of all compounds and synthetic methods.
Supplementary data related to this article can be found at
https://doi.org/10.1016/j.ejmech.2018.09.006.
crystal structures of the NMDA receptor NR1 ligand-binding core, EMBO J. 22
(2003) 2873e2885.
21] T. Kuner, P.H. Seeburg, H.R. Guy, A common architecture for Kþ channels and
ionotropic glutamate receptors? Trends Neurosci. 26 (2003) 27e32, https://
doi.org/10.1016/S0166-2236(02)00010-3.
[22] G. Kohr, P.H. Seeburg, Subtype-specific regulation of recombinant NMDA
receptor-channels by protein tyrosine kinases of the src family, J. Physiol. 492
(
1996) 445e452.
References
[
23] H. Monyer, N. Burnashev, D.J. Laurie, B. Sakmann, P.H. Seeburg, Develop-
mental and regional expression in the rat brain and functional properties of
four NMDA receptors, Neuron 12 (1994) 529e540.
[24] D. Inta, M.A. Vogt, N. Pfeiffer, G. K o€ hr, P. Gass, Dichotomy in the anxiolytic
versus antidepressant effect of C-terminal truncation of the GluN2A subunit of
NMDA receptors, Behav. Brain Res. 247 (2013) 227e231, https://doi.org/
10.1016/j.bbr.2013.03.036.
[
1] R.G.M. Morris, E. Anderson, G.S. Lynch, M. Baudry, Selective impairment of
learning and blockade of long-term potentiation by an N-methyl-D-aspartate
receptor antagonist, AP5, Nature 319 (1986) 774e776, https://doi.org/
10.1038/319774a0.
[
[
[
[
[
[
[
[
2] Y.T. Wang, D. Bieger, R.S. Neuman, Activation of NMDA receptors is necessary
for fast information transfer at brainstem vagal motoneurons, Brain Res. 567
[25] R. Pinacho, N. Villalmanzo, M. Roca, R. Iniesta, A. Monje, J.M. Haro, J.J. Meana,
I. Ferrer, G. Gill, B. Ramos, Analysis of Sp transcription factors in the post-
mortem brain of chronic schizophrenia: a pilot study of relationship to
negative symptoms, J. Psychiatr. Res. 47 (2013) 926e934, https://doi.org/
10.1016/j.jpsychires.2013.03.004.
[26] Y. Sun, X. Cheng, J. Hu, Z. Gao, The role of GluN2A in cerebral ischemia:
promoting neuron death and survival in the early stage and thereafter, Mol.
Neurobiol. 55 (2017) 1208e1216, https://doi.org/10.1007/s12035-017-0395-
8.
[27] E. Bettini, A. Sava, C. Griffante, C. Carignani, A. Buson, A.M. Capelli, M. Negri,
F. Andreetta, S. a Senar-sancho, L. Guiral, F. Cardullo, Identification and
characterization of novel NMDA receptor antagonists selective for NR2A- over
NR2B-containing receptors, J. Pharmacol. Exp. Therapeut. 335 (2010)
636e644, https://doi.org/10.1124/jpet.110.172544.
[28] K.B. Hansen, K.K. Ogden, S.F. Traynelis, Subunit-selective allosteric inhibition
of Glycine binding to NMDA receptors, J. Neurosci. 32 (2012) 6197e6208,
https://doi.org/10.1523/JNEUROSCI.5757-11.2012.
(
1991) 260e266, https://doi.org/10.1016/0006-8993(91)90804-5.
3] E.W. Harris, A.H. Ganong, C.W. Cotman, Long-term potentiation in the hip-
pocampus involves activation of N-methyl-D-aspartate receptors, Brain Res.
Bull. 323 (1984) 132e137.
4] Z.I. Bashir, S. Alford, S.N. Davies, A.D. Randall, G.L. Collingridge, Long-term
potentiation of NMDA receptor-mediated synaptic transmission in the hip-
pocampus, Nature 349 (1991) 156e158, https://doi.org/10.1038/349156a0.
5] J.W. Johnson, P. Ascher, Glycine potentiates the NMDA response in cultured
mouse brain neurons, Nature 325 (1987) 529e531, https://doi.org/10.1038/
325529a0.
6] M.L. Mayer, G.L. Westbrook, P.B. Guthrie, Voltage-dependent block by Mg2þ
of NMDA responses in spinal cord neurones, Nature 309 (1984) 261e263,
https://doi.org/10.1038/309261a0.
7] K. Tsuzuki, S. Mochizuki, M. Iino, H. Mori, M. Mishina, S. Ozawa, Ion perme-
ation properties of the cloned mouse ε2/z1 NMDA receptor channel, Mol.
Brain Res. 26 (1994) 37e46, https://doi.org/10.1016/0169-328X(94)90071-X.
8] M. Iino, S. Ciani, K. Tsuzuki, S. Ozawa, Y. Kidokoro, Permeation Properties of
Naþ and Ca2þ Ions through the Mouse e2/z1 NMDA Receptor Channel
Expressed in Xenopus Oocytes, J. Membr. Biol. 155 (1997) 143e156.
9] A. Frandsen, J. Drejer, A. Schousboe, Direct evidence that excitotoxicity in
cultured neurons is mediated via N-Methyl-D-Aspartate (NMDA) as well as
non-NMDA receptors, J. Neurochem. 53 (1989) 297e299, https://doi.org/
[29] D.H. Hackos, P.J. Lupardus, T. Grand, Y. Chen, T.M. Wang, P. Reynen,
A. Gustafson, H.J.A. Wallweber, M. Volgraf, B.D. Sellers, J.B. Schwarz, P. Paoletti,
M. Sheng, Q. Zhou, J.E. Hanson, Positive allosteric modulators of GluN2A-
containing NMDARs with distinct modes of action and impacts on circuit
function,
Neuron
89
(2016)
983e999,
https://doi.org/10.1016/
j.neuron.2016.01.016.
1
0.1111/j.1471-4159.1989.tb07327.x.
[30] R.A. Volkmann, C.M. Fanger, D.R. Anderson, V.R. Sirivolu, K. Paschetto,
E. Gordon, C. Virginio, M. Gleyzes, B. Buisson, E. Steidl, S.B. Mierau,
M. Fagiolini, F.S. Menniti, MPX-004 and MPX-007: new pharmacological tools
to study the physiology of nmda receptors containing the GluN2A Subunit,
PLoS One 11 (2016) 1e20, https://doi.org/10.1371/journal.pone.0148129.
[31] S.L. Müller, J.A. Schreiber, D. Schepmann, N. Strutz-Seebohm, G. Seebohm,
B. Wünsch, Systematic variation of the benzenesulfonamide part of the
GluN2A selective NMDA receptor antagonist TCN-201, Eur. J. Med. Chem. 129
(2017) 124e134, https://doi.org/10.1016/j.ejmech.2017.02.018.
[32] G.A. Patani, E.J. LaVoie, Bioisosterism: a rational approach in drug design,
Chem. Rev. 96 (1996) 3147e3176, https://doi.org/10.1021/cr950066q.
[33] K.E. Horner, P.B. Karadakov, Chemical Bonding and Aromaticity in Furan,
Pyrrole, and Thiophene : a Magnetic Shielding Study, 2013, https://doi.org/
10.1021/jo401319k.
[
10] D.J.A. Wyllie, M.R. Livesey, G.E. Hardingham, Influence of GluN2 subunit
identity on NMDA receptor function, Neuropharmacology 74 (2013) 4e17,
https://doi.org/10.1016/j.neuropharm.2013.01.016.
11] N.J. Sucher, S. Akbarian, C.L. Chi, C.L. Leclerc, M. Awobuluyi, D.L. Deitcher,
M.K. Wu, J.P. Yuan, E.G. Jones, S.A. Lipton, Developmental and regional
expression pattern of a novel NMDA receptor-like subunit ( NMDAR-L ) in the
rodent brain, J. Neurosci. 15 (1995) 6509e6520.
12] M. Hollmann, J. Boulter, C. Maron, L. Beasley, J. Sullivan, G. Pecht,
S. Heinemann, Zinc potentiates agonist-induced currents at certain splice
variants of the NMDA receptor, Neuron 10 (1993) 943e954, https://doi.org/
[
[
1
0.1016/0896-6273(93)90209-A.
[
13] D.J. Laurie, P.H. Seeburg, Ligand affinities at recombinant N-methyl-D-aspar-
tate receptors depend on subunit composition, Eur. J. Pharmacol. Mol. Phar-
macol. 268 (1994) 335e345, https://doi.org/10.1016/0922-4106(94)90058-2.
14] T. Schüler, I. Mesic, C. Madry, I. Bartholom a€ us, B. Laube, Formation of NR1/NR2
[34] E. Krieger, G. Vriend, YASARA View - molecular graphics for all devices - from
smartphones to workstations, Bioinformatics 30 (2014) 2981e2982, https://
doi.org/10.1093/bioinformatics/btu426.
[
and NR1/NR3 heterodimers constitutes the initial step in N-methyl-D-aspar-
tate receptor assembly, J. Biol. Chem. 283 (2008) 37e46, https://doi.org/
[35] A.S. Konagurthu, J.C. Whisstock, P.J. Stuckey, A.M. Lesk, MUSTANG: a multiple
structural alignment algorithm, Proteins Struct. Funct. Bioinforma. 64 (2006)
559e574, https://doi.org/10.1002/prot.20921.
10.1074/jbc.M703539200.
[
15] A.J. Delaney, P.L. Sedlak, E. Autuori, J.M. Power, P. Sah, Synaptic NMDA re-
ceptors in basolateral amygdala principal neurons are triheteromeric