Design and synthesis of novel quinoxaline-2,3-dione AMPA/Gly(N) receptor antagonists: Amino acid derivatives

Article abstract of DOI:10.1021/jm980455n

PNQX (1,4,7,8,9,10-hexahydro-9-methyl-6-nitropyrido[3,4-f]quinoxaline- 2,3-dione) is a potent AMPA (IC₅₀ = 0.063 μM) and Gly(N) (IC₅₀ = 0.37 μM) receptor antagonist that was developed in our laboratories. While possessing a desirable

Full text of DOI:10.1021/jm980455n

2266
J . Med. Chem. 1999, 42, 2266-2271
Brief Articles
Design a n d Syn th esis of Novel Qu in oxa lin e-2,3-d ion e AMP A/GlyN Recep tor
An ta gon ists: Am in o Acid Der iva tives
Sham S. Nikam,*,† J ohn J . Cordon,‡ Daniel F. Ortwine,† Tycho H. Heimbach,§ Anthony C. Blackburn,§
Mark G. Vartanian,‡ Carrie B. Nelson,‡ Roy D. Schwarz,‡ Peter A. Boxer,‡ and Michael F. Rafferty†
Departments of Chemistry, Neuroscience Therapeutics, and Pharmaceutics and Drug Delivery,
Parke-Davis Pharmaceutical Division/ Warner-Lambert Company, 2800 Plymouth Road, Ann Arbor, Michigan 48105
Received August 4, 1998
PNQX (1,4,7,8,9,10-hexahydro-9-methyl-6-nitropyrido[3,4-f]quinoxaline-2,3-dione) is a potent
AMPA (IC50 ) 0.063 µM) and GlyN (IC50 ) 0.37 µM) receptor antagonist that was developed in
our laboratories. While possessing a desirable in vitro and in vivo activity profile, this compound
suffers from low aqueous solubility. In an effort to improve its potency and physical properties,
we have designed and synthesized novel ring-opened analogues 4, 6, 9, and 11. Modeling
analyses demonstrated that, while the 5-substituent in these analogues was forced to adopt
an out-of-plane conformation due to steric contacts with neighboring substituents, the overall
structure retained a good fit to a previously described AMPA pharmacophore model. This
nonplanar orientation may lessen efficient packing in the solid state, compared to PNQX, leading
to increased water solubility. Indeed, several nonplanar analogues containing appropriate
functionalities, for example, the sarcosine analogue 9, were found to retain AMPA (IC50 ) 0.14
µM) and GlyN (IC50 ) 0.47 µM) receptor affinity and possess improved aqueous solubility
compared to PNQX. The synthesis and the SAR of these compounds are discussed.
In tr od u ction
Glutamic acid is the principal excitatory neurotrans-
mitter in the central nervous system (CNS), and it plays
a significant role in regulating neuronal activity via
various glutamate receptors. Excessive release of glutam-
ic acid can cause overexcitation of these receptors,
leading to neuronal death. Within the ionotropic class
of these receptors, R-amino-3-hydroxy-5-methyl-4-isox-
azolepropionic acid (AMPA), kainate (KA), and N-
methyl-D-aspartate (NMDA) receptors have received
significant attention due to their implication in a variety
of pathophysiological conditions such as global and focal
ischemia, epilepsy, and Parkinson’s, Huntington’s, and
Alzheimer’s diseases.1-9 We have focused on developing
potent antagonists for the AMPA, KA, and NMDA
cator of CNS penetration. Potency is maintained in
animal stroke models, such as middle carotid artery
occlusion (MCAO) and global forebrain ischemia.14 Due
to its excellent in vitro and in vivo profile, PNQX was
selected as a lead compound in our stroke program. The
major disadvantage of PNQX is its low aqueous solubil-
ity (0.0086 mg/mL, Table 2), which is also observed in
several other excitatory amino acid antagonists contain-
ing a quinoxaline-2,3-dione motif. The poor aqueous
solubility leads to the potential of crystallization in the
kidney. Other disadvantages are the short duration of
action following iv administration (<5 min) in the MES
assay and a lengthy linear chemical synthesis, which
precludes rapid analoging. Thus, our primary goal in
the identification of potential backups for PNQX was
to design and synthesize equipotent molecules with
improved aqueous solubility.
associated glycine (GlyN) receptors.
A number of competitive AMPA receptor antagonists
have been developed in the past few years, several of
which have shown neuroprotective activity. Examples
include NBQX,10 NS 257,11 YM 90K,12 and PNQX.13 Of
these, NBQX has been studied extensively due to its
selective affinity for the AMPA receptor.10 The focus of
this study was PNQX, which displays high affinity for
both the AMPA and GlyN receptors.13 PNQX is also
potent in the maximal electroshock assay (MES; ED50
) 0.44 mg/kg, iv), which has been used as a measure of
anticonvulsant activity and a preliminary in vivo indi-
* To whom correspondence should be addressed. Phone: (734)-622-
7452. Fax: (734)-622-5165. E-mail: sham.nikam@wl.com.
† Department of Chemistry.
‡ Department of Neuroscience Therapeutics.
§ Department of Pharmaceutics and Drug Delivery.
10.1021/jm980455n CCC: $18.00 © 1999 American Chemical Society
Published on Web 05/29/1999

Brief Articles
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 12 2267
Ch a r t 1. Nonplanar Analogues of PNQX
Sch em e 1. Hofmann Degradation of the Isoquinoline
Intermediate
Ch a r t 2. Pharmacophore Model for AMPA Receptor
Antagonists
we simplified the model by dividing PNQX into halves:
viz. the northern half and the southern half. On the
basis of the SAR of PNQX and its analogues, we decided
to keep the southern portion intact and make necessary
modifications in the northern portion, particularly the
C-5 and C-6 positions, to improve the physical properties
of the new quinoxaline-2,3-dione derivatives. During the
course of writing this manuscript, new 5-(aminomethyl)-
quinoxaline-2,3-dione analogues were reported.15 How-
ever, these compounds are unsubstituted at C-6, and
the overall rationale and synthesis differ significantly
from this report.
Ch a r t 3. Pharmacophore Model for GlyN Receptor
Antagonists
Ch em istr y
Synthesis of new analogues was designed based on
the intermediates available from the PNQX synthesis.
Thus, the synthetic strategy (Scheme 1) involved Hof-
mann degradation of the substituted tetrahydro-
isoquinoline intermediate 1 with ethyl bromoacetate in
the presence of potassium carbonate in boiling acetone.
The reaction gave a good yield of the desired benzyl-
amine derivative, [(2-amino-5-bromo-3-nitro-6-vinylben-
zyl)methylamino]acetic acid ethyl ester (2), and has
proven to be an excellent route to multifunctional phenyl
rings. Compound 2 was the critical intermediate for the
synthesis of the new acyclic analogues of PNQX, as it
was appropriately substituted with 3-amino-2-nitro
groups necessary for conversion to the corresponding
quinoxaline-2,3-dione derivatives.
The poor solubility of PNQX is due in part to the
relatively planar tricyclic motif of the molecule, which
is likely responsible for the efficient packing in the solid
state, which can in turn decrease solubility. Thus, our
initial strategy to make soluble PNQX analogues was
to open the reduced pyridine ring (Chart 1) to give
nonplanar, C-5- and C-6-substituted quinoxaline-2,3-
dione derivatives with a substituted aminomethyl side
chain at C-5. Polar substituents, such as a carboxymeth-
yl moiety, could then be added onto the amine function-
ality to further improve the aqueous solubility of the
final compounds.
To preserve affinity at the AMPA/GlyN receptors, the
design of new analogues was also based on AMPA/GlyN
pharmacophore models (Charts 2 and 3). The AMPA
pharmacophore model was generated by superimposing
several PNQX analogues onto other active AMPA
antagonists, and it has been described earlier by Bigge
et al.8,13 It was obvious from this model that the 7-nitro-
quinoxaline-2,3-dione motif was very important for good
AMPA/GlyN receptor affinity and that the N-4, C-5, and
C-6 positions were amenable for substitution with
groups of limited polarity and bulk. For design purposes,
The benzylamine derivative 2 was then selectively
reduced using catalytic hydrogenation to give the cor-
responding o-phenylenediamine derivatives (Scheme
2). Reduction of 2 in the presence of RaNi in THF gave
the 6-vinyl-7-bromo-o-phenylenediamine derivative 3,
whereas the use of ethanol as a solvent in the reduction
step gave the corresponding 6-ethyl-7-bromo derivative
5. Replacing RaNi with Pd/C (20%) as the catalyst and
methanol as a solvent gave the desbromo-6-ethyl-o-
phenylenediamine derivative 7. The o-phenylenedi-
amine derivatives 3, 5, and 7, on treatment with oxalic
acid in 2 N HCl at 90 °C, gave the corresponding
quinoxaline-2,3-dione derivatives 4, 6, and 8, respec-
tively. The ethyl sarcosinate analogue 11 was synthe-
sized by cyclizing the intermediate 7 with dimethyl
oxalate in boiling THF to give the corresponding qui-
noxaline-2,3-dione derivative 10. The quinoxaline-2,3-
dione derivative 10 was subsequently nitrated with
KNO3 in sulfuric acid at 0 °C to give the 7-nitro
derivative 11.

2268 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 12
Brief Articles
Sch em e 2. Synthesis of 5-Aminomethyl Derivatives of Quinoxaline-2,3-dionesa
a
(a) RaNi/H2/THF; (b) (CO2/H)2/Cl; (c) RaNi/H2/EtOH; (d) Pd/C (20%), H2/EtOH; (e) KNO3/H2SO4; (f) (CO2Me)2/EtOH.
Ta ble 1. Acyclic PNQX Analogues: Sarcosine Derivatives
IC50 (µM)a
compd
AMPA
GlyN
PNQX
4
6
8
9
11
0.063
1.18 ((0.29)
2.32 ((0.63)
0.37
NT
NT
>10
0.47
>10
>10
0.14 ((0.02)
1.59 ((1.17)
a
NT, not tested. The AMPA and GlyN data are the mean of
three experiments with triplicate incubations in each experiment.
SEM for compound 9 in [3H]GlyN assay (n ) 3) is (0.007.
Resu lts a n d Discu ssion
F igu r e 1. Orthogonal view of the overlay of sarcosine deriva-
tive 9 (colored by atom type) and PNQX (orange) fit to an
AMPA pharmacophore model.8,13 Four receptor interaction site
points (A-D) from the model have been included. Note the
relatively nonplanar orientation of the amino acid side chain,
particularly the N-methyl group.
The in vitro binding data for the sarcosine derivatives
are summarized in Table 1. The nonplanar acyclic
analogue, sarcosine derivative 9, retained the AMPA
and GlyN receptor binding affinity shown by PNQX,
suggesting that the northern portion of the molecule is
indeed amenable to modifications and can tolerate
bigger groups than that found on PNQX. Modeling
studies confirmed that the aminomethyl side chain is
held nearly 90° out of plane of the quinoxaline-2,3-dione
ring system (Figure 1) due to steric contacts with
neighboring groups. Modeling also showed that one
oxygen of the carboxylic acid group, as well as the amine
from the aminomethyl side chain, can interact at the
same H-bond donor site (Chart 2, site D) of the receptor
in the AMPA pharmacophore model as the piperidine
ring nitrogen of PNQX, albeit from different directions.
However, the lack of an enhancement in potency over
PNQX suggests that an additional hydrogen bond to this
site contributes little to overall binding energy. The
overall SAR of the new sarcosine analogues indicates
that the southern portion of the molecule is important
for activity, with the amide carbonyl groups at C-2 and
C-3 participating in a Coulombic interaction with a
positively charged site (Chart 2, site A) at the receptor.
The electron-withdrawing nitro group at C-7 not only
interacts at a hydrogen bond-accepting site (Chart 3,
site C) near the southwestern region of the molecule but
also enhances the Coulombic interaction in the eastern
portion by rendering the proton on the N-4 nitrogen

Brief Articles
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 12 2269
Ta ble 2. Measured Solubility and Predicted pKa’sb
charges as observed in amino acids and the fact that
both the monomers and the zwitterionic dimers that can
form are less planar and less rigid than those of PNQX.
This would result in a reduction in packing efficiency
in the solid state, leading to increased solubility. The
solubility differences between the acyclic PNQX ana-
logues themselves are less dramatic and are most likely
driven by differences in their pKa’s since the most
soluble of those analogues is also predicted to be most
negatively charged at pH 7.4.
compd
solubilitya (µg/mL)
predicted pKa’sb
PNQX
4
6
9
8.6
170
140
420
150
6.9, 7.4, 10.6
2.3, 8.3, 8.8, 11.7
2.3, 8.3, 8.8, 11.9
2.3, 7.8, 8.8, 10.7
5.9, 8.5, 10.6
11
a
Solubility in pH 7.4, 50 mM phosphate buffer. b Predicted pKa’s
using ACD Labs log D suite version 3.00.
more acidic. A nitro group seems optimum here, as the
7-bromo analogue 6 had much less affinity for the
AMPA receptor and the analogue with no substitution
at C-7 (8) was inactive. These observations are consis-
tent with the primary sites of interaction indicated
earlier in the AMPA pharmacophore model based on
PNQX analogues.8,9,13
In conclusion, the strategy to synthesize new nonpla-
nar acyclic analogues of PNQX has resulted in ana-
logues such as the sarcosine derivative 9 that show
comparable affinity for the AMPA receptor with at least
2-fold better affinity for the GlyN receptor and improved
solubility. Work is in progress to delineate other non-
planar analogues with better in vivo activity.
The sarcosine analogue 9 also shows excellent affinity
for the GlyN site, which indicates that the polar car-
boxylic acid group in the aminomethyl side chain has
enough flexibility to interact with the H-bonding inter-
action site at the GlyN receptor, which appears to be
directed in the northeastern portion of the molecule in
comparison to the northwestern orientation for the
AMPA receptor. This northeastern orientation of the
sarcosine side chain is favored due to the presence of
the ethyl group at C-6. This ethyl substituent, in
combination with the quinoxaline ring N-4-H, con-
strains the aminomethyl side chain at C-5 out of the
plane of the quinoxaline ring system. The ethyl sarco-
sinate derivative 11 was synthesized to reduce overall
polarity, since 9 was not active in the MES assay. This
compound was much weaker at the AMPA receptor in
comparison to the sarcosine derivative 9 but showed
weak activity in the MES assay (20% protection @ 30
mg/kg, iv). This is consistent with the ester probably
acting as a prodrug for the required carboxylic acid at
the AMPA receptor.
The strategy of increasing solubility by making the
nonplanar compounds was successful, as sarcosine
derivatives 4, 6, 9, and 11 were much more soluble than
PNQX at pH 7.4 (Table 2). These results can be
explained by examining the difference in ionization
potential as well as the difference in rigidity and
planarity of the PNQX and its acyclic analogues.
The aqueous solubility of PNQX is limited by the fact
that it is predicted (pKa’s, Table 2) to exist as a
zwitterion at pH 7.4. That is, half of the species in
solution at pH 7.4 should be zwitterionic and half should
be negatively charged. Thus, if the predicted pKa’s are
accurate and we assume that the zwitterionic form is
its least soluble form, the solubility of the zwitterionic
form of PNQX is 4.3 µg/mL. This zwitterionic form is
likely to be nearly insoluble because it sets up the
formation of rigid and somewhat planar zwitterionic
dimers that can then pack efficiently and/or utilize
hydrogen bonding to form a robust crystal lattice. These
dimers are planar and rigid due to the planarity and
rigidity of the monomers and due to the charges being
located at opposite ends of the monomers.
Exp er im en ta l Section
[3H ]AMP A, [3H ]Glycin e, a n d [3H ]Ka in a t e R ecep t or
Bin d in g. Receptor binding was measured as percent displace-
ment of [3H]AMPA and [3H]kainate from extensively washed
rat cortical synaptosomal membranes. Using 50 mM Tri Hl
buffer, pH 7.4, tubes containing membranes, 10 nM [3H]-
AMPA, 10 mM KSCN, and test compounds were incubated
on ice for 60 min,16 while other tubes containing 5 nM [3H]-
kainate, 20 mM CaCl2, and test compounds were incubated
on ice for 90 min.17,18 Separately a 20 nM [3H]glycine solution,
in 50 mM Hepes KOH, pH 7.4, was incubated on ice for 30
min.19 Nonspecific binding was determined by the presence of
1.0 mM glutamate. The incubation was terminated by rapid
vacuum filtration through GF/B filters followed by 2 × 2-mL
washes with appropriate ice-chilled buffer. Radioactivity was
determined by liquid scintillation techniques, and IC50 values
were determined by regression.
Ma xim a l Electr osh ock An ticon vu lsa n t Assa y. Male
CF-1 mice, 20-28 g, from Charles River Laboratories Portage,
MI, were used in all experiments. All dose-response testing
was conducted at the previously determined time of peak drug
effect. Groups of 10 mice each were given various intravenous
doses of antagonist dissolved in 5% dextrose in distilled water
and administered into the retro-orbital sinus. Electroshock was
delivered by corneal electrodes with 60-Hz alternating current
(50 mA root-mean-squared for 0.2 s) using a constant-current
stimulator (Wahlquist Instruments, Salt Lake City, UT).
Untreated or vehicle-treated mice had reliable tonic extensor
seizures with rearward extension of hind limbs for more than
3 s. Dose-effect experiments were analyzed by probit analy-
sis.20 These methods are similar to those published else-
where.21
Solu bility Mea su r em en ts. Equilibrium solubilities of
PNQX and its analogues (Table 2) were determined by adding
an excess amount of compound to pH 7.4, 50 mM phosphate
buffer followed by 5 min of sonification and overnight equili-
bration by moderate stirring at ambient temperature. Sample
aliquots were removed and filtered through a 0.45-µm nylon
filter. Solutions with known concentrations (standard solu-
tions) were prepared for each compound to be used to quantify
saturated solutions. Saturated and standard solutions were
then analyzed by UV spectrophotometry with an HP 8452A
diode array spectrophotometer using a quartz cell with a 1-cm
path length. To keep the relationship between absorbance and
concentration linear, some saturated solutions required dilu-
tion. The scanned wavelength range was 230-400 nm. Linear
regression was performed on absorbance and concentration
data of the standard solutions at an appropriately chosen
wavelength, typically near 330 nm. This allowed the concen-
trations of the saturated solutions, i.e., solubilities, to be
calculated at the corresponding wavelength.
None of the analogues (4, 6, 9, and 11) are as planar
or rigid as PNQX. Although all but one is predicted to
be primarily zwitterionic at pH 7.4, the solubilities of
those zwitterions are much higher than that of PNQX
(Table 2). This is likely due to the proximity of the
Ch em istr y. All starting materials were purchased from
Aldrich and used without additional purification. 5-Bromo-2-

2270 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 12
Brief Articles
methyl-7-nitro-1,2,3,4-tetrahydroisoquinolin-8-ylamine (1) was
synthesized as previously described.13 1H NMR spectra studies
were recorded at 400 MHz in CDCl3, DMSO-d6, or CD3OD.
Chemical shifts are referenced to the residual proton signal
for CDCl3 (7.26 ppm), DMSO-d6 (2.49 ppm), or CD3OD (3.3
ppm).
was used without further purification: 1H NMR (DMSO-d6) δ
1.09 (t, 3H, J ) 7.5 Hz), 2.17 (s, 3H), 2.66 (q, 2H, J ) 7.4 Hz,
J ) 11.1 Hz), 3.39 (s, 2H), 3.76 (s, 2H), 6.96 (d, 1H, J ) 8.2
Hz), 7.04 (d, 1H, 8.2 Hz), 11.90 (s, 1H), 11.98 (bs, 1H); MS
(CI) m/z 292 (M + 1).
[(6-Eth yl-7-n itr o-2,3-d ioxo-1,2,3,4-tetr a h yd r oqu in oxa -
lin -5-ylm eth yl)m eth ylam in o]acetic Acid Am m on iu m Salt
(9). A dark solution of 8 (0.368 g) in concentrated H2SO4 (5
mL) was cooled to 5 °C, and KNO3 (0.5 g, 5 mmol) was added
under stirring. The reaction mixture was allowed to warm to
room temperature and quenched with ice after 12 h. The pale-
brown residue was filtered, and the mother liquor was made
basic by bubbling NH3. A pale-green precipitate of the am-
monium salt of 9 was obtained, which was filtered and air-
dried to yield 0.100 g (25.4%): mp >300 °C; 1H NMR (DMSO-
d6) δ 1.12 (t, 3H, J ) 7.3 Hz), 2.01 (s, 3H), 2.8 (q, 2H), 2.92 (s,
2H), 3.70 (s, 2H), 7.61 (s, 1H); MS m/z 337 (M + H).
[(6-Eth yl-2,3-d ioxo-1,2,3,4-tetr a h yd r oqu in oxa lin -5-yl-
m eth yl)m eth yla m in o]a cetic Acid Eth yl Ester (10). A
solution of 7 (0.57 g, 1.9 mmol) and dimethyl oxalate (0.472 g,
4 mmol) in EtOH (5 mL) was heated to reflux for 16 h. The
solvent was partially evaporated, and the reaction mixture was
cooled to give 0.828 g of a buff precipitate. The product was
purified by column chromatography (SiO2, petroleum ether:
EtOAc, 95:5 to 75:25) to give 0.274 g (45%) of 10: 1H NMR
(CDCl3) δ 1.17 (t, 3H, J ) 7.8 Hz), 1.33 (t, 3H, J ) 7.08 Hz),
2.32 (s, 3H), 2.72 (q, 2H, J ) 7.5 Hz, J ) 15.1 Hz), 3.37 (s,
2H), 3.86 (s, 2H), 4.31 (q, 2H, J ) 7.1 Hz, J ) 14.2 Hz), 6.98
(d, 1H, J ) 8.3 Hz), 7.14 (d, 1H, J ) 8.3 Hz), 10.8 (bs, 1H),
12.3 (bs, 1H); MS m/z 320 (M + H).
E t h yl 2-{[(2-Am in o-5-b r om o-3-n it r o-6-vin ylp h en yl)-
m eth yl]m eth yla m in o}a ceta te (2). To a suspension of 1 (2.86
g, 10 mmol) and potassium carbonate (2.76 g, 20 mmol) in
acetone (50 mL) was added bromoacetic acid ethyl ester (1.67
g, 10 mmol). The reaction mixture was stirred under reflux
until TLC (SiO2, petroleum ether:EtOAc, 1:1) indicated comple-
tion. Volatile materials were evaporated under vacuum, and
water (150 mL) was added to the yellow residue. Product was
extracted with EtOAc (2 × 150 mL), and the EtOAc extracts
were washed with water (2 × 50 mL) and dried over MgSO4.
Product was purified by chromatography (SiO2, petroleum
ether:EtOAc, 95:5 to 75:25) to give 2.25 g (60%): mp 68-70
1
°C; H NMR (CDCl3) δ 1.29 (t, 3H, J ) 7.2 Hz), 2.26 (s, 3H),
3.26 (s, 2H), 3.86 (s, 2H), 4.21 (q, 2H), 5.23 (dd, 1H, J ) 1.2
Hz, J ) 18 Hz), 5.69 (dd, 1H, J ) 1.2 Hz, J ) 11.5z), 6.57 (dd,
1H, J ) 11.6 Hz, J ) 18 Hz), 8.05 (bs, 2H), 8.36 (s, 1H); MS
(CI) m/z 373 (M + H).
[(2,3-Dia m in o-5-b r om o-6-vin ylb en zyl)m et h yla m in o]-
a cetic Acid Eth yl Ester (3). A suspension of 2 (0.5 g, 1.34
mmol) and RaNi (0.5 g) in THF (75 mL) was hydrogenated
(H2, 50 psi) in a Parr apparatus. The reaction mixture was
filtered, and the filtrate was evaporated to give 0.45 g of the
diamine 3, which was used without further purification: 1H
NMR (CDCl3) δ 1.28 (t, 3H, J ) 7.2 Hz), 2.27 (s, 3H), 3.22 (s,
2H), 3.79 (s, 2H), 4.03 (bs, 2H), 4.17 (bs, 4H), 4.19 (q, 2H),
5.14 (dd, 1H, J ) 1.8 Hz, J ) 16 Hz), 5.54 (dd, 1H, J ) 1.8 Hz,
J ) 10 Hz), 6.59 (m, 1H), 6.93 (s, 1H).
[(6-Eth yl-7-n itr o-2,3-d ioxo-1,2,3,4-tetr a h yd r oqu in oxa -
lin -5-ylm eth yl)m eth ylam in o]acetic Acid Eth yl Ester (11).
To a stirred solution of 10 (0.175 g, 0.57 mmol) in concentrated
H2SO4 (2 mL) at 5 °C was added KNO3 (0.071 g, 0.71 mmol).
The reaction mixture was stirred for 16 h and poured over ice.
The brown precipitate was filtered, washed with water, and
purified by column chromatography (SiO2, CHCl3:MeOH, 95:5
to 80:20) to give 0.135 g (65%) of 11: mp 110 °C effervescence,
[(7-Br om o-6-vin yl-2,3-d ioxo-1,2,3,4-tetr a h yd r oqu in ox-
a lin -5-ylm eth yl)m eth yla m in o]a cetic Acid Hyd r och lor id e
(4). A solution of 3 (0.45 g) and oxalic acid (0.327 g, 2.6 mmol)
in aqueous 2 N HCl (20 mL) was heated to 80 °C. After 4 h,
the reaction mixture was cooled and triturated with a spatula
to give an off-white precipitate, which was filtered, dried, and
recrystallized from a water:acetone mixture to give 0.162 g of
1
185 °C dec; H NMR (CDCl3) δ 1.22 (t, 3H, J ) 7.3 Hz), 1.28
(t, 3H, J ) 7.3 Hz), 2.28 (s, 3H), 2.77 (q, 2H, J ) 7.3 Hz, J )
14.8 Hz), 3.37 (s, 2H), 3.88 (s, 2H), 4.28 (q, 2H, J ) 7.1 Hz, J
) 14.4 Hz), 7.69 (s, 1H), 11.37 (bs, 1H), 12.66 (s, 1H); MS m/z
365 (M + H).
1
4 (31%): mp >300 °C; H NMR (D2O) δ 2.37 (s, 3H), 2.49 (s,
2H), 3.69 (s, 2H), 4.27 (s, 2H), 5.25 (d, 1H, J ) 17.86 Hz), 5.48
(d, 1H, J ) 11.54 Hz), 6.77 (dd, 1H, J ) 11.21 Hz, J ) 17.62
Hz), 7.45 (s, 1H); MS (CI) 368 (M + H).
[(7-Br om o-6-eth yl-2,3-d ioxo-1,2,3,4-tetr a h yd r oqu in ox-
a lin -5-ylm eth yl)m eth yla m in o]a cetic Acid (6). A solution
of 2 (0.475 g, 1.27 mmol) in EtOH (75 mL) was hydrogenated
(H2, 50 psi) in the presence of RaNi (0.2 g). The reaction
mixture was filtered, and the filtrate was evaporated to give
0.450 g of a white solid, [(2,3-diamino-5-bromo-6-ethylbenzyl)-
methylamino]acetic acid ethyl ester (5), which was used
without further purification. To a stirred solution of 5 (0.450
g) in 5 N HCl (5 mL) was added oxalic acid dihydrate (0.327
g, 2.6 mmol). The reaction mixture was heated to 90 °C for 16
h, after which additional oxalic acid dihydrate (0.163 g, 1.3
mmol) was added. On cooling, an off-white solid separated,
which was filtered, washed with water and methanol, and
Ack n ow led gm en t. The authors acknowledge Dr.
J ing Belfield for comprehensive literature searches and
Don J ohnson for high-pressure hydrogenation experi-
ments. We thank our Analytical Department colleagues
for providing the analytical data and Dr. Christopher
Bigge for useful discussions, particularly on his experi-
ences with PNQX.
Refer en ces
(1) Lees, G. J . Therapeutic potential of AMPA Receptor ligands in
neurological disorders. CNS Drugs 1995, 5, 51.
1
dried to give 0.174 g (34%) of 6: mp >245-248 °C; H NMR
(2) Collingridge, G.; Lester, R. Excitatory amino acid receptors in
the vertebrate nervous system. Pharmcol. Rev. 1989, 40, 145.
(3) Muir, K. W.; Lees, K. R. Clinical experience with excitatory
amino acid antagonist drugs. Stroke 1995, 26, 503.
(4) Doble, A. Excitatory amino acid receptors and neurodegenera-
tion. Therapie 1995, 50, 337.
(CD3OD) δ 1.15 (t, 3H, J ) 7.5 Hz), 2.84 (s, 3H), 3.11 (q, 2H),
4.32 (s, 2H), 4.71 (s, 2H), 7.56 (s, 1H); MS (CI) 326 (M - CO2H
+ 1); HRMS calcd for C14H16BrN3O4 (M + 1)+ ) 370.0402,
found 370.0385.
[(2,3-Dia m in o-6-eth ylben zyl)m eth yla m in o]a cetic Acid
Eth yl Ester (7). A solution of 2 (0.5 g, 1.34 mmol) and KOAc
(0.13 g, 1.34 mmol) in EtOH was hydrogenated (H2, 50 psi)
over Pd/C (20%, 0.1 g) in a Parr apparatus. The reaction
mixture was filtered and the filtrate evaporated to give 0.42 g
of a semisolid, which was used without further purification.
[(6-Eth yl-2,3-d ioxo-1,2,3,4-tetr a h yd r oqu in oxa lin -5-yl-
m eth yl)m eth yla m in o]a cetic Acid (8). A solution of 7 (0.42
g) and oxalic acid dihydrate (0.327 g, 2.6 mmol) in aqueous 2
N HCl (20 mL) was heated to 90 °C. The reaction mixture was
stirred for 16 h and cooled. A buff precipitate separated, which
was filtered and air-dried to yield 0.368 g (84%) of 8, which
(5) Meldrum, B.; Garthwaite, J . Excitatory amino acid neurotoxicity
and neurodegenrative disease. Trends Pharmacol. Sci. 1990, 11,
379.
(6) Nakanishi, S. Molecular diversity of glutamate receptors and
implications for brain function. Science 1992, 258, 597.
(7) Monoghan, D. T.; Bridges, R. D.; Cotman, C. W. The excitatory
amino acid receptors: their classes, pharmacology and distinct
properties in the function of the central nervous system. Annu.
Rev. Pharmacol. Toxicol. 1990, 29, 365-402.
(8) Bigge, C. F.; Boxer, P. A.; Ortwine, D. F. AMPA/KA receptors.
Curr. Pharm. Des. 1996, 2, 397-412.
(9) Bigge, C. F.; Nikam, S. S. AMPA receptor agonists, antagonists
and modulators: their potential for clinical utility. Exp. Opin.
Ther. Patents 1997, 7, 1099-1114.

Brief Articles
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 12 2271
(10) Sheardown, M. J .; Nielson, E. O.; Hanson, A. J .; J acobsen, P.;
Honore, T. 2,3-Dihydroxy-6-nitro-7-sulfamoylbenzo[f]quinoxa-
line: a neuroprotectant for cerebral ischemia. Science 1990, 247,
571-574.
(11) Watjen, F.; Bigge, C. F.; J ensen, L. H.; Boxer, P. A.; Lescosky,
L. J .; Nielsen, E. O.; Malone, T. C.; Campbell, G. W.; Coughen-
our, L. L.; Rock, D. M.; Drejer, J .; Marcoux, F. W. NS 257
(1,2,3,6,7,8-hexahydro-3-(hydroxyimino)-N,N,7-trimethyl-2-
oxobenzo[2,1-b:3,4-c′]dipyrrole-5-sulfonamide is a potent, sys-
temically active AMPA receptor antagonist. Bioorg. Med. Chem.
Lett. 1994, 4, 371-376.
(12) Ohmori, J .; Sakamoto, S.; Kubota, H.; Shimizu-Sasamata, M.;
Okada, M.; Kawasaki, S.; Hidaka, K.; Togami, J .; Furuya, T.;
Murase, K. J . Med. Chem. 1994, 37, 467-475.
(13) Bigge, C. F.; Malone, T. C.; Boxer, P. A.; Nelson, C. B.; Ortwine,
D. F.; Schelkun, R. M.; Retz, D. M.; Lescosky, L. J .; Borosky, S.
A.; Vartanian, M. G.; Schwarz, R. D.; Campbell, G. W.; Ro-
bichaud, L. J .; Watjen, F. J . Med. Chem. 1995, 38, 3720-3740.
(14) Schielke, G. P.; Kupina, N. C.; Boxer, P. A.; Cordon, J . J .; Bigge,
C. F.; Marcoux, F. W. Delayed treatment with the novel AMPA
receptor antagonist PD152247 (PNQX), reduces lesion size
following focal ischemia and reperfusion in the rat. J . Cereb.
Blood Flow Metab. 1995, 15 (Suppl. 1), S372.
(16) Nielson, E. O.; Cha, J . J .; Honore, T.; Penny, J . B.; Young, A. B.
Thiocyanate stabilizes AMPA receptor binding to the quisqualate
receptor. Eur. J . Pharmacol. 1988, 157, 197-203.
(17) Nielson, E. O.; Cha, J . J .; Honore, T.; Penny, J . B.; Young, A. B.
Thiocyanate stabilizes AMPA receptor binding to the quisqualate
receptor. Eur. J . Pharmacol. 1988, 157, 197-203.
(18) J ohansen, T. H.; Drejer, J .; Watjen, F.; Nielsen, E. O. A novel
non-NMDA receptor antagonist shows selective displacement of
low-affinity [3H]kainate binding. Eur. J . Pharmacol. 1993, 246,
195-204.
(19) London, E. D.; Coyle, J . T. Specific binding of [3H]kainic acid to
receptor sites in rat brain. Mol. Pharmacol. 1979, 15, 492-505.
(20) J ones, S. M.; Snell, L. D.; J ohnson, K. M. Characterization of
the binding of radioligands to the NMDA, phencyclidine, and
glycine receptors in buffy coat membranes. J . Pharmacol.
Methods 1989, 21, 161-168.
(21) Litchfield, J . T.; Wilcoxon, F. A simplified method of evaluation
dose-effect experiments. J . Pharmacol. 1949, 96, 99-113.
(22) Krall, R. L.; Penry, J . K.; White, B. G.; Kupferberg, H. J .;
Swinyard, E. A. Antiepileptic drug development: II. Anticon-
vulsant drug screening. Epilepsia 1978, 19, 409-428.
(15) Acklin, P.; Allgeier, H.; Auberson, Y. P.; Bischoff, S.; Ofner, S.;
Sauer, D.; Schmutz, M. Bioorg. Med. Chem. Lett. 1998, 8, 493-
498 and references cited therein.
J M980455N