A new phenylalanine derivative acts as an antagonist at the AMPA Receptor GluA2 and introduces partial domain closure: Synthesis, resolution, pharmacology, and crystal structure

Article abstract of DOI:10.1021/jm200862h

In order to map out molecular determinants for competitive blockade of AMPA receptor subtypes, a series of 2-carboxyethylphenylalanine derivatives has been synthesized and pharmacologically characterized in vitro. One compound in this series, (RS)-3h, showed micromolar affinity for GluA1_o and GluA2(R)_o receptors with an approximately 4-fold preference for GluA1/2 vs GluA3/4. In TEVC electrophysiological experiments (RS)-3h competitively antagonized GluA2(Q)_i receptors. The X-ray structure of the active enantiomer (S)-3h in complex with GluA2-S1S2J showed a domain closure around 8°. Even though the nitro and the carboxyethyl groups of (S)-3h were both anchored to Tyr702 through a water H-bond network, these interactions only induced weak subtype selectivity. In spite of the fact that (S)-3h induced a domain closure close to that observed for partial agonists, it did not produce agonist responses at GluA2 receptors under nondesensitizing conditions. 2-Carboxyethylphenylalanine derivatives provide a new synthetic scaffold for the introduction of substituents that could lead to AMPA receptor subtype-selective ligands.

Full text of DOI:10.1021/jm200862h

ARTICLE
pubs.acs.org/jmc
A New Phenylalanine Derivative Acts as an Antagonist at the AMPA
Receptor GluA2 and Introduces Partial Domain Closure: Synthesis,
Resolution, Pharmacology, and Crystal Structure†
||
Ewa Szymaꢀnska,‡, ,# Karla Frydenvang,‡,# Alberto Contreras-Sanz,§ Darryl S. Pickering,§ Elena Frola,‡
Zorica Serafimoska,‡ Birgitte Nielsen,‡ Jette S. Kastrup,‡ and Tommy N. Johansen*,‡
‡Department of Medicinal Chemistry and §Department of Pharmacology and Pharmacotherapy, Faculty of Pharmaceutical Sciences,
University of Copenhagen, 2 Universitetsparken, DK-2100 Copenhagen, Denmark
S Supporting Information
b
ABSTRACT: In order to map out molecular determinants for
competitive blockade of AMPA receptor subtypes, a series of
2-carboxyethylphenylalanine derivatives has been synthesized
and pharmacologically characterized in vitro. One compound in
this series, (RS)-3h, showed micromolar affinity for GluA1o and
GluA2(R)o receptors with an approximately 4-fold preference
for GluA1/2 vs GluA3/4. In TEVC electrophysiological experi-
ments (RS)-3h competitively antagonized GluA2(Q)i receptors.
The X-ray structure of the active enantiomer (S)-3h in complex
with GluA2-S1S2J showed a domain closure around 8°. Even
though the nitro and the carboxyethyl groups of (S)-3h were
both anchored to Tyr702 through a water H-bond network, these interactions only induced weak subtype selectivity. In spite of the
fact that (S)-3h induced a domain closure close to that observed for partial agonists, it did not produce agonist responses at GluA2
receptors under nondesensitizing conditions. 2-Carboxyethylphenylalanine derivatives provide a new synthetic scaffold for the
introduction of substituents that could lead to AMPA receptor subtype-selective ligands.
’ INTRODUCTION
have been synthesized and characterized pharmacologically, only
few of them have shown high AMPA receptor subunit selectivity.8
Among the intensively studied class of uracil-based AMPA
ligands, (S)-5-fluorowillardiine exhibited 55-fold higher potency
at GluA1 than GluA3 receptors9 and a structurally very similar
3-hydroxypyridazine 1-oxide derivative was reported showing
161-fold selectivity for GluA1 over GluA3 receptors.10 Subunit-
selective agonists are also found in the group of isoxazole-
based analogues of AMPA. One such ligand is the ibotenic
acid derivative (S)-2-amino-3-(4-bromo-3-hydroxyisoxazol-5-yl)
propionic acid [(S)-Br-HIBO, 1] (Figure 1), a GluA1/2 prefer-
ring AMPA receptor agonist with a 70-fold selectivity vs GluA3.11À13
The corresponding chloro analogue Cl-HIBO exhibited a >1000-
fold selectivity.14 On the basis of detailed structural analyses, it
has been claimed13 that at least part of the marked selectivity
among AMPA receptors can be ascribed to interactions between
the isoxazole moiety of Br-HIBO and one particular amino acid
residue, Tyr698 of GluA1. In GluA2, this residue is conserved
(Tyr702), whereas it is a phenylalanine in GluA3 and GluA4.
To the best of our knowledge, similar AMPA receptor subunit
selectivity has not yet been identified among competitive anta-
gonists.8 Even though several structural classes of antagonists are
Ionotropic glutamate receptors (iGluRs) are ligand-gated ion
channels that, based on agonist binding affinities and sequence
similarities, have been divided into three functionally distinct
subclasses: (RS)-2-amino-3-(3-hydroxy-5-methylisoxazol-4-yl)-
propionic acid (AMPA), kainic acid (KA), and N-methyl-D-
aspartic acid (NMDA) receptors. Native AMPA receptors, which
mediate the majority of the fast excitatory amino acid synaptic
transmission in the central nervous system (CNS), are cation-
selective tetrameric homo- or heterooligomers formed by com-
binations of subunits GluA1À4. Across the mammalian brain, a
large heterogeneity of AMPA receptors can be observed, differ-
entiated further by splicing of each subunit in an extracellular
region (flip and flop variants) that modifies channel kinetics.1À5
Additional diversity comes from RNA editing of a residue in the
pore region of GluA2 from glutamine to arginine, which significantly
reduces calcium permeability and decreases single channel conduc-
tance in AMPA GluA2-containing receptors6 and RNA editing of an
R/G site controlling channel desensitization properties.7
The therapeutic significance of the individual AMPA receptor
subunits is not clearly understood, mainly because of lack of
potent and selective pharmacological tools. Subunit-selective
ligands are therefore valuable, at least as pharmacological tools
to investigate the physiological roles of the individual receptor
subtypes. Despite a large number of AMPA receptor agonists that
Received: July 1, 2011
Published: September 19, 2011
r
2011 American Chemical Society
7289
dx.doi.org/10.1021/jm200862h J. Med. Chem. 2011, 54, 7289–7298
|

Journal of Medicinal Chemistry
ARTICLE
Scheme 1a
a (i) NaCN, DMF, 80 °C; (ii) 60% H2SO4, reflux; (iii) LiAlH4, THF,
0 °C f reflux.
The most active compound, (RS)-2-amino-3-(2-(2-carboxyethyl)-
5-chloro-4-nitrophenyl)propionic acid (3h), was also examined
by two-electrode voltage clamp (TEVC) electrophysiology of
AMPA receptors expressed in X. laevis oocytes. To investigate the
activity of individual enantiomers, 3h was resolved into S- and R-
enantiomers using chiral HPLC. Finally, (S)-3h was crystallized
in complex with GluA2-S1S2J to 1.9 Å resolution.
Figure 1. Chemical structures of selected AMPA receptor ligands as
well as a general structure of the target compounds.
Chemistry. The syntheses of the new phenylalanine deriva-
tives are illustrated in Schemes 1À5. Starting materials 7b,d were
commercially available, whereas 2-(4-tert-butylphenyl)ethanol
7c was synthesized in three steps from benzyl bromide 4c
(Scheme 1). The target carboxyethyl-substituted phenylalanines
3aÀd,f,g were prepared in analogy to a strategy described by
Hamilton.16 Key intermediates, isochromans 8bÀd, were pre-
pared from the alcohols 7bÀd which were converted into the
corresponding MEM-acetals and cyclized to the isochromans
using Lewis acids (Scheme 2).
known, a possible reason for the lack of selectivity may be the fact
that none of the antagonists occupy the cavity in the neighbor-
hood of Tyr702 in their binding to GluA2. One intensively
studied class of competitive antagonists is the isoxazole-based
acidic amino acids, of which (S)-2-amino-3-[5-tert-butyl-3-(phos-
phonomethoxy)isoxazol-4-yl]propionic acid [(S)-ATPO, 2] is a
well-known example. Structural information of the binding of
(S)-ATPO to the AMPA receptor GluA2 is known (PDB code
1N0T).15 Analysis of this crystal structure reveals that the
isoxazole ring of (S)-ATPO, when interacting with the GluA2
ligand binding domain (LBD, GluA2-S1S2), mainly acts as a
linker between the α-amino acid group and the phosphonate
chain. Furthermore, the volume of the binding pocket seems to
be large enough to accommodate other spacers, for instance, a
phenyl ring, offering possibilities to introduce substituents
pointing into different areas of the binding pocket, not accessible
when using the isoxazole spacer of (S)-ATPO. 2-Substituted
phenylalanine-based acidic amino acids have been reported to
antagonize AMPA and kainate receptors,16,17 making such class
of compounds a relevant starting point for the design of com-
petitive AMPA receptor antagonists useful for exploring the
structural basis for blockade of individual subtypes of AMPA
receptors at the molecular level. We therefore synthesized a series
of 2-carboxyethylphenylalanine-based acidic amino acids 3
(Figure 1) and solved the X-ray structure of the GluA2 LBD in
complex with the most potent compound within the series, 3h.
The pharmacological data were satisfactorily consistent with
information derived from the X-ray cocrystal structure which
was used to map out some of the molecular determinants defining
AMPA receptor subtype selectivity.
Ring-opening of isochromans 8aÀd using 33% HBr in acetic
acid, following nucleophilic substitution with diethyl acetamido-
malonate under basic conditions, gave bromomalonates 10aÀd,
which were converted to the corresponding nitriles 11aÀd using
sodium cyanide. In the case of 11a,c, dry DMF was used as
solvent, resulting in relatively low yields of the substitution
products and relatively high yields of corresponding elimination
products. For this reason in the cases of 11b,d the solvent was
changed to dry DMSO which markedly increased yields of the
expected nitriles and suppressed the elimination reaction. Hy-
drolysis of 11aÀd using aqueous 48% HBr gave the desired
amino acids 3aÀd as zwitterions either based on treatment with
propylene oxide and subsequent crystallization or alternatively
using HPLC purification. Compound 12, obtained as a result of
radical bromination of the intermediate nitrile 11b with NBS,
was first refluxed with 48% HBr and then, after evaporation of
acid, refluxed for 2 h in water to give amino acid 3e (Scheme 3).
Preparation of isochromans using the asymmetric alcohol 7f
resulted in an inseparable mixture of the two regioisomers 8f,g in
a 1.6:1 ratio (Scheme 4). The mixture of 8f,g was then cleaved by
33% HBr in acetic acid to 9f,g and subsequently treated with
diethyl acetamidomalonate to afford a mixture of 10f,g which
could be separated by fractional crystallization from heptane. 10f
and 10g were finally converted to nitriles 11f,g and hydrolyzed in
analogy to the preparation of 3aÀd.
’ RESULTS
In the target phenylalanines a carboxyethyl chain was intro-
duced into position 2 as a distal acidic group and various sub-
stitutions in positions 4, 5, and 6 (in structure 3 R1, R2, and R3,
respectively) were examined to explore the available space in the
ligand binding cleft. All new target amino acids were pharmaco-
logically characterized by radioligand binding at native AMPA
receptors in rat brain membranes and at homomeric recombi-
nant rat AMPA receptors expressed in Sf9 insect cell membranes.
To obtain compounds 3hÀj, compound 11d was treated with
a mixture of nitric acid and sulfuric acid in dichloromethane,
resulting in nitration exclusively in position 4 of the phenyl ring
(Scheme 5). The nitration was accompanied by addition of water
to the nitrile. Reduction of the nitro group with iron in the
presence of hydrochloric acid gave the corresponding amine 14,
which was converted to 15 in a Sandmeyer reaction using copper
7290
dx.doi.org/10.1021/jm200862h |J. Med. Chem. 2011, 54, 7289–7298

Journal of Medicinal Chemistry
ARTICLE
cyanide. Hydrolysis of intermediates 13, 14, and 15 and purifica-
tion in analogy to the preparation of 3aÀd afforded the target
amino acids 3hÀj.
column,20,21,23 strongly suggests that the first eluting enantiomer
on the Sumichiral column has S-configuration. To strengthen
this configurational assignment, CD spectra were recorded for
both enantiomers and for (S)-Phe as a reference compound.
Similar to (S)-Phe, the less retained enantiomer showed a
positive Cotton effect at 205 nm (Δε = 0.44 m2/mol), which
is in agreement with published data observed for (S)-α-amino
acids in acidic solution.24
Chromatographic Resolution of 3h. Resolution of racemic
3h was performed by chiral HPLC using a Sumichiral OA-5000
ligand-exchange column containing N,S-dioctyl-D-penicillamine18,19
as chiral selector to give the R- and S-enantiomers of 3h with the
satisfactory enantiomeric excess (ee): (R)-3h, 95% ee; (S)-3h,
97% ee. The ee of the isolated enantiomers was determined using
a Chirobiotic T column.20,21 The first eluting enantiomer on the
Sumichiral column was also the first eluting enantiomer on the
Chirobiotic column. This elution order, supported by the pre-
viously reported elution orders of a number of α-amino acids
(including the structurally related amino acids Phe and Tyr)
on both the Sumichiral column19,22 and the Chirobiotic T
Binding Pharmacology. Compounds 3aÀj were evaluated at
native and cloned AMPA receptors in radioligand binding assays
(Tables 1 and 2). Among this series, compound 3h (R1 = NO2;
R2 = Cl; R3 = H; Table 1) was found to have low micromolar
affinity at native AMPA receptors, compounds 3b,dÀf showed
high micromolar affinities, while compounds 3a,c,g did not show
detectable affinity. Compound 3h was subsequently character-
ized at all the cloned rat GluA1À4 (flop isoforms), as well as at
the soluble GluA2 constructs25 GluA2-S1S2J and (Y702F)GluA2-
S1S2J (Table 2). Hill coefficients were near unity at all receptors,
indicating binding to a single population of sites in each case. A
4-fold difference in affinity at GluA1/2 vs GluA3/4 was observed
for 3h (P < 0.05, one way ANOVA, Bonferroni post-test); however,
the affinity was not statistically significantly different between
full-length GluA2(R)o, construct GluA2-S1S2J, or (Y702F)GluA2-
S1S2J (P = 0.233, one way ANOVA). This suggests that, in
contrast to the situation with Br-HIBO, this Y/F amino acid
position is not critical for the observed binding affinity profile of
3h. However, the full-length receptor mutant (Y716F)GluA1o
did exhibit an increased Ki for 3h, suggesting that the Y/F
position can contribute to the GluA1/2 vs GluA3/4 preference
of 3h.
Scheme 2a
Functional Pharmacology. By use of TEVC electrophysiol-
ogy, the antagonist properties of the enantiomers of 3h were
examined at homomeric GluA2(Q)i and GluA3i expressed in X.
laevis oocytes (Table 3). (RS)-3h dose-dependently antagonized
the responses to (S)-Glu (Figure 2) and exhibited only a 2.4-
fold selectivity in the calculated Kb values toward GluA2(Q)i
vs GluA3i (P < 0.001, Student’s t test). Slopes of (RS)-3h
concentrationÀresponse inhibition curves were close to unity.
The functional Kb values of (RS)-3h at GluA2(Q)i and GluA3i
(Table 3) were not statistically significantly different from their
respective Ki values determined from radioligand binding experi-
ments (P > 0.05, one way ANOVA, Bonferroni post-test) (Table 2).
The activity of 3h resides in the S-enantiomer, which had a
calculated Kb at GluA2(Q)i of approximately one-half that of the
racemate, while the low potency observed for (R)-3h (95% ee)
was in good agreement with the observed percentage of the S-
enantiomer left after recrystallization, supporting (R)-3h as being
inactive as an antagonist. A rightward shift in the (S)-Glu
concentrationÀresponse curve with unchanged maximal re-
sponse indicates competitive antagonism by 3h (Figure 3).
a (i)(1) MEM-chloride, DIPEA, CH2Cl2, roomtemp; (2) TiCl4, CH2Cl2,
0 °C; (ii) 33% HBr in AcOH, 110 °C; (iii) NaH, 1.1 equiv, diethyl
acetamidomalonate, 1.1 equiv, DMF, room temp; (iv) NaCN, DMF or
DMSO, 80 °C; (v) 48% HBr, reflux.
Scheme 3a
a (i) NBS, benzoyl peroxide, CCl4, reflux; (ii) (1) 48% HBr, reflux, (2) H2O, reflux.
7291
dx.doi.org/10.1021/jm200862h |J. Med. Chem. 2011, 54, 7289–7298

Journal of Medicinal Chemistry
ARTICLE
Scheme 4a
Gly-Thr, and residues 632À775 from segment S2 (numbering
without signal peptide). The complex was crystallized as a dimer
with the ligand (S)-3h observed in both binding clefts. In both
molA and molB all residues could be modeled, as well as Ala from
the cloning remnant in the N-terminal end of molA. Experi-
mental procedures including a summary of data collection and
structure refinements are presented in Supporting Information.
The binding mode of (S)-3h to the GluA2 receptor construct
(Figure 5a) to some extent resembles that of other amino acid
based competitive antagonists.15,26 The α-carboxyl and α-amino
groups of (S)-3h are observed at exactly the same position in the
GluA2 binding cleft and show the same type of hydrogen
bonding interactions as in case of (S)-ATPO15 and as reported
more recently for two willardiine-based antagonists.26 However,
the position of the carboxyethyl chain of (S)-3h differs from that
observed for phosphonate in the GluA2-S1S2J:(S)-ATPO com-
plex with a bent conformation of the carboxyethyl side chain that
places the distal carboxylate group in a relatively wide hydro-
philic pocket in the vicinity of Thr655, Thr686, and Tyr702
(Figure 5b). Apart from the direct hydrogen bond with Thr655,
the distal carboxylate group is involved in interactions with sur-
rounding residues, mediated by a water-mediated H-bonding
network (Figure 5a). The electron density of the carboxyethyl
moiety is weak and thus indicates flexibility of the substituent, even
when bound to the receptor (Figure 5a). This flexibility results in a
slightly different conformation of the ligand in molA and molB.
The nitro substituent of (S)-3h is accommodated by a cavity
formed by Thr686, Leu704, Glu705, and Met708 (Figure 5b),
directly interacting, by one of its oxygen atoms, with the back-
bone NH of Glu705 (Figure 5a). The same oxygen atom makes a
weak hydrogen bond (3.4 Å) to Thr686 and Tyr702 through a
water molecule that is engaged in H-bond formation between the
carboxylate of (S)-3h and the side chain hydroxyl group of
Tyr702 (Figure 5a). Apparently, the second oxygen atom of the
nitro group is not involved in any hydrogen bond, which might
explain why no electron density is observed for this atom.
The chlorine atom of (S)-3h, located in a pocket between
residues of the domains D1 and D2 (Figure 5b), together with
the nitro group forces the Met708 side chain away from the
binding cleft because of otherwise steric repulsions. In addition,
the presence of both chlorine and nitro substituents prohibits
interdomain “locking”15 by preventing formation of the impor-
tant hydrogen bond between Glu402 in D1 and Thr686 in D2
(distance is 4.5 Å, Figure 5a), which consequently prevents
domain closure.
’ DISCUSSION
a (i) (1) MEM-chloride, DIPEA, CH2Cl2, rt; (2) TiCl4, CH2Cl2, 0 °C;
(ii) 33% HBr in AcOH, 110 °C; (iii) NaH, 1.1 equiv, diethyl acetami-
domalonate, 1.1 equiv, DMF, room temp; (iv) NaCN, DMF, 80 °C;
(v) 48% HBr, reflux.
A new series of phenylalanine-based AMPA receptor antago-
nists has been designed from the structure of the isoxazole-based
acidic amino acid (S)-ATPO and the X-ray structure of (S)-
ATPO cocrystallized with GluA2-S1S2J. Replacing the five-
membered isoxazole heterocycle with a larger phenyl spacer
between the alanyl part and the distal acidic functionality has
allowed for a study of the effect of introducing substituents pointing
into different areas of the binding pocket. This new series of
compounds shows a structureÀactivity relationship with some
similarities to, but also some differences from, that observed for
the well-studied isoxazole-based antagonists.27
Application of 600 μM (S)-3h plus 100 μM cyclothiazide did
not produce any detectable agonist response at GluA2(Q)i
(Figure 4). In some experiments, at very negative holding
potentials, extremely tiny responses to 3h were observed yielding
an estimated efficacy of <0.005 compared to that for (S)-Glu.
X-ray Crystallographic Analysis. The X-ray structure of the
compound (S)-3h bound to GluA2-S1S2J was solved to 1.9 Å
resolution. The GluA2-S1S2J construct comprises a Gly-Ala
cloning remnant, amino acid residues 392À506 from segment
S1 of the membrane-bound receptor, a two amino acid linker
In order to analyze in detail structural factors influencing the
pharmacological profile of the presented series of 2-carboxyethyl-
phenylalanines, the crystal structure was solved for GluA2-S1S2J
7292
dx.doi.org/10.1021/jm200862h |J. Med. Chem. 2011, 54, 7289–7298

Journal of Medicinal Chemistry
ARTICLE
Scheme 5a
a (i) 100% HNO3, 96% H2SO4, CH2Cl2, 0 °C; (ii) 48% HBr, reflux; (iii) Fe, 35% HCl, ethanol, water, 90 °C; (iv) (1) 30% HCl, NaNO2, À10 °C, (2)
CuSO4, KCN, aqueous NH3, 0 °C.
Table 1. Binding Pharmacology at Native Rat Brain Mem-
branes Using [3H]AMPA as Radioliganda
Table 2. Binding Pharmacology of 3h at Cloned rat GluA1À4
Expressed in Sf9 Insect Cell Membranes or As Soluble Ligand
Binding Domain Constructs (S1S2J)a
receptor
Ki (μM)
nH
GluA1o
6.87 ( 0.62
15.5 ( 1.1
4.36 ( 0.53
3.15 ( 0.64
3.06 ( 0.39
22.7 ( 3.2
20.4 ( 3.4
0.92 ( 0.04
1.13 ( 0.23
0.92 ( 0.07
1.02 ( 0.07
1.15 ( 0.03
0.92 ( 0.10
0.85 ( 0.03
(Y716F)GluA1o
GluA2(R)o
GluA2-S1S2J
(Y702F)GluA2-S1S2J
GluA3o
compd
R1
R2
R3
IC50 (μM)
3a
3b
3c
3d
3e
3f
H
H
H
H
H
Cl
H
H
H
H
H
H
H
H
Cl
H
H
H
>100
CH3
74 ( 7
>100
GluA4o
a Mean ( SEM are shown from at least three separate experiments,
conducted in triplicate at 16 ligand concentrations. nH is the Hill
coefficient.
C(CH3)3
Cl
61 ( 4
63 ( 4
19 ( 1
>100
CH2OH
Cl
Cl
Cl
Cl
Cl
3g
3h
3i
nitrogen atoms. However, the position of the carboxyethyl chain
of (S)-3h differs from that observed for the phosphonate moiety
in the GluA2-S1S2J:(S)-ATPO complex. Instead of direct con-
tacts with residues of D2, the distal carboxylate of (S)-3h
interacts directly only with Thr655 and is also involved in an
extensive network of hydrogen bonds formed with the receptor
protein through water molecules.
NO2
3.4 ( 0.3
>100
NH2
3j
COOH
>100
a Mean ( SEM values are indicated for selected compounds from at least
three separate experiments, conducted in triplicate.
The chlorine atom in the R2 position of (S)-3h corresponds to
the cyano group of CNQX and tert-butyl group of (S)-ATPO
(Figure 6); all these groups fill the same, partly polar partly
hydrophobic cavity formed by residues of D1 and D2. The nitro
substituent of (S)-3h nearly occupies the same position in the
binding cleft as the nitro group in CNQX. In contrast to CNQX,
however, one of the oxygen atoms of the nitro substituent of (S)-
3h does not interact directly with Thr686 but makes a weak
hydrogen bond with a water molecule which mediates contacts
with Thr686 and Tyr702.
Analysis of available X-ray structures of GluA2-S1S2J with
various ligands shows that, relative to the apo structure,29 full
agonists such as (S)-Glu and (S)-AMPA induce full domain closure
(approximately 21°),29 while partial agonists such as kainic acid and
bound to the most active compound, (S)-3h. During the past
decade several X-ray complexes of GluA2-S1S2J have been
reported28 in which the receptor construct has been cocrystal-
lized together with a number of structurally different amino acid
based15,26 or quinoxalinedione-related29À32 competitive antago-
nists, such as 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX).
Comparison of the binding interactions of (S)-ATPO and
CNQX with the GluA2 binding pocket suggests that (S)-3h
resembles a hybrid of (S)-ATPO and CNQX (Figure 6). The α-
carboxylate and α-ammonium groups of (S)-3h interact through
hydrogen bonds and ion-pair interactions with the protein in the
same way as described for (S)-ATPO, reproduced in the case of
CNQX by two carbonyl groups and one of the quinoxaline
7293
dx.doi.org/10.1021/jm200862h |J. Med. Chem. 2011, 54, 7289–7298

Journal of Medicinal Chemistry
ARTICLE
Table 3. TEVC Pharmacology of 3h at Cloned Rat AMPA
Receptors Expressed in X. laevis Oocytesa
Kb (μM)
receptor
(RS)-3h
(S)-3h
(R)-3h
GluA2(Q)i
8.59 ( 0.70 (12)
20.4 ( 1.4 (9)
3.43 ( 0.42 (9)
57.7 ( 19.0 (3)
GluA3i
nd
nd
a Shown are mean ( SEM values of duplicate measurements at 8À10
concentrations; number of experiments is in parentheses. nd, not determined.
Figure 4. Lack of agonist response by (S)-3h at GluA2(Q)i under
nondesensitizing conditions (100 μM cyclothiazide, CTZ). Vh = À30 mV.
Drug application duration is 30À45 s.
Figure 2. Antagonism of (S)-Glu (3 μM) responses at GluA2(Q)i and
GluA3i by (RS)-3h. Shown are results from representative TEVC
experiments, conducted with duplicate stimulations. The control re-
sponse is 3 μM (S)-Glu in the absence of antagonist. Vh = À45 mV
(GluA2(Q)i); Vh = À90 mV (GluA3i).
(6À8°),30 (S)-3h holds the binding cleft of GluA2 in a slightly
tighter conformation, showing a domain closure in close range to
those induced by partial agonists. In light of this observation it
should be emphasized that no clear sign of agonism was observed
for (S)-3h in electrophysiological studies performed under
conditions where desensitization was blocked with 100 μM
cyclothiazide (Figure 4), although a very low efficacy partial
agonism cannot be completely ruled out. In this context, it is interesting
to note that CNQX has been shown to be a partial agonist in the
presence of transmembrane AMPA receptor regulatory proteins.30
Further comparison of the binding modes of (S)-ATPO,
CNQX, and (S)-3h shows that the nitro group in the R1 position
and the chlorine atom in the R2 position interfere with the
D1ÀD2 interdomain hydrogen bond between Glu402 and
Thr686, preventing the formation of this D1ÀD2 lock in a way
similar to CNQX. It has been shown that the D1ÀD2 interdomain
lock contact plays an important role in the agonist-induced closed
conformation.13,33 In the crystal structure of (S)-3h the stabilization
of an open LBD conformation is reinforced by direct, as well as
water-mediated, interactions of the distal carboxylate group of (S)-
3h with residues of D2 (Thr655, Thr686, and Tyr702), similar to
interactions seen in case of (S)-ATPO.15
Using the X-ray structure of (S)-3h in complex with GluA2-
S1S2J for a structureÀaffinity analysis reveals possible explana-
tions for the observed binding affinities. In the first series of
compounds with both R1 and R3 positions of the phenyl ring
unsubstituted (3aÀe, Table 1), it appears that substituents of
intermediate size such as methyl (3b), chloro (3d), and hydro-
xymethyl (3e) are equally well accommodated by the receptor.
The chloro substituent in position R2 of 3h is nearly overlying
the tert-butyl group of (S)-ATPO (Figure 6), but whereas the
tert-butyl of (S)-ATPO appears to be optimal for AMPA receptor
affinity,15,27 the same bulky substituent of 3c appears to be too
large (Table 1).
Figure 3. ConcentrationÀresponse curves of (S)-Glu in the absence
(control) and presence of 10 μM (RS)-3h in an oocyte expressing
GluA2(Q)i, which indicate competitive antagonism. Currents are ex-
pressed as a percentage of the maximal control response. Stimulations
were carried out in duplicate at Vh = À60 mV: O, EC50 = 15 μM; b,
EC50 = 30 μM.
Br-HIBO induce domain closure in the range of 12À18°.13,29 On
the other hand, competitive antagonists, as CNQX and (S)-
ATPO, hold the LBD of GluA2 in an open conformation,
showing small domain closures (from 3À8°)15,29À31 or even small
domain opening as 8-methyl-5-(4-(N,N-dimethylsulfamoyl)phenyl)-
6,7,8,9,-tetrahydro-1H-pyrrolo[3,2-h]isoquinoline-2,3-dione-3-O-
(4-hydroxybutyrate-2-yl)oxime (NS1209) (5°).32 The domain
closure induced by (S)-3h is 9° in molA and 7° in molB.
This means that compared to (S)-ATPO (3À5°)15 and CNQX
7294
dx.doi.org/10.1021/jm200862h |J. Med. Chem. 2011, 54, 7289–7298

Journal of Medicinal Chemistry
ARTICLE
Figure 6. Superimposition of the binding mode of three competitive
antagonists. The structures of GluA2-S1S2J with (S)-3h (white), (S)-
ATPO (cyan), and CNQX (orange) have been superimposed on D1
residues (all molA). Water molecules (within 3.5 Å) are displayed in
colors according to the ligands.
likely explanation for the lack of affinity is related to the fact that a
carboxylate group in the R1-position is unfavorable in this part of
the receptor cavity because of a nearby carboxylate (Glu402);
thus, binding of 3j will result in chargeÀcharge repulsion. Finally,
substitution with an additional chlorine atom in the R3 position
(3g) is not tolerated by the receptor, probably because of too
close contacts in the region of Pro478 and Tyr450.
Despite a large number of AMPA receptor ligands that have
been synthesized and characterized pharmacologically, only few
of them have shown high AMPA receptor subunit selectivity.11,12,14
For agonists it has previously been claimed that such a selectivity
profile within AMPA receptors at least partly can be ascribed to
water-mediated interactions between the ligands and Tyr702 in
GluA2 (Tyr698 in GluA1).13 This interaction is unavailable in
GluA3 and GluA4, where the corresponding residue is Phe. For
this reason we decided to carry out detailed binding studies of
(S)-3h both to homomeric GluA1À4 receptors and to the
soluble constructs GluA2-S1S2J and (Y702F)GluA2-S1S2J mutant.
(S)-3h displayed a weak GluA1/GluA2 over GluA3/GluA4
selectivity. However, no difference in binding affinity to both
constructs was observed showing that this single residue alone
cannot account for the difference in binding affinity of 3h
between GluA1/GluA2 and GluA3/GluA4 full-length receptors.
In conclusion, we have identified a competitive AMPA receptor
antagonist, (S)-3h, that induces a domain closure of GluA2
greater than that observed for CNQX. Even though the domain
closure induced by (S)-3h is close to that observed for partial
agonists, such as KA, (S)-3h does not show any agonist response
at GluA2 receptors. However, (S)-3h and its interactions with
GluA2-S1S2J might be useful elements for future studies map-
ping out important molecular interactions and conformational
changes of the receptor protein for receptor activation and blockade.
Figure 5. Binding mode of (S)-3h in the ligand binding site of GluA2-
S1S2J. (A) Omit Fo À Fc electron-density map contoured at the 3σ level.
Polar contacts (dashed lines) to the ligand (S)-3h within 3.2 Å are shown
as well as the water-mediated network between the ligand and Tyr702.
Residues of GluA2-S1S2J (molA) are represented as white sticks and water
molecules as red spheres. (B) Surface of the ligand binding cavity (shown in
white), illustrating the flexibility of the cavity and the extra volume occupied
by water molecules. Water molecules within 3.5 Å from the ligand are shown.
Among the group of phenylalanine derivatives with R2 = Cl,
R3 = H and a variety of substituents in the R1 position (3d, 3f,
3hÀj, Table 1) it was observed that introducing an additional Cl
substituent (3f, R1 = Cl) or a nitro group (3h, R1 = NO2) was
beneficial for the affinity (Table 1), the nitro group being
superior. Whereas the nitro group of 3h is a H-bond acceptor
and contributes to favorable interactions with GluA2 residues
and water molecules (Figure 5a), the R1 chlorine atom of 3f is
also capable of accepting hydrogen bonds; however, the hydro-
gen bonds are expected to be longer and therefore weaker than
for the nitro group. The fact that compounds 3f and 3h accept
hydrogen bonds with a dissimilar spatial orientation might
contribute to the difference in affinity observed for the two com-
pounds. Compound 3d has a hydrogen atom instead of a nitro
group and is therefore unable to take part in a hydrogen bonding
network similar to 3h, which most likely explains the weaker
binding of 3d compared to 3f and 3h. Compounds with an amino
or a carboxy group in position R1 (3i and 3j, respectively) only
showed limited binding to native AMPA receptors (IC50 > 100 μM).
The aromatic amine in 3i can act as a hydrogen bond donor for
the water molecule but cannot function as a hydrogen bond
acceptor (from the backbone NH of Glu705). In the case of 3j, a
’ EXPERIMENTAL SECTION
Chemistry. General Procedures. All solvents and reagents
were purchased from commercial sources and used without further
purification. Analytical thin-layer chromatography (TLC) was performed
7295
dx.doi.org/10.1021/jm200862h |J. Med. Chem. 2011, 54, 7289–7298

Journal of Medicinal Chemistry
ARTICLE
on silica gel 60 F254 aluminum plates (Merck). All compounds were
detected using UV light and a KMnO4 spraying reagent. The target
amino acids were also visualized using a ninhydrin spraying reagent. 1H
NMR spectra were recorded on a 300 MHz Varian Mercury 300BB or
300 MHz Varian Gemini 2000BB spectrometer. D2O or DMSO-d6
(dimethyl-d6-sulfoxide) were used as solvents. Chemical shifts (δ) are
given in parts per million (ppm), and coupling constants (J) are given in
hertz (Hz). The following abbreviations were used: br s = broad singlet,
s = singlet, d = doublet, dd = double doublet, t = triplet, m = multiplet.
Elemental analyses were performed by J. Theiner, Microanalytical
Laboratory, Department of Physical Chemistry, University of Vienna,
Austria, and are within (0.5% of the theoretical values, corresponding to
g95% purity. CD spectra of (R)- and (S)-3h were recorded at ambient
temperature in 1.0 cm cuvettes on an OLIS DSM (digital subtractive
method) 10 CD spectrophotometer dissolved in 1 M HCl/ethanol (1:1).
HPLC Procedures. Preparative chromatography was performed
using HPLC system consisting of a Jasco 880-PU pump, a Rheodyne
7125 injector equipped with a 5.0 mL sample loop, a Shimadzu SPD-6A
detector (220 or 250 nm), and a Merck-Hitachi D-2000 chromato-
integrator. Purification of target amino acids was performed on a
reversed-phase Xterra MS C18 column, 10 μm, 10 mm  300 mm,
Waters. The column was isocratically eluted at 8 mL/min with a mixture
of methanol and 15 mM aqueous AcOH, the concentration of methanol
varying depending on the structure of the target amino acid. Preparative
resolution of the racemic 3h was performed using the Sumichiral OA-
5000 chiral ligand exchange column (10 mm  250 mm) equipped with
a Sumichiral OA-5000 guard column (10 mm  4 mm, Sumika Chemical
Analysis Service, Japan). The column was eluted at 4 mL/min with a
mixture of 2-propanol (15% v/v) and an aqueous solution of ammonium
acetate (50 mM), adjusted to pH 4.7 with glacial AcOH and containing
copper(II) acetate (0.1 mM). The temperature of column was kept at
60 °C. Determination of the enantiomeric excess (ee) of (R)- and (S)-3h
was performed on a Chirobiotic T column (4.6 mm  150 mm, ASTEC)
equipped with a Chirobiotic T guard column (4.6 mm  50 mm,
ASTEC). Elution was performed with 0.5 mL/min of a mixture of water
adjusted to pH 4 with AcOH and ethanol (40:60).
C, 65.51; H, 7.90; N, 4.77. Found: C, 65.18; H, 7.84; N, 4.71. 1H NMR
(300 MHz, DMSO-d6): δ 1.27 (s, 9H, CH3), 2.45À2.55 (m, 2H, CH2),
2.75À2.94 (m, 3H, CH2 and CHH), 3.25 (dd, J1 = 14.7 Hz, J2 = 4.5 Hz,
1H, CHH), 3.43 (dd, J1 = 7.8 Hz, J2 = 4.5 Hz, 1H, CH), 7.12 (d, J = 8.1
Hz, 1H, Ar), 7.20 (dd, J1 = 8.1 Hz, J2 = 2.1 Hz, 1H, Ar), 7.29 (d, J = 1.8
Hz, 1H, Ar). 13C NMR (75 MHz, D2O, as a hydrobromide salt): δ 28.7,
33.2, 35.9, 36.5, 37.2, 56.3, 127.8, 130.3, 131.5, 134.2, 138.6, 152.7,
173.7, 179.7.
(RS)-2-Amino-3-(2-(2-carboxyethyl)-4,5-dichlorophenyl)-
propionic Acid (3f). Yield 96 mg, 63%. Anal. Calcd for C12H13-
NO4Cl2: C, 47.08, H, 4.28, N, 4.58; Found: C, 46.99; H, 4.46; N, 4.46.
1H NMR (300 MHz, DMSO-d6): δ 2.50 (td, J1 = 7.6 Hz, J2 = 2.5 Hz, 2H,
CH2), 2.77À2.87 (m, 3H, CH2 and CHH), 3.16 (dd, J1 = 14.7 Hz, J2 =
5.4 Hz, 1H, CHH), 3.41 (dd, J1 = 8.2 Hz, J2 = 5.4 Hz, 1H, CH), 7.40 (s, 1H,
Ar), 7.45 (s, 1H, Ar). 13C NMR (75 MHz, D2O, as a hydrobromide salt): δ
26.0, 32.3, 34.1, 53.2, 129.7, 131.2, 131.7, 132.8, 139.7, 170.6, 176.6.
(RS)-2-Amino-3-(6-(2-carboxyethyl)-2,3-dichlorophenyl)-
propionic Acid (3g). Yield 115 mg, 75%. Anal. Calcd for C12H13-
NO4Cl2: C, 47.08, H, 4.28, N, 4.58; Found: C, 46.65; H, 4.44; N, 4.46.
1H NMR (300 MHz, DMSO-d6): δ 2.45 (t, J = 7.7 Hz, 2H, CH2), 2.85
(t, J = 7.7 Hz, 2H, CH2), 3.02 (dd, J1 = 13.8 Hz, J2 = 6.9 Hz, 1H, CHH),
3.27 (dd, J1 = 13.8 Hz, J2 = 8.3 Hz, 1H, CHH), 3.44 (t, J = 7.6 Hz, 1H,
CH), 7.13 (d, J = 8.5 Hz, 1H, Ar), 7.38 (d, J = 8.5 Hz, 1H, Ar).
(RS)-2-Amino-3-(2-(2-carboxyethyl)-5-chlorophenyl)pro-
pionic Acid (3d). A mixture of 11d (395 mg, 1 mmol) and 48% HBr in
water (6 mL) was refluxed for 2 h and evaporated under reduced
pressure. Purification by preparative reversed-phase HPLC followed by
recrystallization from water afforded the target amino acid as a white
solid. Yield 173 mg, 64%. Anal. Calcd for C12H14ClNO4 0.1H2O: C,
3
52.70; H, 5.23; N, 5.12. Found: C, 52.68; H, 5.20; N, 5.36. 1H NMR (300
MHz, D2O): δ 2.64 (t, J = 7.5 Hz, 2H, CH2), 2.92 (t, J = 7.5 Hz, 2H,
CH2), 3.04 (dd, J1 = 14.7 Hz, J2 = 8.7 Hz, 1H, CHH), 3.31 (dd, J1 = 14.7
Hz, J2 = 6.8 Hz, 1H, CHH), 3.89 (dd, J1 = 8.7 Hz, J2 = 6.8 Hz, 1H, CH),
7.22À7.29 (m, 3H, Ar). 13C NMR (75 MHz, D2O): δ 26.4, 33.1, 34.8,
53.8, 128.4, 130.2, 131.1, 132.1, 134.5, 138.2, 171.3, 177.5.
(RS)-2-Amino-3-(2-(2-carboxyethyl)-5-(hydroxymethyl)phe-
nyl)propionic Acid (3e). A mixture of 12 (163 mg, 0.36 mmol) in
48% aqueous HBr (2 mL) was refluxed for 1 h, and the solvent was
removed in vacuo. Then water (2 mL) was added and the mixture was
refluxed for additional 2 h and evaporated. Purification by reversed-
phase HPLC afforded the target amino acid as a white solid. Yield 36 mg,
37%. Anal. Calcd for C13H17NO5: C, 58.47; H, 6.97; N, 5.57. Found: C,
58.97; H, 7.04; N, 5.29. 1H NMR (300 MHz, D2O): δ 2.45 (t, J = 7.5 Hz,
2H, CH2), 2.74 (t, J = 7.3 Hz, 2H, CH2), 2.85 (dd, J1 = 15 Hz, J2 = 8.8 Hz,
1H, CHH), 3.14 (dd, J1 = 14 Hz, J2 = 6.4 Hz, 1H, CHH), 3.70 (dd, J1 =
8.8 Hz, J2 = 6.0 Hz, 1H, CH), 4,34 (s, 2H, CH2), 6,99 (s, 1H, Ar), 7.03
(d, J = 8,7 Hz, 1H, Ar), 7.59 (d, J = 8.3 Hz, 1H, Ar).
Synthesis of the Compounds 3hÀj. Compounds 13À15
(0.5 mmol) were refluxed in 48% aqueous HBr and purified by the method
described for compound 3d to give the target amino acids 3hÀjas white solids.
(RS)-2-Amino-3-(2-(2-carboxyethyl)-5-chloro-4-nitrophe-
nyl)propionic Acid (3h). Yield 92 mg, 58%. Anal. Calcd for
Final Steps in the Synthesis of Compounds 3aÀc,f,g. A
mixture of 11 (0.5 mmol) and 48% HBr in water (3 mL) was refluxed for
2 h, then evaporated under reduced pressure. The residue was dissolved
in water (10 mL), washed with diethyl ether (10 mL), and the water
phase was evaporated in vacuo. The remaining solid was dissolved in a
mixture of ethanol and water (7:3), and the pH was raised to the
isoelectric point (pH 4À5) by adding propylene oxide. The precipitated
crystals were filtered, dried, and recrystallized from ethanolÀwater (3:1)
(3a) or from water (3b,c,f,g) to give white crystals of the target amino acid.
(RS)-2-Amino-3-(2-(2-carboxyethyl)phenyl)propionic Acid
(3a). Yield 51 mg, 43%. Anal. Calcd for C12H15NO4 0.65H2O: C,
3
57.89; H, 6.60; N, 5.63. Found: C, 57.86; H, 6.46; N, 5.51. 1H NMR (300
MHz, D2O): δ 2.41 (t, J = 7.2 Hz, 2H, CH2), 2.68 (t, J = 7.2 Hz, 2H,
CH2), 2.92 (dd, J1 = 14.7 Hz, J2 = 8.9 Hz, 1H, CHH), 3.17 (dd, J1 = 14.7
Hz, J2 = 6.5 Hz, 1H, CHH), 4.04 (dd, J1 = 8.9 Hz, J2 = 6.5 Hz, 1H, CH),
7.00À7.11 (m, 4H, Ar). 13C NMR (75 MHz, D2O): δ 26.5, 32.9, 34.4,
53.8, 127.1, 128.2, 129.9, 130.4, 132.1, 139.2, 171.3, 177.9.
C12H13ClN2O6 0.8H2O: C, 43.53; H, 4.44; N, 8.46. Found: C, 43.46;
3
H, 4.28; N, 8.33. 1H NMR (300 MHz, DMSO-d6): δ 2.56 (t, J = 7.7 Hz,
2H, CH2), 2.80À3.05 (m, 3H, CH2 and CHH), 3.27 (dd, J1 = 14.3 Hz,
J2 = 5.2 Hz, 1H, CHH), 3.52 (t, J = 7.4 Hz, 1H, CH), 7.60 (s, 1H, Ar),
7.88 (s, 1H, Ar). 13C NMR (75 MHz, DMSO-d6): δ 26.1, 32.1, 33.9,
52.1, 122.1, 125.7, 132.7, 140.4, 140.9, 146.2.
(RS)-2-Amino-3-(2-(2-carboxyethyl)-5-methylphenyl)pro-
pionic Acid (3b). Yield 84 mg, 67%. Anal. Calcd for C13H17NO4 0.8
3
H2O: C, 58.77; H, 7.06; N, 5.27. Found: C, 58.47; H, 6.97; N, 5.57. 1H
NMR (300 MHz, D2O): δ 2.04 (s, 3H, CH3), 2.42 (t, J = 7.4 Hz, 2H,
CH2), 2.70 (t, J = 7.4 Hz, 2H, CH2), 2.78 (dd, J1 = 14.0 Hz, J2 = 8.9 Hz,
1H, CHH), 3.10 (dd, J1 = 14.0 Hz, J2 = 6.2 Hz, 1H, CHH), 3.69 (dd, J1 =
8.9 Hz, J2 = 6.2 Hz, 1H, CH), 6.85 (s, 1H, Ar), 6.89 (d, J = 8.3 Hz, 1H,
Ar), 6.95 (d, J = 7.8 Hz, 1H, Ar). 13C NMR (75 MHz, D2O): δ 20.2, 26.3,
32.9, 34.8, 53.7, 129.0, 129.3, 131.0, 132.0, 136.0, 137.1, 171.2, 177.4.
(RS)-2-Amino-3-(5-tert-butyl-2-(2-carboxyethyl)phenyl)-
propionic Acid (3c). Yield 69 mg, 47%. Anal. Calcd for C16H23NO4:
(S)-2-Amino-3-(2-(2-carboxyethyl)-5-chloro-4-nitrophenyl)-
propionic Acid [(S)-3h] and (R)-2-Amino-3-(2-(2-carboxyethyl)-
5-chloro-4-nitrophenyl)propionic Acid [(R)-3h]. The chiral
separation of (RS)-3h was performed using a ligand-exchange Sumichir-
al OA500 HPLC column,18,19 as described in HPLC Procedures.
Appropriate fractions were pooled and evaporated. In order to remove
7296
dx.doi.org/10.1021/jm200862h |J. Med. Chem. 2011, 54, 7289–7298

Journal of Medicinal Chemistry
ARTICLE
the buffer salts, the residues were dissolved in 1 M HCl and purified by
reverse-phase FC eluted with methanol/15 mM AcOH. Recrystalliza-
tions from water gave isomers with sufficient purity. The 1H NMR data
of S- and R-enantiomers of 3h were identical with the data for (RS)-3h
reported above. (S)-3h. Yield 17%, 97% ee; Δε (205 nm) = 0.46 m2/mol.
methoxy)isoxazol-4-yl]propionic acid; (S)-Br-HIBO, (S)-2-ami-
no-3-(4-bromo-3-hydroxyisoxazol-5-yl)propionic acid; CNS, central
nervous system; CNQX, 6-cyano-7-nitroquinoxaline-2,3-dione;
CTZ, cyclothiazide; Glu, glutamate; GluA2-S1S2J, ligand bind-
ing domain of the GluA2 receptor subunit; iGluR, ionotropic
glutamate receptor; KA, kainic acid; LBD, ligand binding
domain; NMDA, N-methyl-D-aspartic acid; NS1209, 8-methyl-
5-(4-(N,N-dimethylsulfamoyl)phenyl)-6,7,8,9,-tetrahydro-1H-
pyrrolo[3,2-h]isoquinoline-2,3-dione-3-O-(4-hydroxybutyrate-
2-yl)oxime; TEVC, two-electrode voltage clamp; Vh, TEVC holding
potential
Anal. Calcd for C12H13ClN2O6 1H2O: C, 43.06; H, 4.52; N, 8.37.
3
Found: C, 43.23; H, 4.20; N, 8.15. (R)-3h. Yield 21%, 95% ee; Δε
(205 nm) = À0.44 m2/mol. Anal. Calcd for C12H13ClN2O6 1H2O: C,
3
43.06; H, 4.52; N, 8.37. Found: C, 43.28; H, 4.22; N, 8.30.
(RS)-2-Amino-3-(4-amino-2-(2-carboxyethyl)-5-chlorophe-
nyl)propionic Acid (3i). Yield 64 mg, 45%. Anal. Calcd for
C12H15N2O4Cl 0.35H2O: C, 49.19; H, 5.40; N, 9.56. Found: C, 49.03; H,
3
5.19; N, 9.43. 1H NMR (300 MHz, D2O): δ 2.65 (t, J = 7.2 Hz, 2H, CH2),
2.88 (t, J = 7.2 Hz, 2H, CH2), 2.97 (dd, J1 = 14.8 Hz, J2 = 9.1 Hz, 1H, CHH),
3.26 (dd, J1 = 14.8 Hz, J2 = 5.8 Hz, 1H, CHH), 3.86 (t, J = 8.8 Hz, 1H, CH),
6.86 (s, 1H, Ar), 7.19 (s, 1H, Ar); 13C NMR (75 MHz, DMSO-d6): δ 26.9,
32.9, 35.5, 55.1, 115.1, 115.8, 123.8, 130.3, 139.1, 143.0, 166.9, 173.7.
(RS)-4-(2-amino-2-carboxyethyl)-5-(2-carboxyethyl)-2-chlo-
robenzoic Acid (3j). Yield 55 mg, 35%. Anal. Calcd for C13H14-
’ REFERENCES
(1) Traynelis, S. F.; Wollmuth, L. P.; McBain, C. J.; Menniti, F. S.;
Vance, K. M.; Ogden, K. K.; Hansen, K. B.; Yuan, H.; Myers, S. J.;
Dingledine, R. Glutamate receptor ion channels: structure, regulation,
and function. Pharmacol. Rev. 2010, 62, 405–496.
(2) Bleakman, D.; Alt, A.; Nisenbaum, E. S. Glutamate receptors and
pain. Semin. Cell Dev. Biol. 2006, 17, 592–604.
(3) Black, M. D. Therapeutic potential of positive AMPA modula-
tors and their relationship to AMPA receptor subunits. A review of
preclinical data. Psychopharmacology 2005, 179, 154–163.
(4) Kew, J. N.; Kemp, J. A. Ionotropic and metabotropic glutamate
receptor structure and pharmacology. Psychopharmacology 2005, 179, 4–29.
(5) Rao, V. R.; Finkbeiner, S. NMDA and AMPA receptors: old
channels, new tricks. Trends Neurosci. 2007, 30, 284–291.
NO6Cl 1.4H2O: C, 45.80; H, 4.97; N, 4.11. Found: C, 45.46; H, 4.62;
3
N, 4.03. 1H NMR (300 MHz, D2O): δ 2.73 (t, J = 7.4 Hz, 2H, CH2), 2.99
(t, J=7.4Hz, 2H, CH2), 3.13 (dd, J1 = 14.5 Hz, J2 =8.5Hz, 1H, CHH), 3.41
(dd, J1 = 14.5 Hz, J2 = 5.8 Hz, 1H, CHH), 4.03 (t, J = 8.5 Hz, 1H, CH), 7.38
(s, 1H, Ar), 7.55 (s, 1H, Ar). 13C NMR (75 MHz, DMSO-d6): δ 26.4, 33.4,
34.9, 54.3, 128.9, 130.5, 130.9, 131.6, 138.6, 140.0, 166.9, 169.6, 173.4.
’ ASSOCIATED CONTENT
(6) Hume, R. I.; Dingledine, R.; Heinemann, S. F. Identification of a
site in glutamate receptor subunits that controls calcium permeability.
Science 1991, 253, 1028–1031.
(7) Lomeli, H.; Mosbacher, J.; Melcher, T.; Hoger, T.; Geiger, J. R.;
Kuner, T.; Monyer, H.; Higuchi, M.; Bach, A.; Seeburg, P. H. Control of
kinetic properties of AMPA receptor channels by nuclear RNA editing.
Science 1994, 266, 1709–1713.
S
Supporting Information. Synthesis procedures and
b
NMR data for all compounds except compounds investigated
pharmacologically, experimental procedures for in vitro pharma-
cology, and results from X-ray crystallography. This material is
available free of charge via the Internet at http://pubs.acs.org.
(8) Fleming, J. J.; England, P. M. Developing a complete pharma-
cology for AMPA receptors: a perspective on subtype-selective ligands.
Bioorg. Med. Chem. 2010, 18, 1381–1387.
Accession Codes
†PDB code for GluA2 ligand-binding domain (S1S2J) in com-
plex with the antagonist (S)-2-amino-3-(2-(2-carboxyethyl)-5-
chloro-4-nitrophenyl)propionic acid [(S)-3h] is 3TZA.
(9) Jane, D. E.; Hoo, K.; Kamboj, R.; Deverill, M.; Bleakman, D.;
Mandelzys, A. Synthesis of willardiine and 6-azawillardiine analogs:
pharmacological characterization on cloned homomeric human AMPA
and kainate receptor subtypes. J. Med. Chem. 1997, 40, 3645–3650.
(10) Greenwood, J. R.; Mewett, K. N.; Allan, R. D.; Martin, B. O.;
Pickering, D. S. 3-Hydroxypyridazine 1-oxides as carboxylate bioisos-
teres: a new series of subtype-selective AMPA receptor agonists.
Neuropharmacology 2006, 51, 52–59.
’ AUTHOR INFORMATION
Corresponding Author
*Phone: (+45) 35 33 64 12. Fax: (+45) 35 33 60 01. E-mail:
tnj@farma.ku.dk.
(11) Coquelle, T.; Christensen, J. K.; Banke, T. G.; Madsen, U.;
Schousboe, A.; Pickering, D. S. Agonist discrimination between AMPA
receptor subtypes. NeuroReport 2000, 11, 2643–2648.
(12) Banke, T. G.; Greenwood, J. R.; Christensen, J. K.; Liljefors, T.;
Traynelis, S. F.; Schousboe, A.; Pickering, D. S. Identification of amino
acid residues in GluR1 responsible for ligand binding and desensitiza-
tion. J. Neurosci. 2001, 21, 3052–3062.
Present Addresses
Department of Technology and Biotechnology of Drugs,
Jagiellonian University Medical College, 9 Medyczna Street,
30-688 Krakow, Poland.
Author Contributions
#E.S. and K.F. contributed equally to this work and are joint first
(13) Hogner, A.; Kastrup, J. S.; Jin, R.; Liljefors, T.; Mayer, M. L.;
Egebjerg, J.; Larsen, I. K.; Gouaux, E. Structural basis for AMPA receptor
activation and ligand selectivity: crystal structures of five agonist complexes
with the GluR2 ligand-binding core. J. Mol. Biol. 2002, 322, 93–109.
(14) Bjerrum, E. J.; Kristensen, A. S.; Pickering, D. S.; Greenwood,
J. R.; Nielsen, B.; Liljefors, T.; Schousboe, A.; Brauner-Osborne, H.;
Madsen, U. Design, synthesis, and pharmacology of a highly subtype-
selective GluR1/2 agonist, (RS)-2-amino-3-(4-chloro-3-hydroxy-5-iso-
xazolyl)propionic acid (Cl-HIBO). J. Med. Chem. 2003, 46, 2246–2249.
(15) Hogner, A.; Greenwood, J. R.; Liljefors, T.; Lunn, M. L.;
Egebjerg, J.; Larsen, I. K.; Gouaux, E.; Kastrup, J. S. Competitive
antagonism of AMPA receptors by ligands of different classes: crystal
structure of ATPO bound to the GluR2 ligand-binding core, in
comparison with DNQX. J. Med. Chem. 2003, 46, 214–221.
authors.
’ ACKNOWLEDGMENT
The work was supported by the Danish Medical Research
Council, the Alfred Benzon Foundation, GluTarget, Dansync,
and Polish State Committee for Scientific Research. ESRF,
Grenoble, France, is thanked for providing beamtime.
’ ABBREVIATIONS USED
AMPA, (RS)-2-amino-3-(3-hydroxy-5-methylisoxazol-4-yl)pro-
pionic acid; (S)-ATPO, (S)-2-amino-3-[5-tert-butyl-3-(phosphono-
7297
dx.doi.org/10.1021/jm200862h |J. Med. Chem. 2011, 54, 7289–7298

Journal of Medicinal Chemistry
ARTICLE
(16) Hamilton, G. S. Omega-[2-(alkyl)phenyl]-2-aminoalkanoic
Acids as Antagonists of Excitatory Amino Acids Receptors. US 5,364,
876, 1994.
(33) Robert, A.; Armstrong, N.; Gouaux, J. E.; Howe, J. R. AMPA
receptor binding cleft mutations that alter affinity, efficacy, and recovery
from desensitization. J. Neurosci. 2005, 25, 3752–3762.
(17) Hamilton, G. S.; Huang, Z.; Patch, R. J.; Guzewska, M. E.;
Elliott, R. L.; Borosky, S. A.; Bednar, D. L.; Ferkany, J. W.; Karbon, E. W.
Phosphonoethylphenylalanine derivatives as novel antagonists of non-
NMDA ionotropic glutamate receptors. Bioorg. Med. Chem. Lett. 1992,
2, 1269–1274.
(18) Oi, N.; Kitahara, H.; Kira, R. Direct separation of enatiomers by
high-performance liquid chromatography on a new chiral ligand-
exchange phase. J. Chromatogr., A 1992, 592, 291–296.
(19) Miyazawa, T.; Minowa, H.; Imagawa, K.; Yamada, T. Enantio-
meric separation of non-protein amino acids by chiral ligand-exchange
high-performance liquid chromatography. Anal. Lett. 1997, 30, 867–882.
(20) Armstrong, D. W.; Liu, Y.; Ekborg-Ott, K. H. A covalently
bonded teicoplanin chiral stationary phase for HPLC enantiosepara-
tions. Chirality 1995, 7, 474–497.
(21) Berthod, A.; Liu, Y.; Bagwill, C.; Armstrong, D. W. Facile liquid
chromatographic enantioresolution of native amino acids and peptides using a
teicoplanin chiral stationary phase. J. Chromatogr., A 1996, 731, 123–137.
(22) Møller, E. H.; Egebjerg, J.; Brehm, L.; Stensbøl, T. B.; Johansen,
T. N.; Madsen, U.; Krogsgaard-Larsen, P. Resolution, absolute stereo-
chemistry, and enantiopharmacology of the GluR1À4 and GluR5
antagonist 2-amino-3-[5-tert-butyl-3-(phosphonomethoxy)-4-isoxazolyl]
propionic acid. Chirality 1999, 11, 752–759.
(23) Madsen, U.; Brauner-Osborne, H.; Frydenvang, K.; Hvene, L.;
Johansen, T. N.; Nielsen, B.; Sanchez, C.; Stensbøl, T. B.; Bischoff, F.;
Krogsgaard-Larsen, P. Synthesis and pharmacology of 3-isoxazolol
amino acids as selective antagonists at group I metabotropic glutamic
acid receptors. J. Med. Chem. 2001, 44, 1051–1059.
(24) Legrand, M.; Rougier, M. J. Application of the Optical Activity
to Stereochemical Determinations. In Determination of Configurations by
Dipole Moment, CD or ORD; Kaagan, H. B., Ed.; Stereochemistry:
Fundamentals and Methods, Vol. 2; Georg Thieme Publishers: Stutt-
gart, Germany, 1977; pp 33À183.
(25) Frandsen, A.; Pickering, D. S.; Vestergaard, B.; Kasper, C.;
Nielsen, B. B.; Greenwood, J. R.; Campiani, G.; Fattorusso, C.; Gajhede,
M.; Schousboe, A.; Kastrup, J. S. Tyr702 is an important determinant of
agonist binding and domain closure of the ligand-binding core of GluR2.
Mol. Pharmacol. 2005, 67, 703–713.
(26) Ahmed, A. H.; Thompson, M. D.; Fenwick, M. K.; Romero, B.;
Loh, A. P.; Jane, D. E.; Sondermann, H.; Oswald, R. E. Mechanisms of
antagonism of the GluR2 AMPA receptor: structure and dynamics of the
complex of two willardiine antagonists with the glutamate binding
domain. Biochemistry 2009, 48, 3894–3903.
(27) Madsen, U.; Bang-Andersen, B.; Brehm, L.; Christensen, I. T.;
Ebert, B.; Kristoffersen, I. T.; Lang, Y.; Krogsgaard-Larsen, P. Synthesis and
pharmacology of highly selective carboxy and phosphono isoxazole amino
acid AMPA receptor antagonists. J. Med. Chem. 1996, 39, 1682–1691.
(28) Pøhlsgaard, J.; Frydenvang, K.; Madsen, U.; Kastrup, J. S.
Lessons from more than 80 structures of the GluA2 ligand-binding
domain in complex with agonists, antagonists and allosteric modulators.
Neuropharmacology 2011, 60, 135–150.
(29) Armstrong, N.; Gouaux, E. Mechanisms for activation and
antagonism of an AMPA-sensitive glutamate receptor: crystal structures
of the GluR2 ligand binding core. Neuron 2000, 28, 165–181.
(30) Menuz, K.; Stroud, R. M.; Nicoll, R. A.; Hays, F. A. TARP
auxiliary subunits switch AMPA receptor antagonists into partial ago-
nists. Science 2007, 318, 815–817.
(31) Cruz, L. A.;Estebanez-Perpina, E.;Pfaff, S.; Borngraeber, S.; Bao, N.;
Blethrow, J.; Fletterick, R. J.; England, P. M. 6-Azido-7-nitro-1,4-dihydroqui-
noxaline-2,3-dione (ANQX) forms an irreversible bond to the active site of
the GluR2 AMPA receptor. J. Med. Chem. 2008, 51, 5856–5860.
(32) Kasper, C.; Pickering, D. S.; Mirza, O.; Olsen, L.; Kristensen,
A. S.; Greenwood, J. R.; Liljefors, T.; Schousboe, A.; Watjen, F.; Gajhede,
M.; Sigurskjold, B. W.; Kastrup, J. S. The structure of a mixed GluR2
ligand-binding core dimer in complex with (S)-glutamate and the
antagonist (S)-NS1209. J. Mol. Biol. 2006, 357, 1184–1201.
7298
dx.doi.org/10.1021/jm200862h |J. Med. Chem. 2011, 54, 7289–7298