Synthesis and in Vitro Characterisation of Ifenprodil-Based Fluorescein Conjugates as GluN1/GluN2B N-Methyl-D-aspartate Receptor Antagonists

Article abstract of DOI:10.1002/cbic.201200675

GluN2B-containing NMDA receptors are involved in many important physiological functions and play a pivotal role in mediating pain as well as in several neurodegenerative disorders. We aimed to develop fluorescent probes to target the GluN2B subunit selectively in order to allow better understanding of the relationships between receptor localisation and physiological importance. Ifenprodil, known as the GluNR2B antagonist of reference, was chosen as the template for the elaboration of probes. We had previously reported a fluorescein conjugate that was shown (by confocal microscopy imaging of DS-red-labelled cortical neurons) to bind specifically to GluN2B. To elaborate this probe, we explored the influence of both the nature and the attachment point of the spacer between the fluorophore and the parent compound, ifenprodil. We performed chemical modifications of ifenprodil at the benzylic position and on the phenol ring by introducing secondary amine or amide functions and evaluated alkyl chains from two to 20 bonds either including or not including secondary amide functions as spacers. The previously developed probe was found to display the greatest activity in the inhibition of NMDA-induced Ca²⁺ influx by calcium imaging experiments on HEK293 cells transfected with the cDNA encoding for GluN1-1A and GluN2B. Further investigations revealed that this probe had a neuroprotective effect equivalent to that of ifenprodil in a standard test for neurotoxicity. Despite effects of lesser amplitude with these probes relative to ifenprodil, we demonstrated that they displaced [³H]ifenprodil in mouse brain slices in a similar manner.

Full text of DOI:10.1002/cbic.201200675

CHEMBIOCHEM
FULL PAPERS
DOI: 10.1002/cbic.201200675
Synthesis and in Vitro Characterisation of Ifenprodil-Based
Fluorescein Conjugates as GluN1/GluN2B N-Methyl-d-
aspartate Receptor Antagonists
Martine Dhilly,[a, b, c] Javier Becerril-Ortega,[b] Nathalie Colloc’h,[b, d] Eric T. MacKenzie,[b, d]
Louisa Barrꢀ,[a, b, c] Alain Buisson,[b] Olivier Nicole,[b] and Cꢀcile Perrio*[a, b, c]
GluN2B-containing NMDA receptors are involved in many im-
portant physiological functions and play a pivotal role in medi-
ating pain as well as in several neurodegenerative disorders.
We aimed to develop fluorescent probes to target the GluN2B
subunit selectively in order to allow better understanding of
the relationships between receptor localisation and physiologi-
cal importance. Ifenprodil, known as the GluNR2B antagonist
of reference, was chosen as the template for the elaboration of
probes. We had previously reported a fluorescein conjugate
that was shown (by confocal microscopy imaging of DS-red-la-
belled cortical neurons) to bind specifically to GluN2B. To elab-
orate this probe, we explored the influence of both the nature
and the attachment point of the spacer between the fluoro-
phore and the parent compound, ifenprodil. We performed
chemical modifications of ifenprodil at the benzylic position
and on the phenol ring by introducing secondary amine or
amide functions and evaluated alkyl chains from two to
20 bonds either including or not including secondary amide
functions as spacers. The previously developed probe was
found to display the greatest activity in the inhibition of
NMDA-induced Ca2+ influx by calcium imaging experiments
on HEK293 cells transfected with the cDNA encoding for
GluN1-1A and GluN2B. Further investigations revealed that this
probe had a neuroprotective effect equivalent to that of ifen-
prodil in a standard test for neurotoxicity. Despite effects of
lesser amplitude with these probes relative to ifenprodil, we
demonstrated that they displaced [3H]ifenprodil in mouse
brain slices in a similar manner.
Introduction
The N-methyl-d-aspartate receptor (NMDAR) is a major iono-
tropic glutamate receptor widely expressed in the central nerv-
ous system.[1] NMDARs are crucial under both physiological
and pathological conditions. Activity-dependent Ca2+ flux
through NMDARs is thought to trigger downstream intracellu-
lar signalling events that initiate long-term changes in synaptic
transmission and hence contribute to neuronal development,
behavioural learning and information storage.[2] Overstimula-
tion of NMDARs results in activation of proteases, lipases and
DNAases, mitochondrial energy failure, and/or signalling to
a variety of apoptotic proteins and is thus implicated in neuro-
nal death.[3] NMDARs are therefore ideal candidates in the ini-
tiation of early neuronal dysfunction and learning deficits, as
well as later neuronal loss in several neurodegenerative condi-
tions such as Alzheimer’s and Huntington’s diseases.[4]
Functional NMDAR channels are hetero-tetrameric assem-
blies, each consisting of two GluN1 subunits plus two GluN2 or
GluN3 subunits.[5] The GluN1 subunit contains a glycine-bind-
ing site and is essential for the formation of functional NMDAR
channels, whereas the GluN2A-2D subunit provides a gluta-
mate-binding site and modifies channel properties such as cur-
rent kinetics and channel conductance, as well as their general
pharmacology. The expression of different NMDAR subtypes is
not static, but regulated dynamically in a cell- and synapse-
specific manner within the central nervous system.[6] In addi-
tion to the composition of the subunit, the subcellular localisa-
tion of NMDARs determines whether the resulting signal trans-
duction cascade promotes neuronal plasticity (as in learning)
and neuronal survival or death. Whereas synaptic plasticity and
cell survival are related to the activation of synaptic NMDARs,
stimulation of extrasynaptic NMDARs, which are mainly com-
posed of GluN2B subunits, leads to the promotion of pro-
death gene expression, mitochondrial calcium overload and
activation of cell death signalling pathways.[7]
[a] M. Dhilly, Dr. L. Barrꢀ, Dr. C. Perrio
CNRS, UMR 6301 ISTCT, LDM-TEP, GIP Cyceron
Boulevard Henri Becquerel, BP5229
14074 Caen Cedex (France)
E-mail: perrio@cyceron.fr
[b] M. Dhilly, J. Becerril-Ortega, Dr. N. Colloc’h, Dr. E. T. MacKenzie, Dr. L. Barrꢀ,
Prof. A. Buisson, Dr. O. Nicole, Dr. C. Perrio
Universitꢀ de Caen Basse-Normandie, GIP Cyceron
14074 Caen Cedex (France)
[c] M. Dhilly, Dr. L. Barrꢀ, Dr. C. Perrio
CEA, DSV/I2BM, LDM-TEP, GIP Cyceron
14074 Caen Cedex (France)
[d] Dr. N. Colloc’h, Dr. E. T. MacKenzie
CNRS, UMR 6301 ISTCT, CERVOxy, GIP Cyceron
14074 Caen Cedex (France)
Despite the importance of GluN2B in physiological and
pathological processes, remarkably few tools for studying the
expression, subcellular location and trafficking of GluN2B in
the living cell have been developed. Targeting the GluN2B sub-
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cbic.201200675.
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2013, 14, 759 – 769 759

CHEMBIOCHEM
FULL PAPERS
www.chembiochem.org
Scheme 1. Fluorescent probe 1a and GluN2B NMDAR antagonists 2–6.
unit with a fluorescent probe ought to provide an original ap-
proach to understanding of the purpose and involvement of
the NMDAR GluN2B subunit in situations of health and disease.
Previously, we had reported the fluorescein conjugate 1a
(Scheme 1) as the first probe for which specific binding to
GluN2B-containing NMDARs was revealed, by confocal micro-
scopic imaging of DS-red-labelled cortical neurons.[8] Probe 1a
is derived from ifenprodil (2, Scheme 1)—the lead compound
in the major class of high-affinity, noncompetitive, voltage-in-
dependent inhibitors of NMDARs with marked selectivity for
GluN2B.[9,10] This potent antagonist, as well as its analogue
Ro 25-6981 (3b), have been extensively used for in vitro and in
vivo studies of the expression, properties and distribution of
NMDARs.[10] To elaborate probe 1a, we explored the influence
both of the nature of the spacer and of its attachment point
between the parent (ifenprodil) and the fluorophore (fluores-
cein) on the GLUN2 activity. Here we report our entire pro-
gramme leading to the selection of probe 1a, as well as fur-
ther pharmacological characterisation including evaluation of
its binding properties both in
phenol part of the parent ligand. By analogy with the nitro-
gen-containing analogues 4[13] and 5,[14] reported as potent
GluN2B antagonists, as well as the (chloroacetyl)aniline 6[11]
used for cysteine affinity labelling experiments, we chose to
introduce an amine function onto ifenprodil as an anchoring
point for the fluorophore through a spacer. We then devel-
oped fluorescent conjugates from different series—series A, B
and C (Scheme 2)—as target probes varying both in the an-
chorage point and in the length of the spacer attaching the
fluorophore to the parent molecule.
Results and Discussion
Chemistry
The preparation of fluorescein conjugates from series A, B and
C involved, firstly, the synthesis of modified ifenprodil and, sec-
ondly, ligation to the fluorescein. Ifenprodil exists in the form
of two racemic diastereoisomers—(Æ)-threo-ifenprodil (2a,
neuronal cultures and in murine
brain slices.
Ifenprodil (2) incorporates a
benzylpiperidine moiety con-
nected to a phenol unit by
a short linker. Its binding site is
located near the central, inter-
lobe cleft of the N-terminal
domain (NTD) of the GluN2B
subunit. Modelling and muta-
genesis studies have suggested
that the benzyl moiety is orient-
ed towards the NTD hinge
whereas the phenol ring is lo-
cated at the entrance to the
cleft.[11,12] On the basis of these
topological considerations, we
designed fluorescent ifenprodil-
based probes by anchoring flu-
orescein as a fluorophore exclu-
sively on the (hydroxymethyl)-
Scheme 2. Target fluorescent conjugates based on ifenprodil.
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2013, 14, 759 – 769 760

CHEMBIOCHEM
FULL PAPERS
www.chembiochem.org
Scheme 1) and (Æ)-erythro-ifenprodil (2b)—and only 2b is
commonly used in biological studies. Electrophysiological stud-
ies, based on voltage-clamp recording, revealed no difference
in potency between the two diastereoisomers 2a and 2b at
the GluN1-1A/GluN2B subunit combination, whereas (Æ)-threo-
ifenprodil (2a) was found to be about five times more active
than isomer 2b in preventing glutamate-induced death in hip-
pocampal neurons.[9,15] Accordingly, we focussed principally on
the synthesis of racemic compounds with threo stereochemis-
try.
either of N-propylamine or of a N-Boc-a,w-diamine with an
alkylene chain of two, four or six carbon atoms. The resulting
threo-ethylenediamines 9a–12a were isolated in 60–88%
yields. The configuration was retained, as a result of a mecha-
nism with anchimeric assistance due to the in situ formation of
an aziridinium intermediate. Deprotection of compounds 9a–
12a with HBr in AcOH afforded the threo-phenolamines 14a–
17a (50-57% yields).
Similarly, erythro-phenolamines 14b and 17b were success-
fully prepared from O-benzylated erythro-ifenprodil 8b and N-
propylamine or N-Boc-1,6-hexamethylenediamine.
Phenolamine 18a, characterised by an aminodecyl chain,
was obtained in only 17% yield from protected ifenprodil 8a
by a mesylation/amination/deprotection sequence without the
isolation of the corresponding intermediate ethylenediamine
13a as a pure compound. It is
Modification of ifenprodil to provide access to the series A
probes was achieved by the methodology previously reported
for the fluorescent ligand 1a (Scheme 3).[8] O-Benzylated threo-
ifenprodil 8a, prepared from propiophenone 7, was subjected
to a stereospecific mesylation/amination sequence with use
noteworthy that the above
methodology failed with N-Boc-
a,w-diamines containing eight
and 12 carbon atoms.
Monodeprotection of N-Boc-
O-benzyl-phenolamine
12a
with TFA resulted in the forma-
tion of hexylamine 19a in 82%
yield. Elongation of the hexyla-
mine chain as the spacer was
then carried out by coupling
with N-Boc-aminohexanoic acid,
leading to amide 20a in 45%
yield. Debenzylation of 20a by
catalytic hydrogenation afford-
ed phenol 21a in quantitative
yield. Further treatment of 21a
with TFA led to phenolamidoa-
mine 22a in 90% yield.
Probe precursors for series B
were prepared from erythro- or
threo-aminoethylaniline
23a
and 23b and from threo-aniline
26a (Scheme 4). The starting
anilines 23a, 23b and 26a
were obtained by the previously
reported copper-catalysed ami-
nation of the corresponding
threo- and erythro-bromoar-
enes.[16] Acylation of ethylenedi-
amines 23a and 23b with N-
Boc-aminohexanoic acid led to
the corresponding amides 24a
and 24b in 57 and 75% yields,
respectively. Deprotection of
24a and 24b with TFA afforded
ligands 25a and 25b (66 and
81% yields).
Treatment of N-Boc-amino-
hexanoic acid with aniline 26a
occurred with the concomitant
Scheme 3. Synthesis of series A precursors 14–18a, 14–18b and 22a. a) Methanesulfonyl chloride (MsCl), Et3N,
THF, 08C, then H2N(CH2)nX, Et3N, THF, RT; b) HBraq (47%), AcOH, reflux; c) trifluoroacetic acid (TFA), CH2Cl2, 08C;
d) HO2C(CH2)nNHBoc, Et3N, ClCO2Et, CH2Cl2, 08C; e) H2, Pd/C, MeOH, RT.
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2013, 14, 759 – 769 761

CHEMBIOCHEM
FULL PAPERS
www.chembiochem.org
Scheme 4. Synthesis of series B precursors 25a, 25b and 29a. a) BocNH(CH2)5CO2H, ClCO2Et, Et3N, CH2Cl2, 08C; b) TFA, CH2Cl2, 08C or HCl, EtOH, 08C; c) LiOH,
MeOH, THF, H2O, RT.
acylation of the aniline function and of the hydroxy group,
leading to the ester amide product 27a in 55% yield. The
quantitative cleavage of the ester function with LiOH, followed
by the deprotection of the primary amine with TFA, led to the
desired product 29a (90% yield).
The synthesis of precursors of probes from series C is out-
lined in Scheme 5. Hydroxypropiophenone 7 was converted
into iodophenol 30 with KI and
Fluorescent probes 1a and 42a–45a were prepared by con-
jugation of amino precursors 17a, 22a, 25a, 29a and 37a
with activated (fluoresceincarboxamido)hexanoic acid under
standard conditions (Scheme 6).[8] The fluorescein N-hexyl de-
rivative 46 was obtained from N-hexylamine by a similar proce-
dure and served as the control compound in biological investi-
gations. The characterisation of fluorescent products was per-
I2 in aqueous ammonia in 87%
yield. Iodophenol 30 was trans-
formed into threo-iodoarene
34a via the intermediates 31–
33 by a four-step procedure in-
volving benzylation of the
phenol, bromination in the a-
position of the ketone, substitu-
tion with benzylpiperidine and
reduction of the ketone with
NaBH4 (37% total yield). The
copper-catalysed coupling of io-
doarene 34a with aqueous am-
monia or ethylenediamine in
the presence of 2-acetylcyclo-
hexanone and Cs2CO3 in DMF
yielded anilines 35a and 38a in
58 and 60% yields, respectively.
Deprotection of the phenol in
38a with BBr3 gave compound
39a, which was highly unstable.
No other transformations, such
as acylation of the ligand 39b,
were successful. Acylation of
35a and 38a with N-Boc-ami-
nohexanoic acid formed amides
36a and 40a in 47 and 92%
yields, respectively. Catalytic hy-
drogenolysis followed by treat-
ment with TFA led to the final
Scheme 5. Synthesis of series C precursors 37a, 39a and 41a. a) I2, KI, NH4OHaq (28%), RT; b) BnBr, K2CO3, DMF,
608C; c) Br2, CCl4, RT; d) 4-benzylpiperidine, Et3N, EtOH, reflux; e) NaBH4, EtOH, RT; f) NH4OH, CuI, 2-acetylcyclohex-
anone, Cs2CO3, DMF, 908C; g) BocNH(CH2)5CO2H, ClCO2Et, Et3N, CH2Cl2, 08C; h) H2, Pd/C, MeOH, RT; i) TFA, CH2Cl2,
08C, or HCl, EtOH, 08C; j) NH2CH2CH2NH2, CuI, K3PO4, DMF, 908C; k) BBr3, CH2Cl2, À788C.
ligands 37a and 41a in moder-
ate yields of 53 and 34%, re-
spectively.
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2013, 14, 759 – 769 762

CHEMBIOCHEM
FULL PAPERS
www.chembiochem.org
Scheme 6. Synthesis of fluorescent probes 1a and 42a–45a, as well as fluorescein hexylamine conjugate 46. a) 6-[Fluorescein-5(6)-carboxamido]hexanoic
acid N-hydroxysuccinimidyl ester, DIEA, NMP, RT, 18 h.
formed by MS analysis. Fluorescence measurements were in
fied compounds 14a, 14b, 15a–18a, 22a, 23a, 23b, 25a,
25b, 26a, 29a, 37a and 41a and the fluorescent conjugates
1a, 42a–45a and 46 were tested at 10 mm to identify active
NMDAR [Ca2+] influx inhibitors and compared to (Æ)-erythro-
ifenprodil (1b, also studied at 10 mm). Results for each com-
pound tested are expressed as % inhibition of [Ca2+]i influx rel-
ative to that determined in the absence of an antagonist
(Figure 1). The inhibition of Ca2+ influx by ifenprodil (2b) was
85% of the NMDA-induced response. This percentage of inhib-
ition was taken as the reference value to compare the effica-
cies of the newly synthesised ifenprodil derivatives.
accordance with that characteristic for fluorescein (lex/lem
=
540/560 nm).
Evaluation of NMDA-induced calcium influx inhibition
For the in vitro biological evaluation of the compounds as se-
lective GluN2B-containing NMDAR antagonists, we used the
previously reported functional assay based on videomicroscop-
ic monitoring of the intracellular calcium concentrations ([Ca2+
]i) in human embryonic kidney cells (HEK293) transfected with
the cDNA encoding for GluN1-1a and GluN2B subunits.[8] Appli-
cation of NMDA (100 mm for 30 s) in the presence of glycine
(10 mm) to the transfected cell induced a rapid and sustained
increase in intracellular Ca2+ influx. In contrast, co-application
of NMDA (in the presence of glycine) and an antagonist de-
creased [Ca2+]i in a manner depending both on the concentra-
tion and on the affinity of the antagonist. The ifenprodil-modi-
Viewed overall, there were three classes of inhibitors. The
first corresponds to those compounds with inhibitory poten-
tials (80–100%) comparable to that of their parent ifenprodil
(2b). The second group of compounds display moderate inhib-
itory properties (50Æ15%). Finally, inhibition in the last class
did not exceed 30%, which indicates weak affinities of the
tested compounds.
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2013, 14, 759 – 769 763

CHEMBIOCHEM
FULL PAPERS
www.chembiochem.org
Figure 1. Determination of inhibition of NMDA-induced [Ca2+] influx in HEK293 cells transfected with the cDNA encoding for GluN1-1A and GluN2B after ap-
plication of new ifenprodil derivatives (10 mm), relative to the activity of the parent compound ifenprodil (2b). Responses to ligand application were normal-
ised to the mean response obtained by stimulation with NMDA (100 mm) in the presence of glycine, (10 mm) taken as the control experiment. Each bar repre-
sents mean values ÆSEM. from three independent experiments Æs.e.m. (n=30). * indicates significantly different from ifenprodil (2b, p<0.05).
Globally, the transformation at the benzylic position (series A
compounds) maintained the receptor binding better than
a modification on the phenol ring (series B and C compounds).
However, the inhibitory properties of series A ifenprodil-modi-
fied precursors depended strongly on the nature of the substi-
tution on the benzylic amine. The propylamine 14a was found
to be nearly inactive (Ca2+ influx inhibition of 29%), whereas
the alkane-a,w-diamine analogues 15a–17a were found to be
potent inhibitors (>80% inhibition). The length of the alkyli-
dene chain was essential for meaningful inhibition, with C6
(17a: 97%)>C2 (15a: 91%)@C4 (16a: 65%)@C10 (18a:
18%). Compound 22a, corresponding to 17a elongated with
an aminohexanecarboxy residue, and thus bearing a chain
longer than that of the decamethylenediamine 18a, main-
tained a considerable inhibitory potency (90%). It is notable
that 1,6-hexamethylenediamine 17b and, to a lesser extent,
ethylenediamine 15a, as well as compound 22a, were found
to be slightly more potent than ifenprodil (2b). The benzylic
hydroxy group in ifenprodil can thus be advantageously re-
placed by a secondary amine provided that this function bears
a labile proton in an appropriate position. The importance of
another protic function has been explained in previous studies
by analogy with polyamines such as spermine and spermidine,
which are non-competitive NMDAR inhibitors.[8]
37a was found to be more active than its analogue 29a from
series B.
With respect to stereochemical influences, only a few diaste-
reoisomer pairs—14a and 14b, 17a and 17b, 23a and 23b,
and 25a and 25b—were tested. The inhibition data speak
largely in favour of the threo forms (14a: 29% vs. 14b: 6%,
17a: 97% vs. 17b: 85%, 25a: 82% vs. 25b: 56%), except in
the case of anilines 23. In this last case, the difference between
the two isomers 23a and 23b was not significant (23a: 61%
vs. 23b: 64%). These results agree with those observed with
the ifenprodil diastereoisomers.[15]
Finally, the most active ifenprodil-modified ligands were the
threo isomers 17a and 22a from series A, 25a and 29a from
series B and 37a from series C. We performed conjugation to
fluorescein only with these precursors (conjugation was also
attempted with ligand 41a but it led to numerous unidentified
fluorescent compounds, probably due to unselective coupling).
Inhibitory data for the four fluorescent probes 42a–45a were
about 39Æ2%, lower than those for the previously reported
probe 1a (68%).[8] The attachment of the fluorescein moiety, in
all cases, reduced the inhibitory efficacy. The presence of an
extended spacer between ligand and fluorescein in probe 42a
did not result in an improved binding. Although moderate, the
inhibitory potentials of the probes were still significant, be-
cause the fluorescein N-hexyl derivative 46 was essentially
without any inhibitory potential (5%).
Series B anilines 23a and 26a had previously been described
as moderate and potent inhibitors (61% and 81%, respective-
ly).[16] Elongation of aniline 23a with the aminohexanecarboxy
residue resulted in ligand 25a, which displayed an increased
inhibitory activity (82% inhibition). After a similar elongation of
the unsubstituted aniline 26a, leading to the derivative 29a,
the inhibitory potential fell considerably (43%).
Finally, probe 1a remained that of choice for further experi-
ments. The already published full pharmacological characteri-
sation of probe 1a (as well as of its precursor 17a) by dose–re-
sponse studies and electrophysiology confirmed their affinities
and selectivities for the NTD of the GluN2B subunit.[8]
The two series C phenols 37a and 41a were estimated to
be moderate inhibitors (65 and 74%, respectively). Ligand 41a,
with the ethylenediamine pattern, was slightly less active than
its analogue 25a from series B. In contradistinction, compound
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2013, 14, 759 – 769 764

CHEMBIOCHEM
FULL PAPERS
www.chembiochem.org
Ifenprodil binding site structural analysis
nists (10 mm). A lactate dehydrogenase (LDH) assay was then
performed to determine the percentage of neuronal death. Ap-
plication of NMDA to the bath induced 37% neuronal death,
which was attenuated by the co-application of the nonselec-
tive NMDAR blocker MK801 (Figure 3). This result confirmed
Recent crystallographic studies have shown that ifenprodil
binds at the interface between GluN1 and GluN2B and partici-
pates in two polar interactions and many hydrophobic interac-
tions.[17] The oxygen atom of the benzylic hydroxy group is in-
volved in one hydrogen bond with a glutamine residue from
GluN2B, and points toward the surface of the GluN1/GluN2B
dimer (Figure 2A). With this orientation, the introduction of flu-
Figure 2. Ifenprodil in binding site: surfaces of GluN1 and of GluN2B with
ifenprodil buried at the interface. A) View of the oxygen atom at the benzyl-
ic position. B) View of the phenol ring.
Figure 3. Evaluation of neuronal death induced by 24 h exposure to NMDA
(30 mm) in the presence or in the absence of MK801, ifenprodil (2b), precur-
sor 17a and probe 1a (10 mm). Each bar represents a mean percentage
value ÆSEM (n=30).
orescein, by means of a spacer, at this benzylic position—such
as in series A compounds (probes 1a and 42a)—can be seen
to be possible. The differences in inhibition of NMDA-induced
[Ca2+] influx may be explained by the length of the spacers,
the shortest (1a) being likely to bind more tightly at the inter-
face of the dimer.
that the observed neuronal death is due to overactivation of
NMDARs. An assay with ifenpodil (2b) resulted in 21% neuro-
nal death, a weaker effect than that of MK801, which might
indicate that selective blockade of GluN2B-containing NMDARs
can protect partially against excitotoxicity. The two ifenprodil
derivatives 17a and 1a exerted a neuroprotective effect similar
to that of ifenprodil (2b) in this standard test of neurotoxicity
(25 and 20% neuronal death, respectively). Neither chemical
modifications nor conjugation thus materially altered the neu-
roprotective activity of parent ifenprodil (2b).
The oxygen atom of the phenol ring in ifenprodil is involved
in a hydrogen bond with a glutamate residue from GluN2B,
and is buried within a pocket formed at the interface between
the two subunits (Figure 2B). The insertion of a long chain
from this position as in series B fluorescent probes 43a and
44a does not seem favourable for binding in the cleft between
GluN1 and GluN2B in the way ifenprodil does. This observation
correlates with the fact that probes and precursors from ser-
ies B display weaker potencies in the inhibition of NMDA-in-
duced [Ca2+] influx.
Displacement studies of [3H]ifenprodil in mouse brain slices
The carbon atom in the ortho position of the phenol ring is
oriented towards the opposite surface of the GluN1/GluN2B
dimer relative to the benzylic hydroxy function (Figure 2B).
Such a location offers space to anchor the fluorophore, as in
series C probe 45a. However, the peptide bond in 45a intro-
duces rigidity that might explain the lesser inhibition of the
NMDA-induced [Ca2+] influx by probe 45a relative to probe
1a.
In order to assess the subunit specificity and selectivity of
probe 1a and precursor 17a, as well as to obtain information
about the pattern of distribution of the compounds within the
central nervous system, displacement studies of [3H]-ifenprodil
in murine brain slices were undertaken. Competition experi-
ments were carried out in the presence of (R)-(+)-3-PPP (D2
agonist and s1 antagonist), GBR 12909 (dopamine reuptake in-
hibitor and s ligand) and GBR 12935 (norepinephrine and dop-
amine reuptake inhibitor) to avoid binding in all ifenprodil tar-
gets except the GluN2B NMDAR. In the absence of antagonist
(noted as the total condition), autoradiography experiments
showed a heterogeneous distribution of radioactivity in the
brain slices after incubation with [3H]ifenprodil (40 nm) for
90 min at 48C. As previously reported,[19] a preferential accu-
mulation of the radioactivity was observed in regions rich in
GluN2B-containing NMDARs in the following order: hippocam-
pus> caudoputamen> cortex> thalamus (Figure 4). In con-
trast, density of labelling was low in cerebellum and non-exis-
tent in white matter. When the incubation was performed in
Quantification of neuronal death in primary cortical cultures
We pursued the investigation of GluN2B antagonistic proper-
ties of the probe 1a and its highly potent inhibitor precursor
17a by evaluation of their neuroprotective activities in cortical
neuronal cultures exposed to NMDA. Primary cortical neuronal
cultures were used at 14 days in vitro (DIV). At this stage of
development, neurons in culture display mixed expression of
GluN2A and GluN2B.[18] Neuronal cultures were treated for 24 h
with NMDA (15 mm) alone or in the presence of NMDA antago-
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2013, 14, 759 – 769 765

CHEMBIOCHEM
FULL PAPERS
www.chembiochem.org
Figure 4. Autoradiographic binding of [3H]ifenprodil in coronal sections of the mouse brain in the absence of antagonists (“total conditions”) or in the pres-
ence of MK801, fluorescein derivative 46, ifenprodil (2b), Ro 25-6981, precursor 17a and probe 1a.
the presence of any selective GluN2B inhibitor (40 mm), much
lower densities and a homogenous distribution of radioactivity
were observed. In vitro [3H]ifenprodil binding to the mouse
brain regions selected as structures of interest was thus inhibit-
ed by ifenprodil (2b) itself, Ro 25-6981, precursor 17a and
probe 1a. In contrast, MK801 and fluorescein derivative 46
were inactive.
binding in all regions (hippocampus: 70%, caudoputamen:
65%, cortex: 56%, thalamus: 40%, cerebellum: 41%) whereas
probe 1a displayed less marked values (hippocampus: 39%,
The results expressed as percentage displacement of
[3H]ifenprodil binding allowed us to refine the comparison of
ligands (Figure 5). The binding inhibition did not exceed 5%
for MK801 in the regions of interest. Values for the fluorescein
derivative 46 were in the low range of 11–18% whatever the
brain structure considered. Ifenprodil (2b) and Ro 25-6981
show structure-specific inhibition of equivalent potency. The
binding inhibition was found to be the greatest in the brain re-
gions rich in GluN2B-containing NMDARs: that is, hippocampus
(56–64%), caudoputamen (51–56%), cortex (45%) and thala-
mus (21–30%), relative to cerebellum (18–13%). Similar profiles
were observed for ifenprodil-modified 17a and probe 1a; this
confirms their recognition by a GluN2B subunit. However, ifen-
prodil-modified 17a showed the most marked inhibition of
Figure 5. Displacement studies of [3H]-ifenprodil in selected regions in
mouse brain slices. Abbreviations: cerebellum (CB), thalamus (THL), cortex
(CX), caudoputamen (CPu), hippocampus (HPC). Each bar represents mean
values ÆSEM (n=8). * indicates significantly different from ifenprodil (2b,
p<0.05). # indicates significantly different from fluorescein derivative 46
(p<0.05).
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2013, 14, 759 – 769 766

CHEMBIOCHEM
FULL PAPERS
www.chembiochem.org
3 mLminÀ1; System C: MeCN 0% (5 min), then linear gradient from
0 to 90% (45 min), then 90% (10 min), then linear gradient from
90 to 0% (10 min) and 0% (10 min) at a flow rate of 3 mLminÀ1].
UV absorption and fluorescence emission spectra were registered
with a Perkin–Elmer LS 55 spectrophotometer with the ligands at
0.003 mm in H2O solution containing DMSO (0.3%).
caudoputamen: 30%, cortex: 26%, thalamus: 14%, cerebel-
lum: 20%). Although inhibition was weak, the results for
probe 1a differed significantly from those for the fluorescein
derivative 46. The inhibitory effects observed with ifenprodil-
like 17a and probe 1a on [3H]ifenprodil ([3H]-2b) binding in
mouse brain slices could be directly related to the results ob-
tained with the functional assay showing high and moderate
affinities for 17a and 1a, respectively, relative to ifenprodil
(2b).
Synthesis of amine-functionalised ifenprodil derivatives: Ifenpro-
dil-based ligands containing either a free primary amine function
at the terminus (15a–18a, 17b, 22a, 24a, 24b, 25a, 25b, 29a,
37a, 39a and 41a) or an aminopropyl group (14a and 14b) were
prepared as shown in Schemes 1–3. Detailed procedures for the
synthesis of individual compounds and their characterisation data
are described in the Supporting Information.
Conclusions
Our objective was to develop a fluorescent probe with a high
affinity for GluN2B-containing NMDARs, with ifenprodil as the
ligand template. This study adds further weight to the argu-
ment that the design of an ifenprodil conjugate needs to
accommodate the influences both of the position and of the
nature of the associated linker to refine the pharmacological
profile of the final product. Chemical modification of the ben-
zylic hydroxy function in ifenprodil was preferable to modifica-
tion on the phenol moiety. A spacer bearing secondary amine
or amide functions was necessary to maintain affinity. Further-
more, our results confirmed that probe 1a, despite having
only a moderate affinity, displayed considerable selectivity for
the GluN2B subunit, and its distribution in the mouse brain re-
gions parallels those rich in GluN2B-containing NMDARs. With
further work, a fluorescent probe, such as one of those de-
scribed here, could become a significant tool for understand-
ing of the GluN2B “complex” in the normal and the diseased
brain.
General protocol for the synthesis of ifenprodil-based conju-
gates: A solution of amine-functionalised ifenprodil (1–1.5 equiv)
and N,N-diisopropylethylamine (DIEA, 2–3 equiv) in N-methyl-2-pyr-
rolidinone (NMP, 200 mL) was added to a solution of 6-[fluorescein-
5(6)-carboxamido]hexanoic acid N-hydroxysuccinimidyl ester
(1 equiv) in NMP (150–300 mL). The reaction mixture was protected
from the light with aluminium foil and stirred at RT for 18 h. It was
then diluted with water (2–3 mL). Purification by semipreparative
HPLC yielded the titled probe as a yellow powder after lyophilisa-
tion.
Probe 1a: This compound was obtained in 52% yield (4.5 mg)
from modified ifenprodil 17a (4.6 mg, 10.8 mmol) and 6-[fluores-
cein-5(6)-carboxamido]hexanoic acid N-hydroxysuccinimidyl ester
(5 mg, 8.5 mmol). MS (ESI+): m/z (%): 895 (60), 588 (20), 308 (100);
HRMS (ESI+): m/z calcd for C54H63N4O8: 895.4646 [M+H]+; found:
895.4666.
Probe 42a: This compound was obtained in 27% yield (3.3 mg)
from modified ifenprodil 22b (6 mg, 11.2 mmol) and 6-[fluorescein-
5(6)-carboxamido]hexanoic acid N-hydroxysuccinimidyl ester
(6.5 mg, 11 mmol). MS (ESI+): m/z (%): 504 (100), 308 (80); HRMS
(ESI+): m/z calcd for C60H73N5O9: 504.7782 [M+2H]2+; found:
504.7794.
Experimental Section
General: Dry THF and Et2O were obtained by use of an MBraun
SPS 800 solvent purification system. All other chemicals, obtained
from commercial suppliers, were used as received. Reaction prod-
ucts were purified by chromatography column with Merck silica
gel (Geduran Si 60, 0.040–0.063 nm). TLC was carried out on silica
gel 60 F254 (1.1 mm, Merck) with spot detection under UV light
and use of an appropriate staining reagent (I2, KMnO4 or ninhy-
drin). Melting points (m.p.s) were obtained with an Electrothermal
IA9000 capillary apparatus and are uncorrected. NMR spectra were
recorded at RT with Bruker DPX 250 or DRX 400 spectrometers. The
chemical shifts are calibrated to TMS (dH =0.00 ppm) or to residual
proton and carbon resonances of the solvent CDCl3 (dH =7.26 and
dC =77.16 ppm). IR spectra were acquired with a Thermo Nico-
let 380 FT-IR instrument. Mass spectra were recorded with QTOF
Micro Waters instruments. Elemental analyses were performed with
a C, H, N, S, O Thermoquest apparatus. HPLC analyses and semipre-
parative purifications were carried out with a Waters system, a 600
gradient pump, a photodiode array detector (l 210 to 750 nm) and
a reversed-phase MS C18 Waters Bondapak column (7.8ꢂ300 mm,
10 mm); elution was with MeCN and aqueous trifluoroacetic acid
(TFA, 0.1%, pH 2) [System A: MeCN 20% (15 min), then linear gradi-
ent from 20 to 80% (20 min), then 80% (5 min), then linear gradi-
ent from 80 to 20% (5 min) and 20% (5 min) at a flow rate of
3 mLminÀ1; System B: MeCN 25% (20 min), then linear gradient
from 25 to 85% (30 min), then 85% (10 min), then linear gradient
from 85 to 25% (10 min) and 25% (10 min) at a flow rate of
Probe 43b: This compound was obtained in 55% yield (4 mg)
from modified ifenprodil 25b (5.3 mg, 11 mmol) and 6-[fluorescein-
5(6)-carboxamido]hexanoic acid N-hydroxysuccinimidyl ester (4 mg,
6.8 mmol). MS (ESI+): m/z (%): 476 [M+H]2+ (100), 467
[M+HÀH2O]2+ (90); HRMS (ESI+): m/z calcd for C56H67N5O9:
476.7469 [M+2H]2+; found: 476.7459.
Probe 44a: This compound was obtained in 53% yield (3.7 mg)
from modified ifenprodil 29b (4.9 mg, 11.2 mmol) and 6-[fluores-
cein-5(6)-carboxamido]hexanoic acid N-hydroxysuccinimidyl ester
(4 mg, 6.8 mmol). MS (ESI+): m/z (%): 455 [M+2H]2+ (40), 446
[M+2HÀH2O]2+ (100); HRMS (ESI+): m/z calcd for C54H62N4O9:
455.2258 [M+2H]2+; found: 455.2255.
Probe 45a: This compound was obtained in 57% yield (4 mg)
from modified ifenprodil 37b (7.1 mg, 15.7 mmol) and 6-[fluores-
cein-5(6)-carboxamido]hexanoic acid N-hydroxysuccinimidyl ester
(4 mg, 6.8 mmol). MS (ESI+): m/z (%): 463 [M+2H]2+ (60), 454
[M+2HÀH2O]2+ (100); HRMS (ESI+): m/z calcd for C54H62N4O10:
463.2233 [M+2H]2+; found: 463.2225.
Ifenprodil binding site analysis: The structure of the heterodimer
GluN1/GluN2B in complexation with ifenprodil (PDB ID: 3QEL) was
used for the structural analysis of the binding site of ifenprodil.
Figure 2 was drawn with PYMOL (DeLano Scientific, 2006, San
Carlo, USA).
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2013, 14, 759 – 769 767

CHEMBIOCHEM
FULL PAPERS
www.chembiochem.org
Primary cultures of cortical neurons: Cultures of murine cortical
neurons were prepared from 14- to 15-day-old embryos as de-
scribed previously.[20] Cerebral cortices were dissected, dissociated
and cultured on 24-plates coated with poly-d-lysine (Sigma,
0.1 mgmLÀ1) and laminin (Invitrogen, 0.02 mgmLÀ1) in Dulbecco’s
modified Eagle’s medium (DMEM, Invitrogen) containing foetal
bovine serum (FBS, 5%), horse serum (5%) and glutamine (2 mm)
(all from Sigma). Cultures were kept at 378C in a humidified atmos-
phere containing CO2 (5%). After 3–4 days in vitro (DIV), cytosine
b-arabinoside furanoside (Ara-C, 10 mm) was added to inhibit prolif-
eration of non-neuronal cells. Experiments were performed on neu-
ronal cultures after 14 DIV. Under these conditions, about 98% of
the cells were considered neuronal, as judged by microtubule-asso-
ciated protein-2 staining.
is the ratio measured in a saturated Ca2+ solution, Kd =135 nm (the
dissociation constant of fura-2), and F0 and Fs are the fluorescence
intensities measured in a Ca2+-free solution at 380 nm or in a satu-
rated Ca2+ solution at 380 nm, respectively.[21] We assessed the in-
crease in intracellular calcium ([Ca2+]i) by measuring the area under
the curve for two minutes after the start of the stimulation. Be-
cause of the uncertainties inherent in calibration procedures, the
concentrations presented here should be considered “estimated”.
Quantification of neuronal cell death by lactate dehydrogenase
(LDH) release: At 14 DIV, the culture medium of primary cortical
neurons was replaced with serum-free DMEM. Primary cortical neu-
rons were then treated for 24 h with NMDA (15 mm) in the pres-
ence or the absence of the tested compounds (10 mm). After treat-
ment, an aliquot (50 mL) was removed from the neuronal culture
medium, and neurotoxicity was quantified by measuring the
amount of LDH released by dead and dying cells. LDH activity was
determined by recording the decrease in absorbance at 340 nm
corresponding to the oxidation of NADH into NAD+. The LDH level
associated with 100% neuronal death was determined in associat-
ed cultures exposed to Triton X-100 (0.1%) for 10 min. Background
LDH release was measured in cultures subjected to a control wash
and was subtracted from the experimental values. Results were ex-
pressed in percentages of neuronal cell death on a scale between
0 and 100%.
HEK293 cell culture and plasmid transfection: cDNAs encoding
the full-length sequences of GluN1-1a and GluN2B-GFP were a gen-
erous gift from Dr. Stefano Vicini, (Georgetown University Medical
Center, Washington). Human embryonic kidney 293 (HEK293) cells
were grown in DMEM supplemented with l-glutamine (1%), strep-
tomycin/penicillin/neomycin (1%) and FBS (10%). Exponentially
growing cells were plated on 12 mm glass coverslips (Biospace,
MatTek Corporation) coated with poly-d-lysine and grown at 378C
in a humidified atmosphere containing CO2 (5%). Cells, 24 h after
plating, were transfected with cDNA in Transfast Reagent (7.5 mL,
Promega). Co-transfections of GluN1-1a and GluN2B-GFP were per-
formed at a ratio of 1:1. In addition, upon transfection, cells were
treated with (2R)-amino-5-phosphonovaleric acid (d-APV, 100 mm)
to prevent cell death. Studies of the recombinant receptors were
performed 24 h after transfection.
Autoradiography and displacement studies: Adult male Swiss
mice weighing about 22 g were obtained from the adjacent Uni-
versity of Caen’s Centre for Biological Resources (CURB, authorisa-
tion for breeding and housing experimental animals: A14 118 015).
Anaesthesia was induced with isoflurane (5%) in oxygen and ni-
trous oxide (30/70) and the mice were killed by decapitation (au-
thorisation: M. D. 14 54). Brains were rapidly removed, frozen in
pre-cooled (dry ice) isopentane and stored at À208C. Brains were
cut with a cryomicrotome (Leica cryostat CM 3050) in 20 mm ros-
trocaudal adjacent sections containing cerebral structures with dif-
ferent densities of GluN2B-containing NMDARs (i.e., cortex, hippo-
campus, thalamus, caudoputamen and cerebellum). The coronal
sections were applied to glass slides with low non-specific binding
(SuperFrost Plus, Menzel-Glaser GmbH, Braunschweig, Germany)
and stored at À208C until utilisation. Adjacent sections were pre-
incubated twice (5 min, 48C) in Tris·HCl buffer solution (Trizma
Base, Sigma–Aldrich, 50 mm) containing NaCl (100 mm), bovine
Calcium imaging: Cultures on glass-bottomed dishes (N8 0, cover-
glass 0.085–0.13 mm, Mattek Corporation, New Jersey, USA) were
loaded for 45 min with fura-2-acetoxymethyl ester (10 mm, AM, F-
1201, Invitrogen) and pluronic acid (0.2%) and were then incubat-
ed for an additional 15 min at RT in HEPES and bicarbonate-buf-
fered saline solution (HBBSS) containing NaCl (116 mm), KCl
(5.4 mm), CaCl2 (1.8 mm), MgSO4 (0.8 mm), NaH2PO4 (1.3 mm),
HEPES (12 mm), glucose (5.5 mm), bicarbonate (25 mm) and glycine
(10 mm) at pH 7.45. Experiments were performed at RT with perfu-
sion at 2 mLminÀ1 (peristaltic pump) on the stage of a Nikon
Eclipse inverted microscope with use of a quartz light pathway,
a xenon lamp (100 W) and an oil immersion objective (Nikon CFI
40X NA. 1.3 optimised for fluorescence imaging). An intensified
charge-coupled device camera (Coolsnap EZ) recorded the fluores-
cence emissions (>510 nm) evoked by excitation through narrow
band-pass filters (340Æ5 nm/380Æ6.5 nm) housed in a computer-
controlled filter wheel. Fluorescence images were acquired with
a PC-based computer at a maximal time resolution of 2 s per suc-
cessive measurement with use of Metafluor 6.3 (Universal Imaging
Corporation). Images were background-subtracted as follows:
a frame was taken at each excitation wavelength from an area of
the coverslip devoid of cells. Measurements were obtained simulta-
neously from ꢀ30 HEK transfected cells or ꢀ150 neurons per
probe. Calibration of the fura-2 signal was achieved by determin-
ing both the maximum 340/380 ratio and the minimum 340/380
ratio of fura-2. The maximum fura-2 ratio was achieved with satu-
rating levels of Ca2+, and the minimum fura-2 ratio was obtained
in the presence of the Ca2+ chelator ethyleneglycol tetraacetic acid
with no added Ca2+, both in the presence of ionomycin. Ca2+ con-
centrations measured at intermediate ratios were calculated ac-
cording to the relationship described for fura-2: [Ca2+]i =Kd-
[(RÀRmin)/(RmaxÀR)] F0/Fs, where R is the measured ratio of 340/380
fluorescence, Rmin is the ratio measured in a Ca2+-free solution, Rmax
serum albumin (BSA, 1 gLÀ1
, Sigma–Aldrich) and bacitracin
(20 mgLÀ1, ICN Biomedicals, Irvine, CA, USA) adjusted to a final pH
of 7.4. The slides were then incubated at 48C for 90 min in buffer
(700 mL) containing [3H]ifenprodil (40 nm, specific radioactivity
40 CimmolÀ1, Perkin–Elmer). [3H]Ifenprodil binding was displaced
in the presence of ifenprodil (Sigma–Aldrich) or tested compounds
(all at 40 mm). Additional sections were incubated in the presence
of (R)-(+)-3-PPP [(R)-(+)-3-(3-hydroxyphenyl)-N-propylpiperidine,
100 mm, Sigma–Aldrich, Saint Quentin Fallavier, France; D2 agonist
and s1 antagonist], GBR 12909 (30 mm, Tocris Bioscience; dopa-
mine reuptake inhibitor and s ligand) and GBR 12935 (30 mm,
Tocris Bioscience; norepinephrine and dopamine reuptake inhibi-
tor) to avoid binding in all ifenprodil targets except the GluN2B
NMDAR. Ifenprodil was dissolved in Tris·HCl buffer and DMSO at
10 mm. (+)-3-PPP, GBR 12909 and GBR 12935 were dissolved in
Tris·HCl buffer. After incubation, slices were subjected to four suc-
cessive washes (30 s) at 48C in Tris·HCl buffer (first with BSA and
the three others without BSA). The excess buffer salt was removed
in a final 30 s wash at 48C in distilled water. Sections were dried
overnight at RT. Slices were apposed to a storage phosphor screen
(Perkin–Elmer) in a film cassette. Cassettes were wrapped in
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2013, 14, 759 – 769 768

CHEMBIOCHEM
FULL PAPERS
www.chembiochem.org
18803; d) B. Laube, J. Kuhse, H. Betz, J. Neurosci. 1998, 18, 2954–2961;
e) D. J. Goebel, M. S. Poosch, Mol. Brain Res. 1999, 69, 164–170; f) P.
Paoletti, Eur. J. Neurosci. 2011, 33, 1351–1365.
stretchable film to protect slides against dampness and kept at
À208C to decrease non-specific binding. After 4 days of exposure,
the screens were removed from cassettes under dim lighting and
immediately scanned (at a resolution of 600 dpi) in a Cyclone Stor-
age Phosphor system (Perkin–Elmer). Before and after the reading,
screens were erased by exposure to white light on a standard ra-
diographic light box. Anatomical regions of interest of surface area
sufficient to provide a total count of adequate statistical power
were identified in the Watson and Paxinos atlas[22] and then de-
lineated in the digital autoradiograms with OptiQuant v03.00 soft-
ware (Packard Instrument Co.). The bound radioligand signal was
obtained in a DLU/mm2 (Digital Light Unit) with Cyclone and the
OptiQuant software.
[6] a) J. M. Loftis, A. Janowsky, Pharmacol. Ther. 2003, 97, 55–85; b) C. C.
Wang, R. G. Held, S. C. Chang, L. Yang, E. Delpire, A. Ghosh, B. J. Hall,
Neuron 2011, 72, 789–805.
[7] a) G. E. Hardingham, Y. Fukunaga, H. Bading, Nat. Neurosci. 2002, 5,
405–414; b) S. G. Cull-Candy, D. N. Leszkiewicz, Sci. STKE 2004, 255,
re16; c) C. M. Gladding, L. A. Raymond, Mol. Cell. Neurosci. 2011, 48,
308–320.
[8] P. Marchand, J. Becerril-Ortega, L. Mony, C. Bouteiller, P. Paoletti, O.
Nicole, L. Barrꢀ, A. Buisson, C. Perrio, Bioconjugate Chem. 2012, 23, 21–
26.
[9] a) A. P. Tamiz, E. R. Whittemore, Z.-L. Zhou, J.-C. Huang, J. A. Drewe, J.-C.
Chen, S.-X. Cai, E. Weber, R. M. Woodward, J. F. W. Keana, J. Med. Chem.
1998, 41, 3499–3506; b) S. S. Nikam, L. T. Meltzer, Curr. Pharm. Des.
2002, 8, 845–855; c) J. A. McCauley, Expert Opin. Ther. Pat. 2005, 15,
389–407; d) K. R. Gogas, Curr. Opin. Pharmacol. 2006, 6, 68–74; e) M. E.
Layton, M. J. Kelly III., K. J. Rodzinak, Curr. Top. Med. Chem. 2006, 6,
697–709; f) I. Borza, G. Domany, Curr. Top. Med. Chem. 2006, 6, 687–
695; g) J. A. McCauley, Expert Opin. Ther. Pat. 2006, 16, 863–870; h) M.
Koller, S. Urwyler, Expert Opin. Ther. Pat. 2010, 20, 1683–1702.
[10] a) I. J. Reynold, R. Miller, Mol. Pharmacol. 1989, 36, 758–765; b) B. L.
Chenard, I. A. Shalaby, B. K. Koe, R. T. Ronau, T. W. Butler, M. A. Proch-
niak, A. W. Schmidt, C. B. Fox, J. Med. Chem. 1991, 34, 3085–3090; c) K.
Williams, Mol. Pharmacol. 1993, 44, 851–859; d) V. Mutel, D. Buchy, A.
Klingelschmidt, J. Messer, Z. Bleuel, J. A. Kemp, J. G. Richards, J. Neuro-
chem. 1998, 70, 2147–2154; e) K. Williams, Curr. Drug Targets 2001, 2,
285–298; f) T. W. Rosahl, P. B. Wingrove, V. Hunt, R. L. Fradley, J. M. K.
Lawrence, R. P. Heavens, P. Treacey, M. Usala, A. Macauley, T. P. Bonnert,
P. J. Whiting, K. A. Wafford, Mol. Cell. Neurosci. 2006, 33, 47–56.
[11] L. Mony, L. Krzaczkowski, M. Leonetti, A. Le Goff, K. Alarcon, J. Neyton,
H.-O. Bertrand, F. Acher, P. Paoletti, Mol. Pharmacol. 2009, 75, 60–74.
[12] a) T. Masuko, K. Kashiwagi, T. Kuno, N. D. Nguyen, A. J. Pahk, J. Fukuchi,
K. Igarashi, K. Williams, Mol. Pharmacol. 1999, 55, 957–969; b) F. Perin-
Dureau, J. Rachline, J. Neyton, P. Paoletti, J. Neurosci. 2002, 22, 5955–
5965; c) L. Marinelli, S. Cosconati, T. Steinbrecher, V. Limongelli, A. Berta-
mino, E. Novellino, D. A. Case, ChemMedChem 2007, 2, 1498–1510;
d) F.-M. Ng, M. T. Geballe, J. P. Snyder, S. F. Traynelis, C.-M. Low, Mol.
Brain 2008, 1, 16; e) L. Mony, J. N. C. Kew, M. J. Gunthorpe, P. Paoletti,
Br. J. Pharmacol. 2009, 157, 1301–1317.
[13] J. A. McCauley, C. R. Theberge, J. J. Romano, S. B. Billings, K. D. Ander-
son, D. A. Claremon, R. M. Freidinger, R. A. Bednar, S. D. Mosser, S. L.
Gaul, T. M. Connolly, C. L. Condra, M. Xia, M. E. Cunningham, B. Bednar,
G. L. Stump, J. J. Lynch, A. Macaulay, K. A. Wafford, K. S. Koblan, N. J. Liv-
erton, J. Med. Chem. 2004, 47, 2089–2096.
[14] G. Barta-Szalai, I. Borza, E. Bozꢃ, C. Kiss, B. Agai, A. Proszenyꢄk, G. M.
Keseru, A. Gere, S. Kolok, K. Galgꢃczy, C. Horvꢄth, S. Farkas, G. Domꢄny,
Bioorg. Med. Chem. Lett. 2004, 14, 3953–3956.
[15] P. Avenet, J. Leonardon, F. Besnard, D. Graham, J. Frost, H. Depoortere,
S. Z. Langer, B. Scatton, Eur. J. Pharmacol. 1996, 296, 209–213.
[16] C. Bouteiller, J. Becerril-Ortega, P. Marchand, O. Nicole, L. Barrꢀ, A. Buis-
son, C. Perrio, Org. Biomol. Chem. 2010, 8, 1111–1120.
[17] E. Karakas, N. Simorowski, H. Furukawa, Nature 2011, 475, 249–254.
[18] F. El Gaamouch, A. Buisson, O. Moustiꢀ, M. Lemieux, S. Labrecque, B.
Bontempi, P. De Koninck, O. Nicole, J. Neurosci. 2012, 32, 10767–10779.
[19] a) J. Benavides, B. Peny, J. Allen, B. Scatton, J. Pharmacol. Exp. Ther.
1992, 260, 896–901; b) V. Mutel, D. Buchy, A. Klingelschmidt, J. Messer,
Z. Bleuel, J. A. Kemp, J. G. Richards, J. Neurochem. 1998, 70, 2147–2155;
c) S. Grimwood, P. Richards, F. Murray, N. Harrison, P. B. Wingrove, P. H.
Hutson, J. Neurochem. 2000, 75, 2455–2463.
Statistical analyses: Data for displacement studies of [3H]ifenprodil
are presented as means Æstandard errors of means (SEMs). Statisti-
cal analyses were performed with Sigma Plot 11.2Tm (Jandel Scien-
tific, Germany). Multiple comparisons versus control group (ifen-
prodil or compound 46) for each cerebral area were performed by
the Bonferroni t-test. All p values <0.05 are considered to be stat-
istically significant.
Acknowledgements
Financial support for this research was provided by the MERNT
(Ministꢁre de la Recherche et des Nouvelles Technologies), CNRS
(Centre National de la Recherche Scientifique), CEA (Commissariat
ꢂ l’Energie Atomique), the “CRUNCH” Network (Pꢃle Universitaire
Normand de Chimie Organique), the “Rꢀgion Basse-Normandie”,
and the European Union (FEDER funding).
Keywords: fluorescent probes · GluN2B · ifenprodil · NMDA
receptors · structure–activity relationships
[1] a) R. Dingledine, K. Borges, D. Bowie, S. F. Traynelis, Pharmacol. Rev.
1999, 51, 7–61; b) P. Paoletti, J. Neyton, Curr. Opin. Pharmacol. 2007, 7,
39–47.
[2] a) G. L. Collingridge, W. Singer, Trends Pharmacol. Sci. 1990, 11, 290–
296; b) T. V. P. Bliss, G. L. Collingridge, Nature 1993, 361, 31–39; c) L.
Kaczmarek, M. Kossut, J. Skangiel-Kramska, Physiol. Rev. 1997, 77, 217–
255; d) S. Cull-Candy, S. Brickley, M. Farrant, Curr. Opin. Neurobiol. 2011,
11, 327–335; e) G. E. Hardingham, H. Bading, Trends Neurosci. 2003, 26,
81–89; f) P. Paoletti, J. Neyton, Curr. Opin. Pharmacol. 2007, 7, 39–47;
g) W. E. Childers, Jr., R. B. Baudy, J. Med. Chem. 2007, 50, 2557–2562.
[3] a) S. A. Lipton, P. A. Rosenberg, N. Engl. J. Med. 1994, 330, 613–622;
b) M. Arundine, M. Tymianski, Cell Calcium 2003, 34, 325–337.
[4] a) S. Boyce, A. Wyatt, J. K. Webb, R. O’Donnell, G. Mason, M. Rigby, D.
Sirinathsinghji, R. G. Hill, N. M. Rupniak, Neuropharmacology 1999, 38,
611–623; b) K. Steece-Collier, L. K. Chambers, S. S. Jaw-Tsai, F. S. Menniti,
J. T. Greenamyre, Exp. Neurol. 2000, 163, 239–243; c) L. Li, M. Fan, C. D.
Icton, N. Chen, B. R. Leavitt, M. R. Hayden, T. H. Murphy, L. Raymond,
Neurobiol. Aging 2003, 24, 1113–1121; d) P. L. Chazot, Curr. Med. Chem.
2004, 11, 389–396; e) L. Li, T. H. Murphy, M. R. Hayden, L. A. Raymond,
J. Neurophysiol. 2004, 92, 2738–2746; f) W. Guo, F. Wei, S. Zou, M. T.
Robbins, S. Sugiyo, T. Ikeda, J. C. Tu, P. F. Worley, R. Dubner, K. Ren, J.
Neurosci. 2004, 24, 9161–9173; g) C. X. Wang, A. Shuaib, Curr. Drug Tar-
gets: CNS Neurol. Disord. 2005, 4, 143–151; h) G. A. Higgins, T. M. Bal-
lard, M. Enderlin, M. Haman, J. A. Kemp, Psychopharmacology 2005, 179,
85–98; i) C. Beinat, S. Banister, I. Moussa, A. J. Reynolds, C. S. McErlean,
M. Kassiou, Curr. Med. Chem. 2010, 17, 4166–4190.
[20] F. Lꢀveillꢀ, F. El Gaamouch, E. Gouix, M. Lecocq, D. Lobner, O. Nicole, A.
Buisson, FASEB J. 2008, 22, 4258–4271.
[21] G. Grynkiewicz, M. Poenie, R. Y. Tsien, J. Biol. Chem. 1985, 260, 3440–
3450.
[22] C. Watson, G. Paxinos, Chemoarchitectonic Atlas of The Mouse Brain, 1st
ed., Elsevier, 2010.
[5] a) N. Brose, G. P. Gasic, D. E. Vetter, J. M. Sullivan, S. F. Heinemann, J. Biol.
Chem. 1993, 268, 22663–22671; b) T. Priestley, P. Laughton, J. Myers, B.
Le Bourdelles, J. Kerby, P. J. Whiting, Mol. Pharmacol. 1995, 48, 841–
848; c) E. Meddows, B. Le Bourdelles, S. Grimwood, K. Wafford, S.
Sandhu, P. Whiting, R. A. McIlhinney, J. Biol. Chem. 2001, 276, 18795–
Received: October 29, 2012
Published online on March 26, 2013
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2013, 14, 759 – 769 769