(S)-2-amino-3-(3-hydroxy-7,8-dihydro-6H-cyclohepta[d]isoxazol-4-yl)propionic acid, a potent and selective agonist at the GluR5 subtype of ionotropic glutamate receptors. Synthesis, modeling, and molecular pharmacology

Article abstract of DOI:10.1021/jm0204441

We have previously described (RS)-2-amino-3-(3-hydroxy-7,8-dihydro-6H-cyclohepta[d]isoxazol- 4-yl)propionic acid (4-AHCP) as a highly effective agonist at non-N-methyl-D-aspartate (non-NMDA) glutamate (Glu) receptors in vivo, which is more potent than (RS

Full text of DOI:10.1021/jm0204441

1350
J . Med. Chem. 2003, 46, 1350-1358
(S)-2-Am in o-3-(3-h yd r oxy-7,8-d ih yd r o-6H-cycloh ep ta [d ]isoxa zol-4-yl)p r op ion ic
Acid , a P oten t a n d Selective Agon ist a t th e Glu R5 Su btyp e of Ion otr op ic
Glu ta m a te Recep tor s. Syn th esis, Mod elin g, a n d Molecu la r P h a r m a cology
Lotte Brehm,† J eremy R. Greenwood,† Kasper B. Hansen,§ Birgitte Nielsen,† J an Egebjerg,§ Tine B. Stensbøl,‡
Hans Bra¨uner-Osborne,† Frank A. Sløk,† Tine T. A. Kronborg,† and Povl Krogsgaard-Larsen*,†
Department of Medicinal Chemistry, The Danish University of Pharmaceutical Sciences, 2 Universitetsparken, DK-2100
Copenhagen, Denmark, Department of Molecular Pharmacology, H. Lundbeck A/ S, DK-2500 Valby, Denmark, and
Department for Molecular and Structural Biology, University of Aarhus, DK-8000 Aarhus C, Denmark
Received October 8, 2002
We have previously described (RS)-2-amino-3-(3-hydroxy-7,8-dihydro-6H-cyclohepta[d]isoxazol-
4-yl)propionic acid (4-AHCP) as a highly effective agonist at non-N-methyl-D-aspartate (non-
NMDA) glutamate (Glu) receptors in vivo, which is more potent than (RS)-2-amino-3-(3-hydroxy-
5-methylisoxazol-4-yl)propionic acid (AMPA) but inactive at NMDA receptors. However, 4-AHCP
was found to be much weaker than AMPA as an inhibitor of [3H]AMPA binding and to have
limited effect in a [3H]kainic acid binding assay using rat cortical membranes. To shed light
on the mechanism(s) underlying this quite enigmatic pharmacological profile of 4-AHCP, we
have now developed a synthesis of (S)-4-AHCP (6) and (R)-4-AHCP (7). At cloned metabotropic
Glu receptors mGluR1R (group I), mGluR2 (group II), and mGluR4a (group III), neither 6 nor
7 showed significant agonist or antagonist effects. The stereoisomer 6, but not 7, activated
cloned AMPA receptor subunits GluR1o, GluR3o, and GluR4o with EC50 values in the range
4.5-15 µM and the coexpressed kainate-preferring subunits GluR6 + KA2 (EC50 ) 6.4 µM).
Compound 6, but not 7, proved to be a very potent agonist (EC50 ) 0.13 µM) at the kainate-
preferring GluR5 subunit, equipotent with (S)-2-amino-3-(5-tert-butyl-3-hydroxyisothiazol-4-
yl)propionic acid [(S)-Thio-ATPA, 4] and almost 4 times more potent than (S)-2-amino-3-(5-
tert-butyl-3-hydroxyisoxazol-4-yl)propionic acid [(S)-ATPA, 3]. Compound 6 thus represents a
new structural class of GluR5 agonists. Molecular modeling and docking to a crystal structure
of the extracellular binding domain of the AMPA subunit GluR2 has enabled identification of
the probable active conformation and binding mode of 6. We are able to rationalize the observed
selectivities by comparing the docking of 4 and 6 to subtype constructs, i.e., a crystal structure
of the extracellular binding domain of GluR2 and a homology model of GluR5.
In tr od u ction
made up of GluR1-4 subunits and kainate receptors of
GluR5-7 and KA1,2 subunits.14,15 Whereas (S)-AMPA
(2, Figure 1) shows low or no affinity for kainate
receptors, kainic acid (1) exerts a nondesensitizing
response at AMPA receptors, resulting in large current
flows.16 The observed effects of 1 at native iGluRs in
some cases probably reflect activation of AMPA recep-
tors,17 and a number of questions regarding the phar-
macology of 1 and kainate receptors remain unan-
swered.
Glutamic acid (Glu) is the major excitatory amino acid
neurotransmitter in the central nervous system (CNS)
and is implicated in pathological processes leading to a
number of neurological and psychiatric diseases.1-3 The
transmitter functions of Glu are mediated by a hetero-
geneous group of receptors comprising both G-protein-
coupled metabotropic receptors (mGluRs) and ionotropic
receptors (iGluRs).4-6 On the basis of structural, func-
tional, and pharmacological characteristics, the eight
mGluRs7-9 are divided into three groups: group I
(mGluR1,5), group II (mGluR2,3), and group III
(mGluR4,6,7,8).8-10 The iGluRs are subdivided into
three pharmacologically distinct classes of recep-
tors, namely, N-methyl-D-aspartate (NMDA), (RS)-2-
amino-3-(3-hydroxy-5-methylisoxazol-4-yl)propionic acid
(AMPA), and kainate receptors.11,12 AMPA and kainate
receptors were previously termed collectively non-
NMDA receptors.13
Initial pharmacological studies indicated that the
AMPA analogue (RS)-2-amino-3-(5-tert-butyl-3-hydroxy-
isoxazol-4-yl)propionic acid (ATPA) was a relatively
weak AMPA receptor agonist.18 However, more recent
studies have revealed that ATPA is a selective and
highly potent agonist at homomerically expressed GluR5
receptors and that [3H]ATPA binds to cloned human
GluR5 with a Kd of 13 nM.19,20 These effects of ATPA
reside with the (S)-enantiomer 3,21 and (S)-2-amino-3-
(5-tert-butyl-3-hydroxyisothiazol-4-yl)propionic acid [(S)-
Thio-ATPA, 4] has recently been characterized as a
selective GluR5 agonist of even greater potency than
(S)-ATPA (3).22 Together with (S)-5-iodowillardiine (5,23,24
Figure 1), these compounds form a group of selective
and highly potent GluR5 agonist ligands possessing
The iGluRs are homo- or heteromeric assemblies of
subunits forming ion channels, AMPA receptors being
* To whom correspondence should be addressed. Phone: (+45) 35
30 65 11. Fax: (+45) 35 30 60 40. E-mail: ano@dfh.dk.
† The Danish University of Pharmaceutical Sciences.
§ University of Aarhus.
‡ H. Lundbeck A/S.
10.1021/jm0204441 CCC: $25.00 © 2003 American Chemical Society
Published on Web 03/06/2003

A Propionic Acid as Agonist
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 8 1351
Ta ble 1. Binding Data on Ionotropic Glutamate Receptors in
Rat Cortical Membranes
IC50 a (µM)
[3H]AMPA
[3H]kainic acid [3H]CGP39653
AMPAb 0.039 (0.037, 0.041)
>100
>100c
3b
4b
6
1.8 (1.6, 2.0)
23 (22, 24)
14 (11, 18)d
1.6 (1.3, 1.9)d
>100
>100c
0.27 (0.22, 0.34)
0.28 (0.25, 0.32)
33 (29, 38)
>100c
>100
7
28 (25, 32)e
a
Values are expressed as the antilog of the mean of the log of
at least three individual experiments. The numbers in parentheses
(min, max) indicate (SEM according to a logarithmic distribution.
b
d
Data from ref 22. c Tested in [3H]CPP binding assay. Bmax1 for
4 and 6 were 53% and 30%, respectively. e Ki value. Kd ) 6 nM
(from ref 55).
readily separable by column chromatography (CC). The
1H NMR chemical shifts of the tert-butyl groups of this
type of diastereomeric product containing a PEA moiety
have been empirically used to assign the stereochem-
istry at the R-amino acid centers.29,31,32 Thus, a reso-
nance signal for the tert-butyl groups in the region δ
1.3-1.4 ppm indicates (S,R)- or (R,S)-configurations,
whereas a signal at about δ 1.1 ppm indicates (S,S)- or
1
(R,R)-configurations. On the basis of H NMR spectro-
scopic analyses, compounds 18 (δ 1.39 ppm) and 19 (δ
1.12 ppm) were assigned the (2S)- and (2R)-configura-
tions, respectively. Compounds 18 and 19 were depro-
tected in two steps. Formic acid debenzylation provided
the respective (S)-isomer 20 (99.8% ee) and (R)-isomer
21 (99.5% ee) after recrystallization. Subsequently 20
and 21 were treated with a mixture of aqueous HCl and
AcOH followed by ion exchange. The fully deprotected
crude products of 6 (60% ee) and 7 (40% ee) showed
pronounced racemization. Preparative chiral HPLC of
these crude products gave pure 6 (99.9% ee) and 7
(98.8% ee). For a large number of R-amino acids, it has
been shown that the (S)-enantiomer is less retained
than the (R)-enantiomer on Chirobiotic T33 and Crown-
pak CR(-) columns.34 The elution order of 6 and 7 on
these two columns was in accordance with the assigned
configurations based on 1H NMR studies. The elution
order of 6 and 7 was reversed using a Crownpak CR(+)
column. CD spectra of 6 and 7 are mirror images
demonstrating the enantiomeric relationship. The re-
spective positive and negative Cotton effects observed
for 6 and 7 are in agreement with what has been
previously observed for R-amino acids containing an
isoxazole moiety, (S)- and (R)-enantiomers respectively
showing positive and negative Cotton effects of around
210-220 nm in acidic solutions.29,35-37
In Vitr o P h a r m a cology. The affinities of 6 and 7
for native AMPA, kainate, and NMDA receptors were
established using the radioligands [3H]AMPA, [3H]-
kainic acid, and [3H]CGP 39653, respectively (Table 1).
Compound 6 was found to be the eutomer at AMPA
receptors, displacing [3H]AMPA with relatively high
affinity (0.28 µM) and displacing [3H]kainic acid with
low affinity (Table 1). Compound 6 only displaced 70%
of the bound [3H]kainic acid, indicating that the com-
pound is able to discriminate between different [3H]-
kainic acid labeled receptor sites, as previously reported
for (S)-Thio-ATPA (4).22 Compound 7 showed very low
displacement of bound [3H]AMPA, [3H]kainic acid, and
[3H]CPG 39653.
F igu r e 1. Structures of glutamic acid, a number of agonists
at AMPA and kainate receptors, the eutomer (S)-4-AHCP (6),
and the distomer (R)-4-AHCP (7).
bulbous and relatively bulky substituents in equivalent
positions, considered to be an important structural
determinant for effective activation of GluR5 receptors.
The bicyclic amino acid (RS)-2-amino-3-(3-hydroxy-7,8-
dihydro-6H-cyclohepta[d]isoxazol-4-yl)propionic acid (4-
AHCP) has previously been described as a highly active
non-NMDA receptor agonist in vivo showing limited
affinity for [3H]AMPA or [3H]kainic acid binding sites.25,26
We here describe the synthesis of (S)-4-AHCP (6) and
(R)-4-AHCP (7) and the pharmacological characteriza-
tion of 6 as a selective and very potent agonist at cloned
GluR5 receptors. The results of molecular modeling and
docking experiments have enabled identification of the
probable active conformations and binding modes of 4
and 6 at a homology model of the GluR5 recognition site.
Resu lts
Ch em istr y. The syntheses of (S)-4-AHCP (6) and (R)-
4-AHCP (7) are outlined in Scheme 1. Compound 827
was protected using isopropyl bromide and potassium
carbonate in DMF to give the O-alkylated compound 9
(70-92%), contaminated by a minor amount of the
corresponding N-alkylated derivative 10. Chromic acid
oxidation of 9 gave the desired ketone 11 in 47% yield
together with 12 (27% yield). By reaction with triethyl
phosphonoacetate, ketone 11 was converted into a
mixture of isomeric esters 13 (61%) and a 1:1 mixture
of 14 and 15 (29%, from 1H NMR). Compound 13 was
converted into the isomeric ester 15 by treatment with
potassium tert-butoxide. Subsequent DIBAL-H reduc-
tion of 15 provided aldehyde 17 (72%). Under “Ugi four-
component condensation” reaction conditions,28-30 with
(R)-1-phenylethylamine [(R)-PEA] as a chiral auxiliary,
aldehyde 17 was converted into the (2S)- and (2R)-
diastereomeric compounds 18 and 19, respectively.
Compounds 18 (99.9% de) and 19 (99.5% de) were

1352 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 8
Brehm et al.
Sch em e 1a,b
a
Reagents and conditions: (a) (CH3)2CHBr, K2CO3, DMF; (b) Na2Cr2O7, AcOH, concentrated H2SO4; (c) EtOOCCH2PO3(Et); (d) t-BuOK,
DME; (e) DIBAL-H, toluene; (f) (R)-PEA, benzoic acid, tert-butylisonitrile, MeOH; (g) column chromatography; (h) 98% HCOOH; (i) 23%
b
aqueous HCl, AcOH; (j) chiral HPLC. Compound: 16 is the tert-butyl ester analogue of 15.
Ta ble 2. Electrophysiological Data on Oocytes Expressing Different Homomeric and Heteromeric AMPA and Kainate Receptorsa
3b
4b
6
EC50 (µM)
n
rel Imax
EC50 (µM)
n
rel Imax
EC50 (µM)
n
rel Imax
GluR1o
GluR3o
GluR4o
GluR5
22 (18, 27)
7.9 (6.6, 9.2)
7.6 (5.6, 9.6)
0.48 (0.44, 0.52) 1.0
nr
1.6
0.33 ( 0.11
5.2 (4.8, 5.6)
0.64 0.78 ( 0.026 4.5 (3.9, 5.3)
1.1 (1.0, 1.2) 0.67 ( 0.055
1.4 (1.4, 1.5) 1.04 ( 0.084
1.3 (1.3, 1.3) 1.64 ( 0.22
0.84 0.054 ( 0.002 32 (31, 33)
0.67 0.34 ( 0.051
0.90 1.2 ( 0.015
7.2 (6.6, 7.9)
15 (13, 17)
20 (15, 27)
0.10 (0.090, 0.12) 1.5
nr
0.8
1.7 ( 0.095
1.0(0.027
1.0 ( 0.083
0.13 (0.12, 0.14) 1.0 (1.0, 1.1) 1.61 ( 0.13
GluR6
nr
GluR6 + KA2 61 (58, 64)
2.0
0.17 ( 0.011
4.9 (4.6, 5.3)
1.9
0.41 ( 0.040 6.4 (5.8, 6.9)
0.9 (0.9, 1.0) 0.53 ( 0.0.064
a
Values are expressed as the antilog of the mean of the log of at least three individual experiments. The numbers in parantheses (min,
max) indicate (SEM according to a logarithmic distribution. The Imax values are relative to maximal steady-state currents evoked by
kainic acid for all the receptor combinations except GluR6 + KA2 heteromers, where the amplitudes are relative to the maximal currents
evoked by AMPA. (R)-4-AHCP (7) did not show any agonistic or antagonistic activity that could not be ascribed to contamination with
b
(S)-4-AHCP (6). nr ) no response at 100 µM. Data from ref 22.
Compounds 6 and 7 were tested for pharmacological
activity on mGluR1R, mGluR2, and mGluR4a express-
ing CHO cells, representing groups I, II, and III mGluR
receptors, respectively, using conventional second mes-
senger assays. At 1 mM concentrations, no agonist or
antagonist (inhibition of 30 µM Glu response) activity
was observed.
In Vitr o Electr op h ysiology. Compounds 6 and 7
were applied on two-electrode-voltage-clamped Xenopus
oocytes, each expressing one of the following: homo-
meric AMPA receptors composed of GluR1o, GluR3o, or
GluR4o; homomeric kainate receptors composed of
GluR5 or GluR6; heteromeric GluR6 + KA2 kainate
receptors (Table 2). Compound 6 was shown to be an
agonist at all of the AMPA receptor subtypes, with EC50
values ranging from 4.5 to 15 µM, being most potent at
GluR1o. At GluR5 receptors, however, compound 6 was
found to be 35-115 times more potent than at AMPA
receptors, having an EC50 of 0.13 µM and thus being
equipotent with 4 at GluR5. The amplitudes of the

A Propionic Acid as Agonist
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 8 1353
mitted to a series of restrained minimizations using the
impref module in Impact 1.8 (OPLS-AA force field).38
Dock in g. Both complexes (experimental kainic acid:
GluR2-S1S2J and the model of kainic acid:GluR5-
S1S2J ) were prepared according to the method recom-
mended prior to docking using Glide 1.8.38 Compounds
1, 4, and 6 were docked to both models. The results were
clear-cut, with all three ligands given highest scores and
lowest energies in binding poses corresponding to those
observed experimentally for similar ligands.40,42 As an
internal test, the four highest-scoring (Glidescore38)
poses of 1 docked to GluR2, which also had the four
lowest interaction energies (Emodel38), had rms devia-
tions from the crystal structure of around 0.5 Å or less.
In addition, the docked position of 4 in GluR2 closely
resembles the binding mode that has been depicted for
the experimental 4:GluR2-S1S2J complex,43 giving
confidence in Glide’s ability to also predict the binding
mode of 6. The results of docking 4 and 6 to GluR2 and
GluR5 are depicted in Figures 4 and 5, respectively.
F igu r e 2. Two views of compounds 1 (green), 4 (orange), and
6 (magenta) superimposed in low-energy conformations on the
basis of the five isosteric N and O atoms.
maximal steady-state currents (Imax) relative to the
maximal responses evoked by kainic acid (1) were
highest for GluR4o and GluR5 (Table 2). Application of
up to 100 µM of 7 resulted in very small responses that
could be ascribed to small amounts of 6 present.
Molecu la r Mod elin g. Liga n d An a lysis. Com-
pounds 1, 4, and 6, modeled as tri-ionized species, were
subjected to a Monte Carlo conformation search using
the MMFFs force field in Macromodel 8.0, including GB/
SA treatment of aqueous solvation,38 and were super-
imposed by least-squares fitting of the five isosteric
oxygen and nitrogen atoms to the known binding mode
of 1, which is also its global energy minimum (Figure
2). The minimum energy conformation of 6 matched 1
closely, apart from a rotation of the R-carboxylate group
of 6; rotating this group to match 1 costs 2.4 kJ /mol.
The seven-membered ring prefers to adopt a conforma-
tion anti to the R-amino acid side chain. As commonly
seen for flexible ionized ligands, intramolecular hydrogen-
bonded structures dominate the low-energy conformers
of 4 according to the mechanics potential; however, the
lowest energy extended conformation matched 1 (at 10.2
kJ /mol above the global minimum for 4).
Discu ssion
The bicyclic Glu homologue (RS)-2-amino-3-(3-hy-
droxy-7,8-dihydro-6H-cyclohepta[d]isoxazol-4-yl)propi-
onic acid (4-AHCP), derived from the standard AMPA
receptor agonist (S)-AMPA (2) (Figure 1), was originally
designed25 as a tool for indirect mapping of AMPA
receptor recognition sites.44 Using microelectrophoretic
techniques, 4-AHCP was shown to be a highly potent
agonist at non-NMDA receptors in vivo, markedly more
potent than AMPA.25 However, 4-AHCP displayed only
limited affinity for [3H]kainic acid25 and [3H]AMPA26
binding sites. For a number of years, attempts to
improve the very low-yield synthesis of 4-AHCP25 failed,
and furthermore lack of appropriate pharmacological
tools prevented studies on the mechanism(s) underlying
this quite enigmatic pharmacological profile of 4-AHCP.
We have previously shown that homologues of Glu
and Glu analogues typically show agonist or antagonist
effects at subtypes of mGluRs,29,45-47 encouraging the
investigation of 4-AHCP as a potential mGluR ligand.
In addition, cloned AMPA receptors (GluR1-4) and
kainate receptors (GluR5-7, KA1,2) are now available
for pharmacological characterization of Glu receptor
ligands. Prompted by these pharmacological advances
in the Glu receptor field and by the accumulating
evidence of the important role of kainate receptors17
including GluR5 receptors48 in CNS disorders such as
epilepsy, we decided to synthesize (S)-4-AHCP (6) and
(R)-4-AHCP (7) with a view to further pharmacological
characterization.
Hom ology Mod elin g. Sequences were first aligned
using BLAST,39 and an S1S2 sequence construct was
built for GluR5 corresponding to GluR2-S1S2J (53%
identity; higher within the binding site) (Figure 3). A
homology model of GluR5-S1S2J was built on the basis
of the 1:GluR2-S1S2J crystal structure40 using Swiss-
model, first approach mode.41 Compound 1 was rein-
troduced to the binding site. Because it was apparent
that Swissmodel had modeled the side chain of M733
in a conformation unfavorable for binding, a restrained
Monte Carlo search (MMFFs force field38) was per-
formed on this residue and the hydrophobic side chains
in contact with it. Once a suitable low-energy side chain
packing had been found, the entire complex was sub-
F igu r e 3. Comparison of binding site residues (bold) in GluR2 and GluR5, based on alignment of the S1S2 extracellular ligand
binding domains (53% identity). Numbering is derived from ref 40. “lock” ) agonist state interdomain hydrogen bond. “Pkt” )
binding site pocket. “-NH3+”, “-COO-”, and “distal anion” bind the respective charged moieties of the ligands.

1354 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 8
Brehm et al.
of this area of Glu receptor pharmacology, we have
carried out comparative molecular modeling studies on
compounds 1, 4, and 6, mapping out the agonist binding
site of GluR5 in detail for the first time.
Analysis of the ligands shows that although the
skeleton of 6 has one extra carbon between its R-amino
acid and distal anion moieties compared with Glu, 1 and
4, the side chain is held in an ideal orientation for
recognition by iGluRs (but not mGluRs) because of the
unsaturated seven-membered ring. In addition, 4 and
6 occupy remarkably similar volumes, in concord with
their similar pharmacological profiles. Although 3 also
occupies a similar volume, its isoxazole ring oxygen does
not protrude as far as the sulfur atom of 4, nor is it as
lipophilic.
F igu r e 4. Compounds 4 (orange) and 6 (magenta) docked onto
the X-ray crystal structure of the 1:GluR2-S1S2J complex.
With such a high degree of binding-site homology, we
can interpret the results of the docking study with
confidence. The residues recognizing the R-amino acid
and distal anion moieties of agonists at GluR2 and
GluR5 are identical. The receptors differ only in the size,
shape, and polarity of the cavity beside and “below” the
volume occupied by glutamate upon binding (Figures 4
and 5).40 Compounds 6 and 4 both possess moieties that
protrude below the plane of their heterocyclic rings,
whereas high-potency AMPA receptor agonists do not.42
As with 1, L650 in GluR2 experiences steric repulsion
from 6 and 4, limiting the degree to which the receptor
can close tightly around the ligands. Although the
relationship between affinity and the macroscopic do-
main closure measured from the crystal structures is
by no means linear,42 agonists that bind tightly and are
strongly desensitizing display tight domain closure.40,42
In addition to L650, 4 and 6 come close to M708 and
T686 in GluR2; the receptor must adapt upon domain
closure to accommodate these ligands, with a corre-
sponding drop in potency at GluR2 compared with
AMPA and AMPA agonists. Finally, a water molecule
observed experimentally in all agonist:GluR2 complexes
to date links Y702, T686, and the ligand by a hydrogen
bonding network42,50 (not shown in Figure 4). The
hydrophobic volumes of 6 and 4 destabilize this water
molecule compared to the effect of potent GluR2 ago-
nists. Ligand selectivity seems to be highly sensitive to
the ligand’s disposition toward the hydrogen bond
network in the direction of Y702,50 and the reduced
selectivity of 3 versus 4 may be linked to the positioning,
reduced size, and hydrophobicity of the isoxazole ring
oxygen versus sulfur. By contrast with GluR2, the cavity
in GluR5 beside and below the binding position of the
distal carboxylate of Glu is larger and more hydrophobic
overall. The shortening of L650 in GluR2 to V679 in
GluR5 allows for better van der Waals contact and
tighter domain closure around the bulkier ligands, as
does the replacement of T686 by S715. Interestingly,
not only does the substitution of M708 (GluR2) for S737
(GluR5) alter the shape of the cavity (more space around
the tert-butyl group of 4 or the seven-membered ring of
6) it also enables a new interdomain hydrogen bond
network through Y438 (GluR5) upon agonist binding,
either directly or through an extra water molecule (not
included in the homology model). At the same time the
volume between Y450, P478, Y723, and Y405, known
to accommodate heteroaromatic substituents in GluR2,42
seems to be reduced in size and hydrophobicity in
F igu r e 5. Compounds 4 (orange) and 6 (magenta) docked onto
a homology model of GluR5-S1S2J built from the X-ray crystal
structure of the 1:GluR2-S1S2J complex.
Interestingly however, neither 6 nor 7 showed detect-
able agonist or antagonist effects at mGluR1R, mGluR2,
or mGluR4a, representatives of group I, II, and III
mGluR, respectively, perhaps because the conforma-
tional constraints imposed by the seven-membered ring
do not allow 4-AHCP to mimic the extended binding
conformation of Glu at mGluRs. Whereas 7 turned out
to be essentially inactive at the iGluRs tested (Tables 1
and 2), 6 was shown to be a highly potent agonist at
GluR5, equipotent with (S)-Thio-ATPA (4)22 and more
potent than (S)-ATPA (3).21 As observed for 3 and 4
(Table 2), 6 showed much weaker agonist effects at the
homomeric AMPA receptors, GluR1o, GluR3o, and
GluR4o. The affinity of 6 for kainate receptor sites
labeled by [3H]kainic acid (IC50 ) 1.6 µM) (Table 1) is
somewhat higher than that reported previously for (RS)-
4-AHCP,25 but still this affinity is surprisingly low
compared to the very potent GluR5 agonist potency of
6 (Table 2).
(S)-4-AHCP (6) represents a new structural class of
GluR5 receptor agonists and may be an important lead
structure for the design of new GluR5 agonists and
antagonists. It is GluR5 antagonists that have attracted
particular interest as therapeutic agents.17,48,49 How-
ever, since kainate receptors such as GluR5 desensitize
very rapidly and since GluR5 expression is restricted
compared with other kainate subunits,14 the therapeutic
prospects are open for developing GluR5 agonists as
“functional antagonists”.24 To further the development

A Propionic Acid as Agonist
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 8 1355
CR(+) column (150 mm × 10.0 mm) equipped with a guard
column (10 mm × 4.0 mm) was performed at 30 °C using an
LKB 2155 HPLC column oven, and the products were eluted
at 6.0 mL/min. For both columns, products were eluted with
aqueous HClO4 (pH 2.0), and the columns were connected to
the same instrumentation as described above for the Chiro-
biotic T column with a detection wavelength of 200 nm. The
CR(+) and CR(-) columns were used for the determinations
of stereochemical purities of 6 and 7, respectively. The % ee
values were calculated from peak areas.
3-Isop r op oxy-5,6,7,8-tetr a h yd r o-4H-cycloh ep ta [d ]isox-
a zole (9) a n d 2-Isop r op yl-5,6,7,8-tetr a h yd r o-4H-cyclo-
h ep ta [d ]isoxa zolin -3-on e (10). Potassium carbonate (23.1
g, 167 mmol) dried in vacuo was added to a solution of 5,6,7,8-
tetrahydro-4H-cyclohepta[d]isoxazol-3-ol (8)27 (23.3 g, 152
mmol) in dry DMF (200 mL), and the reaction mixture was
left to stir for 30 min at 60 °C. Isopropyl bromide (21.3 mL,
227 mmol) was then added, and the resulting reaction mixture
was stirred at 60 °C for 3 h (followed by TLC). After the
mixture was cooled, water (150 mL) was added and the
mixture was extracted with CH2Cl2 (3 × 100 mL). The
combined organic phases were evaporated, and water (100 mL)
was added to the residue. Extraction with Et2O (3 × 100 mL),
drying, evaporation, and FC [light petroleum/AcOEt (9:1)] gave
9 (27 g, 92%) as an oil. A sample was distilled in a Kugelrohr
apparatus: bp 185-195 °C/1.3 kPa; 1H NMR δ 1.37 (6H, d, J
) 6.2 Hz), 1.62-1.84 (6H, m), 2.31 (2H, m), 2.74 (2H, m), 4.87
(1H, heptet, J ) 6.2 Hz); 13C NMR δ 20.6, 21.8, 25.6, 28.0,
28.3, 30.3, 72.5, 106.7, 169.7, 171.0. Anal. (C11H17NO2) C, H,
N. Different batches gave yields of 9 in the range 70-92%.
GluR5. Finally, replacement of Y702 by L731 in GluR5
means that instead of a water molecule bound in a fairly
rigid hydrogen bond network,40,42 GluR5 agonists face
a more hydrophobic and flexible tunnel in this direction.
In summary, the so-called hydrophobic pocket44 of
agonized GluR2 has changed shape and character in
GluR5 as well as shifting positionssmaller and less
hydrophobic on the left, larger and substantially more
hydrophobic below and on the right (Figures 4 and 5).
It would also seem that neither 4 nor 6 have exhausted
the hydrophobic space available within the binding
cavity. The current model may be of assistance in
further structure-based design of higher-affinity GluR5
specific agonists. In this context, (S)-4-AHCP (6) pre-
sents a new scaffold for the design of kainate subtype
specific ligands.
Exp er im en ta l Section
Ch em istr y. Gen er a l P r oced u r es. Analytical thin-layer
chromatography (TLC) was performed on silica gel F254 plates
(Merck). All compounds were detected using UV light and a
KMnO4 spraying reagent. Compounds containing the 3-isox-
azolol moiety were also visualized using an FeCl3 spraying
reagent. Compounds containing amino groups were also
visualized using a ninhydrin spraying reagent. Flash chro-
matography (FC) was performed on silica gel Matrex LC 60A
(0.035-0.070 mm). Column chromatography (CC) was per-
formed on silica gel Matrex LC 60A (0.070-0.200 mm). The
light petroleum used had a boiling point of 80-100 °C. Organic
phases were dried using MgSO4. Melting points were deter-
mined in capillary tubes and are uncorrected. 1H, 13C, and
associated proton test (APT) spectra were recorded on a Varian
Gemini-2000 BB spectrometer (300 MHz). Unless otherwise
stated, CDCl3 was used as a solvent. Chemical shifts are given
in ppm (δ) using TMS (1H NMR) or the CDCl3 peak [13C NMR
(δ 77.00)] as internal standards. Elemental analyses were
performed at Analytical Research Department, H. Lundbeck
A/S, Denmark, or by J . Theiner, Microanalytical Laboratory,
Department of Physical Chemistry, University of Vienna,
Austria, and are within (0.4% of the theoretical values, unless
otherwise stated. Optical rotations were measured in thermo-
stated cuvettes on a Perkin-Elmer 241 polarimeter. Circular
dichroism (CD) spectra were recorded in 0.1 M HCl solution
in 1.0 cm cuvettes at room temperature on a J asco J -720
spectropolarimeter.
Preparative chiral HPLC was performed on a Chirobiotic T
column (500 mm × 10.0 mm) equipped with a Chirobiotic T
guard column (50 mm × 4.6 mm). The column was connected
to an HPLC instrumentation consisting of a J asco 880 pump,
a Rheodyne 7125 injector equipped with a 5.0 mL loop, a
Shimadzu SPD-6A detector (220 nm), and a Hitachi D-2000
Chromato-Integrator. The column was eluted at 1.5 mL/min
with aqueous AcONH4/AcOH (15 mM, pH 4.0) and EtOH (60:
40). Preparative reversed-phase HPLC was performed using
a Vydac C18 column (250 mm × 22 mm; 10 µm), and the
products were eluted with AcOH (15 mM) and MeOH (80:20)
at 6.0 mL/min. The column was connected to the same
instrumentation as described above except using a TSP
detector. Achiral HPLC analyses of 18 and 19 were performed
on a Knauer Vertex Sperisorb ODS2 column (120 mm × 4.0
mm; 5 µm) using a Waters M510 pump, a Waters U6K injector,
and a Waters 991 photodiode array detector (190 nm). The
column was eluted at 1.0 mL/min with H2O and MeOH (30:
70). Chiral HPLC analyses of 20 and 21 were performed using
a chiral-AGP column (150 mm × 4.0 mm), and the products
were eluted with aqueous KH2PO4/K2HPO4 (50 mM, pH 7.0)
and 2-PrOH (80:20) at 0.5 mL/min. The column was connected
to the same instrumentation as described above for the Vydac
Further elutions gave in different batches yields of the
N-alkylated isomer 10 as an oil in the range trace-20%: 1H
NMR δ 1.30 (6H, d, J ) 6.6 Hz), 1.60-1.84 (6H, m), 2.36 (2H,
t, J ) 5.7 Hz), 2.61 (2H, t, J ) 5.6 Hz), 4.58 (1H, heptet, J )
6.6 Hz); 13C NMR δ 19.7, 21.3, 25.6, 27.7, 28.5, 30.3, 48.6, 110.9,
167.8, 169.3.
3-Isop r op oxy-5,6,7,8-tetr a h yd r ocycloh ep ta [d ]isoxa zol-
4-on e (11) a n d 3-Isop r op oxy-5,6,7,8-tetr a h yd r ocycloh ep -
ta [d ]isoxa zol-8-on e (12). To a solution of 9 (7.56 g, 38.7
mmol) in AcOH (99.9%, 10 mL) and concentrated H2SO4 (6
mL) kept at 0-10 °C was added, over 20 min, a solution of
Na2Cr2O7‚2H2O (17.3 g, 58.1 mmol) in AcOH (50 mL). After
stirring for 3 h at room temperature, the mixture was
neutralized by addition of concentrated NaOH. Extraction with
Et2O (4 × 200 mL), drying, evaporation, and CC [toluene/
AcOEt (9:1 f 6:1)] gave crude unreacted 9 (2 g, 26%). Further
elution gave 12 (2.2 g, 27%): mp 70.0-71.0 °C [toluene/light
petroleum]; 1H NMR δ 1.40 (6H, d, J ) 6.1 Hz), 1.96 (4H, m),
2.60 (2H, t, J ) 6.1 Hz), 2.77 (2H, m), 4.94 (1H, heptet, J )
6.1 Hz); 13C NMR δ 21.7, 21.9, 22.3, 25.7, 43.2, 73.9, 115.8,
160.3, 169.7, 190.2. Anal. (C11H15NO3) C, H, N.
Later fractions contained 11 (3.8 g, 47%): mp 64.5-66.0 °C
[toluene/light petroleum]; 1H NMR δ 1.43 (6H, d, J ) 6.1 Hz),
1.88-2.09 (4H, m), 2.72 (2H, m), 3.03 (2H, t, J ) 6.2 Hz), 4.94
(1H, heptet, J ) 6.1 Hz); 13C NMR δ 21.5, 22.3, 23.7, 28.5,
44.8, 74.0, 109.5, 169.1, 177.8, 194.1. Anal. (C11H15NO3) H, N.
C: calcd 63.14; found 62.55.
Eth yl (E)- a n d Eth yl (Z)-(3-Isop r op oxy-5,6,7,8-tetr a h y-
d r ocycloh ep ta [d ]isoxa zol-4-ylid en e)a ceta te (13 a n d 14)
a n d Eth yl (3-Isop r op oxy-7,8-d ih yd r o-6H-cycloh ep ta [d ]-
isoxa zol-4-yl)a ceta te (15). Triethyl phosphonoacetate (5.1
mL, 25.8 mmol) was added dropwise to a stirred solution of
potassium tert-butoxide (2.4 g, 21.5 mmol) in dry THF (20 mL),
and the mixture was stirred at room temperature for 30 min.
A solution of 11 (900 mg, 4.30 mmol) in dry THF (5 mL) was
then added. After the mixture was stirred for 22 h at 80 °C, a
saturated solution of ammonium chloride (10 mL) and H2O
(25 mL) was added. Extraction with AcOEt (3 × 100 mL),
drying, and evaporation followed by CC [toluene/AcOEt (9:1)]
gave 13 (730 mg, 61%) as an oil: 1H NMR δ 1.31 (3H, t, J )
7.1 Hz), 1.43 (6H, d, J ) 6.2 Hz), 1.88 (4H, m), 2.84 (2H, m),
3.12 (2H, m), 4.19 (2H, q, J ) 7.1), 4.97 (1H, heptet, J ) 6.2
Hz), 6.25 (1H, s); 13C NMR δ 14.1, 21.7, 24.3, 25.8, 26.7, 30.0,
59.6, 73.6, 109.0, 116.3, 147.2, 167.0, 167.8, 173.9. Anal.
column, though with
a detector wavelength of 200 nm.
Analytical chiral HPLC using a Crownpak CR(-) column (150
mm × 4.0 mm) was performed at room temperature with a
flow rate of 0.40 mL/min. Analytical HPLC using a Crownpak

1356 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 8
Brehm et al.
(C15H21NO4) C, H, N. Further elution gave 14 and 15 as a 1:1
mixture (350 mg, 29%). 14: 1H NMR δ 1.24 (3H, t, J ) 7.2
Hz), 1.34 (6H, d, J ) 6.2 Hz), 1.78-1.90 (4H, m), 2.41 (2H,
m), 2.83 (2H, m), 4.13 (2H, q, J ) 7.2 Hz), 4.86 (1H, heptet, J
) 6.2 Hz), 5.83 (1H, s). 15: 1H and 13C NMR spectra of pure
15 (see synthesis below).
2.71 (1H, dd, J ) 9.0 and J ) 13.5 Hz), 2.89 (2H, t, J ) 6.6
Hz), 3.15 (1H, dd, J ) 5.5 and J ) 13.5 Hz), 4.53 (1H, m),
4.97 (1H, heptet, J ) 6.2 Hz), 5.77 (1H, t, J ) 6.2 Hz), 6.02
(1H, s br), 6.65 (1H, d br, J ) 7.8 Hz), 7.39-7.55 (3H, m),
7.68-7.74 (2H, m);
13C NMR δ 21.7, 21.9, 22.4, 27.8, 28.49,
28.53, 37.3, 51.2, 53.8, 73.5, 104.9, 125.1, 126.9, 128.4, 131.5,
131.7, 134.0, 167.2, 168.2, 170.7, 171.7. Anal. (C25H33N3O4) C,
H, N.
Eth yl a n d ter t-Bu tyl (3-Isop r op oxy-7,8-d ih yd r o-6H-
cycloh ep ta [d ]isoxa zol-4-yl)a ceta te (15 a n d 16). A solution
of 13 (2.32 g, 8.31 mmol) in dry DME was added to a stirred
solution of potassium tert-butoxide (0.93 g, 8.29 mmol) in dry
DME (12 mL) at 0 °C and under a N2 atmosphere. The reaction
was left to stir for 1 h (followed by TLC) at room temperature.
The reaction mixture was then neutralized by addition of HCl
(4 M). Extraction with CH2Cl2 (2 × 30 mL), drying, evapora-
tion, and CC [toluene/AcOEt (18:1 f 9:1)] gave crude unre-
acted 13, ca. 9%, contaminated with <1% of the analogous tert-
butyl ester (16). Further elution gave 16 (120 mg, 4.7%) as an
oil: 1H NMR δ 1.38 (6H, d, J ) 6.2 Hz), 1.42 (9H, s), 1.90 (2H,
m), 2.32 (2H, m), 2.94 (2H, t, J ) 6.5), 3.37 (2H, s), 4.91 (1H,
heptet, J ) 6.2 Hz), 5.65 (1H, t, J ) 6.2 Hz). Later fractions
gave 15 (1.00 g, 46%) as an oil: 1H NMR δ 1.25 (3H, t, J ) 7.2
Hz), 1.34 (6H, d, J ) 6.1 Hz), 1.90 (2H, m), 2.35 (2H, m), 2.95
(2H, t, J ) 6.5), 3.40 (2H, s), 4.12 (2H, q, J ) 7.2 Hz), 4.91
(1H, heptet, J ) 6.1 Hz), 5.65 (1H, t, J ) 6.2 Hz); 13C NMR δ
14.0, 21.2, 21.7, 28.5, 29.0, 41.1, 60.3, 73.2, 105.4, 122.4, 131.6,
168.3, 170.9, 172.4. Anal. (C15H21NO4) C, H, N.
(R)-2-Ben za m id o-N-ter t-b u t yl-3-(3-isop r op oxy-7,8-d i-
h yd r o-6H-cycloh ep ta [d ]isoxa zol-4-yl)p r op ion a m id e (21).
Compound 21 was synthesized as described for the synthesis
of 20 from 19 (913 mg, 1.68 mmol). Yield: 375 mg (51%) (97%
ee based on HPLC). A small sample was recrystallized [CH2-
Cl2/light petroleum]: mp 172-173 °C; 99.5% ee. 1H and 13C
NMR spectra were virtually identical with those of the
enantiomer compound 20. Anal. (C25H33N3O4) C, H, N.
(S)- a n d (R)-2-Am in o-3-(3-h yd r oxy-7,8-d ih yd r o-6H-cy-
cloh ep ta [d ]isoxa zol-4-yl)p r op ion ic Acid (6 a n d 7). Com-
pound 20 (325 mg, 0.74 mmol) was dissolved in a mixture of
HCl (35%, 20 mL), H2O (10 mL), and AcOH (99.9%, 10 mL),
and the mixture was heated at 130 °C for 1.5 h. After
evaporation, the brownish residue was dissolved in H2O (10
mL) and washed with CH2Cl2 (2 × 10 mL), and the pooled
CH2Cl2 phases were extracted with H2O (10 mL). The H2O
phases were pooled and evaporated. To remove ca. 1 equiv of
(CH3)3CNH2‚HCl, the residue was subjected to ion exchange
[Amberlite IR-400 (OH); eluent AcOH (2 M)]. A solution of
crude racemized 6 (135 mg, 66%; 60% ee) in H2O (15.5 mL)
was filtered through a Millex HV filter (0.45 µm, Millipore)
and resolved on a Chirobiotic T column (0.6 mL injections).
All fractions containing the first eluting enantiomer (6) were
evaporated and re-evaporated twice from H2O. To remove
AcONH4, the dried oily residue was redissolved in H2O (5 mL)
and subjected to a Vydac (RP-18) column in five injections.
Appropriate fractions were pooled, evaporated, re-evaporated
from H2O, and dried in vacuo to give 6 as white crystals (52
mg): >99.7% ee. All fractions containing the second eluting
enantiomer were evaporated, re-evaporated twice from H2O,
and dried in vacuo. The semicrystalline residue was dissolved
in water (1 mL), and AcONH4 was removed on a Vydac column
in three injections. Appropriate fractions were pooled, evapo-
rated, re-evaporated from H2O, and dried in vacuo to give 7
(11.5 mg): mp 174 °C dec; 97% ee. Compound 21 (350 mg,
0.80 mmol) was treated as 20. The deprotection led to crude
racemized 7 (40% ee), which after resolution gave 6 (26 mg),
99.6% ee, and 7 (57.5 mg), 98.8% ee. 6: Recrystallization [52
mg, >99.7% ee; 26 mg, 99.6% ee; H2O (5-10 mL)] gave 57 mg
as light-mauve crystals; mp 180 °C dec; 99.9% ee; 1H NMR
(D2O, CH3CN, δ 2.06) δ 1.83-1.92 (2H, m), 2.27-2.35 (2H, m),
2.84 (1H, dd, J ) 13.8 and 8.1 Hz), 2.90 (2H, t, J ) 6.5 Hz),
3.15 (1H, dd, J ) 6.0 and 13.8 Hz), 3.82 (1H, dd, J ) 6.0 and
(3-Isop r op oxy-7,8-d ih yd r o-6H-cycloh ep ta [d ]isoxa zol-
4-yl)eth a n a l (17). Compound 15 (1.85 g, 6.62 mmol) was
dissolved in dry toluene (40 mL) under a N2 atmosphere, and
the solution was cooled to -78 °C. DIBAL-H hydride (5.5 mL,
1.2 M in hexane, 6.62 mmol) was added dropwise. After the
mixture was stirred for 10 min, the reaction mixture was
quenched with EtOH (10 mL) followed by addition of a
saturated aqueous solution of ammonium dihydrogen phos-
phate (20 mL). The reaction mixture was allowed to warm to
room temperature, and water (40 mL) was added. Extraction
with CH2Cl2 (3 × 100 mL), drying, evaporation, and CC
1
[toluene/AcOEt (9:1)] gave 17 (1.12 g, 72%); H NMR δ 1.35
(6H, d, J ) 6.1 Hz), 1.91 (2H, m), 2.39 (2H, m), 2.97 (2H, t, 6.5
Hz), 3.44 (2H, s br), 4.92 (1H, heptet, J ) 6.1 Hz), 5.68 (1H, t,
J ) 6.0 Hz), 9.71 (1H, t, J ) 1.4 Hz); 13C NMR δ 21.2, 21.7,
28.6, 29.0, 49.8, 73.7, 105.0, 121.0, 132.6, 167.5, 171.3, 200.7.
Anal. (C13H17NO3) C, H, N.
(2S)- a n d (2R)-N-ter t-Bu t yl-3-(3-isop r op oxy-7,8-d i-
h yd r o-6H-cycloh ep ta [d ]isoxa zol-4-yl)-2-{N-[(R)-1-p h en yl-
eth yl]ben za m id o}p r op ion a m id e 18 a n d 19. A solution of
17 (1.12 g, 4.76 mmol) and (R)-1-phenylethylamine (606 µL,
4.76 mmol) in dry MeOH (20 mL) was refluxed for ca. 1.5 h.
After the mixture was cooled to 10-15 °C, benzoic acid (581
mg, 4.76 mmol) was added, and the reaction mixture was
stirred at 10-15 °C for 5 min. Tert-butylisonitrile (538 µL, 4.76
mmol) was added to the mixture, and stirring was continued
at room temperature overnight. Evaporation followed by CC
[toluene/AcOEt (9:1)] gave 374 mg (14.5%) of 18 (99.9% de
based on HPLC): 1H NMR δ 1.35 (3H, d br, J ) 6.3 Hz), 1.39
(9H, s), 1.47 (3H, d br, J ) 6.0 Hz), 1.54 (3H, d br, J ) 6.6
Hz), 1.81 (2H, m), 2.24 (2H, m), 2.82 (2H, m), 3.23 (2H, m),
3.84 (1H, m), 4.77 (1H, m), 4.98 (1H, m), 5.59 (1H, m), 7.1-
7.2 (5H, m), 7.4-7.5 (5H, m), 8.18 (1H, s br). Further elution
gave 399 mg (15.4%) of 19 (99.5% de based on HPLC): 1H
NMR δ 1.12 (9H, s), 1.27 (3H, d br, J ) 6.0 Hz), 1.42 (3H, d
br, J ) 6.3 Hz), 1.51 (3H, d br, J ) 6.0 Hz), 1.93 (2H, m), 2.35
(2H, m), 2.98 (2H, m), 3.36 (2H, m), 3.91 (1H, m), 5.04 (2 ×
1H, m), 5.76 (1H, t, J ) 6.0 Hz), 7.0-7.3 (5H, m), 7.4-7.5 (5H,
m).
8.1 Hz), 5.88 (1H, t, J ) 6.2 Hz); [R]25 -0.8° (c 0.38, 0.1 M
D
HCl); ∆ꢀ (210 nm) ) +0.2 m2/mol. Anal. (C11H14N2O4‚H2O) C,
H, N. 7: White crystals (57.5 mg); mp 174 °C dec; 98.8% ee;
[R]25 +1.0° (c 0.40, 0.1 M HCl); ∆ꢀ (210 nm) ) -0.2 m2/mol.
D
The 1H NMR spectrum was identical with that of the (S)-
enantiomer 6. Anal. (C11H14N2O4‚H2O) H, N. C: calcd 51.56;
found 51.09.
Cell Cu ltu r e a n d Secon d Messen ger Assa ys. Chinese
hamster ovary (CHO) cell lines expressing mGluR1R, mGluR2,
and mGluR4a were maintained as previously described.9,10,51
Briefly, cells were maintained in a humidified 5% CO2/95%
air atmosphere at 37 °C in DMEM containing a reduced
concentration of (S)-glutamine (100 mg/mL) and 10% dialyzed
fetal calf serum (all GIBCO, Paisley, Scotland). The day before
the inositol phosphate assay, two million mGluR1R-expressing
cells were divided into the wells of a 96-well plate in inositol-
free culture medium containing 4 µCi/mL [3H]inositol. The day
before the cyclic AMP-assay, two million mGluR2- or mGluR4a-
expressing cells were divided into the wells of a 96-well plate
in culture medium. Measurements of inositol phosphate
generation and cyclic AMP inhibition were determined by ion-
exchange chromotography and scintillation proximity assays
as previously described.47
(S)-2-Ben za m id o-N-ter t-b u t yl-3-(3-isop r op oxy-7,8-d i-
h yd r o-6H-cycloh ep ta [d ]isoxa zol-4-yl)p r op ion a m id e (20).
A solution of 18 (754 mg, 1.39 mmol) in formic acid (98%, 20
mL) was stirred at room temperature for 30 min and then 5.5
h at 60 °C. Evaporation followed by FC [toluene/AcOEt (4:1)]
gave 352 mg (58%) of 20 (95% ee based on HPLC). A small
sample was recrystallized [CH2Cl2/light petroleum]: mp 173-
1
174 °C; 99.8% ee; H NMR δ 1.34 (9H, s), 1.40 (3H, d, J ) 6
Hz), 1.41 (3H, d, J ) 6 Hz), 1.67-1.94 (2H, m), 2.25 (2H, m),

A Propionic Acid as Agonist
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 8 1357
Recep tor Bin d in g Assa ys. The membrane preparations
used in all receptor binding experiments were prepared
according to Ransom and Stec52 with slight modifications as
previously described.21 Affinities for AMPA, kainic acid, and
NMDA receptor sites were determined using 5 nM [3H]-
AMPA,53 5 nM [3H]kainic acid54 in the absence of CaCl2, and
2 nM [3H]CGP 39653,55 respectively. The binding assays were
carried out with the modifications previously described.21 The
amount of bound radioactivity was determined using a Pack-
ard TOP-COUNT microplate scintillation counter.
Electr op h ysiology. In vitro cRNA transcription was per-
formed as described previously.22 Xenopus laevis oocytes were
isolated and injected as described previously.22 Currents were
recorded 3-14 days after injection using a two-microelectrode
clamp (OC-725C oocyte clamp, Warner Instrument). Experi-
ments were performed in low Ca2+-Ringer (10 mM Na-HEPES,
pH 7.5, 1165 mM NaCl, 0.1 mM CaCl2, 2.5 mM KCl, 1.8 mM
MgCl2) to avoid activation of the endogenous Ca2+-dependent
Cl- channel. The electrodes were filled with 3 M KCl and had
a resistance of 0.7-2 MΩ. Oocytes were clamped at -100 to
-20 mV, and recordings on homomeric GluR5 and GluR6 were
performed after a 10 min treatment with concanavalin A (1
mg/mL, Sigma type IV) to minimize receptor desensitization.
Da ta An a lysis. Binding data were analyzed by the non-
linear curve-fitting program GRAFIT 3.0.56 Data were fit to
the following equation:
Refer en ces
(1) Excitatory Amino Acids and Synaptic Transmission; Wheal, H.
V., Thomson, A. M., Eds.; Academic Press: London, 1995.
(2) The Ionotropic Glutamate Receptors; Monaghan, D. T., Wenthold,
R. J ., Eds.; Humana Press: Totowa, NJ , 1997.
(3) Parsons, C. G.; Danysz, W.; Quack, G. Glutamate in CNS
disorders as a target for drug development: an update. Drug
News Perspect. 1998, 11, 523-569.
(4) The Metabotropic Glutamate Receptors; Conn, P. J ., Patel, J .,
Eds.; Humana Press: Totowa, NJ , 1994.
(5) Dingledine, R.; Borges, K.; Bowie, D.; Traynelis, S. F. The
glutamate receptor ion channels. Pharmacol. Rev. 1999, 51,
7-61.
(6) Bra¨uner-Osborne, H.; Egebjerg, J .; Nielsen, E. Ø.; Madsen, U.;
Krogsgaard-Larsen, P. Ligands for glutamate receptors: design
and therapeutic prospects. J . Med. Chem. 2000, 43, 2609-2645.
(7) Kno¨pfel, T.; Kuhn, R.; Allgeier, H. Metabotropic glutamate
receptors: novel targets for drug development. J . Med. Chem.
1995, 38, 1417-1426.
(8) Monn, J . A.; Schoepp, D. D. Metabotropic glutamate receptor
modulators: recent advances and therapeutic potential. Annu.
Rep. Med. Chem. 2000, 35, 1-10.
(9) Tanabe, Y.; Masu, M.; Ishii, T.; Shigemoto, R.; Nakanishi, S. A
family of metabotropic glutamate receptors. Neuron 1992, 8,
169-179.
(10) Aramori, I.; Nakanishi, S. Signal transduction and pharmaco-
logical characteristics of a metabotropic glutamate receptor,
mGluR1, in transfected CHO cells. Neuron 1992, 8, 757-765.
(11) Ionotropic Glutamate Receptors in the CNS; J onas, P., Monyer,
H., Eds.; Springer: Berlin, 1999.
(12) Egebjerg, J .; Schousboe, A.; Krogsgaard-Larsen, P. Glutamate
and GABA Receptors and Transporters. Structure, Function and
Pharmacology; Taylor and Francis: London, 2002.
(13) Hollmann, M.; Heinemann, S. Cloned glutamate receptors.
Annu. Rev. Neurosci. 1994, 17, 31-108.
100 × [inhibitor]n
B ) 100 -
n
IC50 + [inhibitor]n
(14) Dingledine, R.; Borges, K.; Bowie, D.; Traynelis, S. F. The
glutamate receptor ion channels. Pharmacol. Rev. 1999, 51,
7-61.
where B is the binding as a percentage of total specific binding
and n is the Hill coefficient.
Data on the cloned iGluR subtypes were acquired using
Clampex 7.0 (Axon Instruments) and subsequently processed
with Sigmaplot 3.0 (J andel Scientific, SPSS). Dose-response
data from individual oocytes were fit to the equation
(15) Rosenmund, C.; Stern-Bach, Y.; Stevens, C. F. The tetrameric
structure of a glutamate receptor channel. Science 1998, 280,
1596-1599.
(16) Fletcher, E. J .; Lodge, D. New developments in the molecular
pharmacology of R-amino-3-hydroxyisoxazole propionate and
kainate receptors. Pharmocol. Ther. 1996, 70, 65-89.
(17) Lees, G. J . Pharmacology of AMPA/kainate receptor ligands and
their therapeutic potential in neurological and psychiatric
disorders. Drugs 2000, 59, 33-78.
(18) Lauridsen, J .; Honore´, T.; Krogsgaard-Larsen, P. Ibotenic acid
analogues. Synthesis, molecular flexibility, and in vitro activity
of agonists and antagonists at central glutamic acid receptors.
J . Med. Chem. 1985, 28, 668-672.
[agonist]n
I ) Imax
n
EC50 + [agonist]n
where I is the current, Imax is the maximum current in response
to the action of the agonist, and n denotes the Hill coefficient.
Relative maximal currents (relative Imax) were calculated
as Imax ligand/Imax kainic acid or AMPA where Imax ligand is the fitted Imax
according to the equation (see above). Imax kainic acid or AMPA is
calculated from the current evoked by kainic acid (for GluR1o,
GluR3o, GluR4o, and GluR5) or AMPA (for GluR6 + KA2),
and the EC50 values previously defined in our laboratory for
the two agonists. The (kainic acid EC50 values)/(n values) for
homomeric GluR1o, GluR3o, GluR4o, and GluR5 were deter-
mined to be (55.1 µM)/1.28, (27.1 µM)/2.30, (29.8 µM)/1.80, (9.6
µM)/1.38, respectively, and the (AMPA EC50 value)/(n value)
for heteromeric GluR6 + KA2 was determined to be (110.0
µM)/1.84.
(19) Clarke, V. R. J .; Ballyk, B. A.; Hoo, K. H.; Mandelzys, A.;
Pellizzari, A.; Bath, C. P.; Thomas, J .; Sharpe, E. F.; Davies, C.
H.; Ornstein, P. L.; Shoepp, D. D.; Kamboj, R. K.; Collingridge,
G. L.; Lodge, D.; Bleakman, D. A hippocampal GluR5 kainate
receptor regulating inhibitory synaptic transmission. Nature
1997, 389, 599-603.
(20) Hoo, K.; Legutko, B.; Rizkalla, G.; Deverill, M.; Hawes, C. R.;
Ellis, G. J .; Stensbøl, T. B.; Krogsgaard-Larsen, P.; Skolnick, P.;
Bleakman, D. [3H]ATPA:
a high affinity ligand for GluR5
kainate receptors. Neuropharmacology 1999, 38, 1811-1817.
(21) Stensbøl, T. B.; J ohansen, T. N.; Egebjerg, J .; Madsen, U.; Ebert,
B.; Krogsgaard-Larsen, P. Resolution, absolute stereochemistry
and molecular pharmacology of the enantiomers of ATPA. Eur.
J . Pharmacol. 1999, 380, 153-162.
(22) Stensbøl, T. B.; J ensen, H. S.; Nielsen, B.; J ohansen, T. N.;
Egebjerg, J .; Frydenvang, K.; Krogsgaard-Larsen, P. Stereo-
chemistry and molecular pharmacology of (S)-thio-ATPA, a new
potent and selective GluR5 agonist. Eur. J . Pharmacol. 2001,
411, 245-253.
(23) Thomas, N. K.; Hawkins, L. M.; Miller, J . C.; Troop, H. M.;
Roberts, P. J .; J ane, D. E. Pharmacological differentiation of
kainate receptors on neonatal rat spinal motoneurones and
dorsal roots. Neuropharmacology 1998, 37, 1223-1237.
(24) Chittajallu, R.; Braithwaite, S. P.; Clarke, V. R. J .; Henley, J .
M. Kainate receptors: subunits, synaptic localization and func-
tion. Trends Pharmacol. Sci. 1999, 20, 26-35.
(25) Krogsgaard-Larsen, P.; Nielsen, E. Ø; Curtis, D. R. Ibotenic acid
analogues. Synthesis and biological and in vitro activity of
conformationally restricted agonists at central excitatory amino
acid receptors. J . Med. Chem. 1984, 27, 585-591.
(26) Ebert, B.; Madsen, U.; Lund, T. M.; Holm, T.; Krogsgaard-
Larsen, P. Molecular pharmacology of cortical and spinal AMPA
receptors. Mol. Neuropharmacol. 1992, 2, 47-49.
The mean and SEM are calculated assuming a logarithmic
distribution of the EC50 or IC50 values. Here, numbers in
parentheses (min, max) indicate ( SEM according to a
logarithmic distribution of EC50 and IC50 values.57
Ack n ow led gm en t. This work was supported by
grants from the Danish Medical Research Council, the
Novo Nordisk Foundation, and the Lundbeck Founda-
tion. The technical assistance of Karen J ørgensen,
Department of Chemistry, University of Copenhagen,
with the recording of CD spectra, Ulla Geneser with the
syntheses, and the secretarial assistance of Anne Nordly
are gratefully acknowledged.
(27) Krogsgaard-Larsen, P.; Natova, L.; Christensen, S. B. Muscimol
analogues. I. Syntheses of 4- and 8-aminocyclohepteno[1,2-d]-
isoxazol-3-ols and 4-(3-aminopropyl)-5-methyl-3-isoxazolol. Acta
Chem. Scand. 1977, 31, 577-583.
Su p p or tin g In for m a tion Ava ila ble: Homology model of
GluR5 as Protein Data Bank file. This material is available
free of charge via the Internet at http://pubs.acs.org.

1358 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 8
Brehm et al.
(28) Ugi, I.; Offermann, K.; Herlinger, H.; Marquarding, D. Die
Umsetzung von (S)-R-Phenyla¨thylamin und Isobutyraldehyd mit
Benzoesa¨ure und tert-Butyl-isocyanid als Modellreaktion fu¨r
stereoselektive Peptid-Synthesen mittels Vierkomponenten-
Kondensationen (The reaction of (S)-R-phenylethylamine and
isobutyraldehyde with benzoic acid and tert-butyl isocyanide as
a model reaction for stereoselective peptide synthesis through
a four-component condensation). Leibigs Ann. Chem. 1967, 709,
1-10.
(29) Ahmadian, H.; Nielsen, B.; Bra¨uner-Osborne, H.; J ohansen, T.
N.; Stensbøl, T. B.; Sløk, F. A.; Sekiyama, N.; Nakanishi, S.;
Krogsgaard-Larsen, P.; Madsen, U. (S)-Homo-AMPA, a specific
agonist at the mGlu6 subtype of metabotropic glutamic acid
receptors. J . Med. Chem. 1997, 40, 3700-3705.
(30) Do¨mling, A.; Ugi, I. Multicomponent reactions with isocyanides.
Angew. Chem., Int. Ed. 2000, 39, 3168-3210.
(31) Marquarding, D.; Hoffmann, P.; Heitzer, H.; Ugi, I. Stereose-
lective four-component condensations of R-ferrocenylethylamine
and its absolute configuration. J . Am. Chem. Soc. 1970, 92,
1969-1971.
(32) Semple, J . E.; Wang, P. C.; Lysenko, Z.; J oullie´, M. M. Total
synthesis of (+)-furanomycin and stereoisomers. J . Am. Chem.
Soc. 1980, 102, 7505-7510.
(33) 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 . Chro-
matogr. A 1996, 731, 123-137.
(34) Shinbo, T.; Yamaguchi, T.; Nishimura, K.; Sugiura, M. Chro-
matographic separation of racemic amino acids by use of chiral
crown ether-coated reversed-phase packings. J . Chromatogr.
1987, 405, 145-153.
(35) J ohansen, T. N.; Ebert, B.; Falch, E.; Krogsgaard-Larsen, P.
AMPA receptor agonists: Resolution, configurational assignment
and pharmacology of (+)-(S)- and (-)-(R)-2-amino-3-[3-hydroxy-
5-(2-pyridyl)isoxazol-4-yl]propionic acid (2-Py-AMPA). Chirality
1997, 9, 274-280.
(36) Falch, E.; Brehm, L.; Mikkelsen, I.; J ohansen, T. N.; Skjærbæk,
N.; Nielsen, B.; Stensbøl, T. B.; Ebert, B.; Krogsgaard-Larsen,
P. Heteroaryl analogues of AMPA. 2. Synthesis, absolute ster-
eochemistry, photochemistry, and structure-activity relation-
ships. J . Med. Chem. 1998, 41, 2513-2523.
(37) J ohansen, T. N.; Ebert, B.; Bra¨uner-Osborne, H.; Didriksen, M.;
Christensen, I. T.; Søby, K. K.; Madsen, U.; Krogsgaard-Larsen,
P.; Brehm, L. Excitatory amino acid receptor ligands: resolution,
absolute stereochemistry, and enantiopharmacology of 2-amino-
3-(4-butyl-3-hydroxyisoxazol-5-yl)propionic acid. J . Med. Chem.
1998, 41, 930-939.
(38) Macromodel 7.2-8.0 (2001) and Glide 1.8 (2002); Schro¨dinger,
Inc., 1500 SW First Avenue, Suite 1180, Portland, OR 97201.
(39) Altschul, S. F.; Gish, W.; Miller, W.; Myers, E. W.; Lipman, D.
J . Basic local alignment search tool (BLAST). J . Mol. Biol. 1990,
215, 403-410.
(40) 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.
(41) http://www.expasy.ch/swissmod/SWISS-MODEL.html. Schwede,
T.; Diemand, A.; Guex, N.; Peitsch, M. C. Protein structure
computing in the genomic era. Res. Microbiol. 2000, 151, 107-
112.
(42) Hogner, A.; Kastrup, J . S.; J in, 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.
(43) Lunn, M.-L. Ph.D. Thesis, The Royal Danish School of Pharmacy,
Copenhagen, Denmark, 2002.
(44) Brehm, L.; J ørgensen, F. S.; Hansen, J . J .; Krogsgaard-Larsen,
P. Agonists and antagonists for central glutamic acid receptors.
Drug News Perspect. 1988, 1, 138-144.
(45) Bra¨uner-Osborne, H.; Sløk, F. A.; Skjærbæk, N.; Ebert, B.;
Sekiyama, N.; Nakanishi, S.; Krogsgaard-Larsen, P. A new
highly selective metabotropic excitatory amino acid agonist:
2-amino-4-(3-hydroxy-5-methylisoxazol-4-yl)butyric acid. J . Med.
Chem. 1996, 39, 3188-3194.
(46) Bra¨uner-Osborne, H.; Krogsgaard-Larsen, P. Pharmacology of
(S)-homoquisqualic acid and (S)-2-amino-5-phosphonopentanoic
acid [(S)-AP5] at cloned metabotropic glutamate receptors. Br.
J . Pharmacol. 1998, 123, 269-274.
(47) Bra¨uner-Osborne, H.; Nielsen, B.; Krogsgaard-Larsen, P. Mo-
lecular pharmacology of homologues of ibotenic acid at cloned
metabotropic glutamic acid receptors. Eur. J . Pharmcol. 1998,
350, 311-316.
(48) Lerma, J .; Paternain, A. V.; Rodr´ıguez-Moreno, A.; Lo´pez-Garc´ıa,
J . C. Molecular physiology of kainate receptors. Physiol. Rev.
2001, 81, 971-998.
(49) Smolders, I.; Bortolotto, Z. A.; Clarke, V. R. J .; Warre, R.; Khan,
G. M.; O’Neill, M. J .; Ornstein, P. L.; Bleakman, D.; Ogden, A.;
Weiss, B.; Stables, J . P.; Ho, K. H.; Ebinger, G.; Collingridge,
G. L.; Lodge, D.; Michotte, Y. Antagonists of GLUK5-containing
kainate receptors prevent pilocarpine-induced limbic seizures.
Nat. Neurosci. 2002, 5, 796-804.
(50) Banke, T. G.; Greenwood, J . R.; Christensen, J . K.; Liljefors, T.;
Schousboe, A.; Pickering, D. S. Identification of amino acid
residues in GluR1 responsible for ligand binding and desensi-
tisation. J . Neurosci. 2001, 21, 3052-3062.
(51) Tanabe, Y.; Nomura, A.; Masu, M.; Shigemoto, R.; Mizuno, N.;
Nakanishi, S. Signal transduction, pharmacological properties,
and expression patterns of two rat metabotropic glutamate
receptors, mGluR3 and mGluR4. J . Neurosci. 1993, 13, 1372-
1378.
(52) Ransom, R. W.; Stec, N. L. Cooperative modulation of [3H]MK-
801 binding to the N-methyl-D-aspartate receptor ion channel
complex by L-glutamate, glycine and polyamines. J . Neurochem.
1988, 51, 830-836.
(53) Honore´, T.; Nielsen, M. Complex structure of quisqualate-
sensitive glutamate receptors in rat cortex. Neurosci. Lett. 1985,
54, 27-32.
(54) Braitman, D. J .; Coyle, J . T. Inhibition of [3H]kainic acid receptor
binding by divalent cations correlates with ion affinity for the
calcium channel. Neuropharmacology 1987, 26, 1247-1251.
(55) Sills, M. A.; Fagg, G.; Pozza, M.; Angst, C.; Brundish, D. E.; Hurt,
S. D.; Wilusz, E. J .; Williams, M. [3H]CGP 39653:
a new
N-methyl-D-aspartate antagonist radioligand with low nanomo-
lar affinity in rat brain. Eur. J . Pharmacol. 1991, 192, 19-24.
(56) Leatherbarrow, R. J . GraFit, version 3.0; Erithacus Software:
Staines, U.K., 1992.
(57) Christopoulus, A. Assessing the distrubution of parameters in
models of ligand-receptor interactions: to log or not to log.
Trends Pharmacol. Sci. 1998, 19, 351-357.
J M0204441