Structure-activity relationships of 1,4-dihydro-(1H,4H)-quinoxaline-2,3-diones as N-methyl-D-aspartate (glycine site) receptor antagonists. 1. Heterocyclic substituted 5-alkyl derivatives

Article abstract of DOI:10.1021/jm001124p

A series of 6,7-dichloro-1,4-dihydro-(1H, 4H)-quinoxaline-2,3-diones (1-17) were prepared in which the 5-position substituent was a heterocyclylmethyl or 1-(heterocyclyl)-1-propyl group. Structure-activity relationships were evaluated where binding affini

Full text of DOI:10.1021/jm001124p

J . Med. Chem. 2001, 44, 1951-1962
1951
Str u ctu r e-Activity Rela tion sh ip s of 1,4-Dih yd r o-(1H,4H)-qu in oxa lin e-2,3-d ion es
a s N-Meth yl-D-a sp a r ta te (Glycin e Site) Recep tor An ta gon ists. 1. Heter ocyclic
Su bstitu ted 5-Alk yl Der iva tives
M. J onathan Fray,* David J . Bull, Christopher L. Carr, Elisabeth C. L. Gautier, Charles E. Mowbray, and
Alan Stobie
Department of Discovery Chemistry, Pfizer Global Research and Development, Sandwich, Kent CT13 9NJ , U.K.
Received November 27, 2000
A series of 6,7-dichloro-1,4-dihydro-(1H, 4H)-quinoxaline-2,3-diones (1-17) were prepared in
which the 5-position substituent was a heterocyclylmethyl or 1-(heterocyclyl)-1-propyl group.
Structure-activity relationships were evaluated where binding affinity for the glycine site of
the N-methyl-D-aspartate (NMDA) receptor was measured using the specific radioligand [3H]-
L-689,560, and functional antagonism was demonstrated by inhibition of NMDA-induced
depolarizations of rat cortical wedges. The ability to prevent NMDA-induced hyperlocomotion
in mice in vivo was measured for selected compounds. Binding affinity increased significantly
if the heterocyclic group, e.g. 1,2,3-triazol-1-yl could participate in accepting a hydrogen bond
from the receptor. It was difficult to obtain compounds with adequate aqueous solubility and
strategies to improve it were investigated. The most potent compound in this series, 6,7-dichloro-
5-[1-(1,2,4-triazol-4-yl)propyl]-1,4-dihydro-(1H, 4H)-quinoxaline-2,3-dione (17) (binding IC50
)
2.6 nM; cortical wedge EC50 ) 90 nM), inhibited NMDA-induced hyperlocomotion in mice (6/9
protected at 20 mg/kg iv). Pharmacokinetic parameters, including extent of brain penetration,
for 11 and 17 are reported.
Ch a r t 1
In tr od u ction
L-Glutamic acid is an important excitatory amino acid
in the central nervous system. It is essential at N-
methyl-D-aspartate (NMDA) receptors for both glycine
and glutamic acid to bind to their respective sites for
receptor activation to occur, and thus glycine antago-
nists are able to function as noncompetitive modulators
of glutamic acid-induced responses in vitro and in
vivo.1-5 Overstimulation of NMDA receptors has been
implicated in a number of pathological conditions in-
volving neuronal death and degeneration and has driven
the search for NMDA antagonists as possible therapies
for thrombo-embolic stroke, traumatic head injury,
Parkinson’s, Huntington’s, and Alzheimer’s diseases,
epilepsy, and schizophrenia.1,2,6,7 Antagonists at the
glycine binding site of the NMDA receptor may have
advantages over competitive antagonists of glutamic
acid (e.g. selfotel (CGS 19755) and D-CPPene (SDZ EAA
494)) and drugs which recognize the ion-channel binding
site (e.g. dizocilpine (MK801) and aptiganel (CNS 1102))
due to a more attractive side effect profile.6 However,
it has, in general, been very difficult to identify com-
pounds with sufficient potency in animal models of
stroke to warrant further progression, primarily because
glycine antagonists are acidic and the structural fea-
tures which favor high potency tend to limit blood-brain
barrier penetration and hence in vivo activity. Never-
theless, several glycine antagonists have been pro-
gressed to clinical trials (see Chart 1). Satisfactory
Phase IIa safety and toleration studies with ACEA 1021
(licostinel) and GV150526A (gavestinel) have been
reported,8,9 but recently both compounds have been
dropped from development.
A large number of papers describe structure activity
relationships (SARs) for glycine antagonists and several
structural series with nanomolar binding affinities have
been discovered.2,10-12 These studies have helped de-
velop a model of the glycine binding site (Figure 1). This
diagram shows schematic binding interactions for 6,7-
dichloro-1,4-dihydro-2,3-quinoxalinedione (DCQX) and
5,7-dichlorokynurenic acid. The charge-charge interac-
tion and optimum filling of the lipophilic pocket are very
important for achieving high potency. SARs of both
quinoxalinediones and kynurenic acids can be rational-
ized using the same receptor model by accommodation
of the chlorines in the 6 and 7 positions of the quinoxa-
linedione binding to the same region of the receptor
surface as the 5-chlorine and 6-hydrogen of the kynurenic
10.1021/jm001124p CCC: $20.00 © 2001 American Chemical Society
Published on Web 05/05/2001

1952 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 12
Fray et al.
Ch a r t 2
F igu r e 1. Schematic model of the glycine binding site with
6,7-dichloroquinoxalinedione (DCQX) (top) and 5,7-dichlo-
rokynurenic acid docked into place.
acid. This overlap led us to speculate whether some of
the features of both series could be combined, and we
recently reported the SARs of a series of 3-hydroxy-
2(1H)-quinolinones13 in which replacement of one of the
nitrogens of a quinoxalinedione by a carbon permitted
the attachment of a suitable substituent from the
4-position without loss of the acidic functionality (see
Chart 2).14
dimethoxy derivative to solve these problems and permit
a variety of chemical manipulations at the 5-position.
Compounds 1-17 (Table 1) were readily prepared by
acidic hydrolysis of 18-34 (Scheme 1, Method A).
Intermediates 18-25 were most conveniently pre-
pared by reaction of the benzylic bromide 35 with
heterocycles such as imidazole, pyrazole, and triazole
(Scheme 2, Method B). 3-Methyl-1,2,4-triazole afforded
a mixture of regioisomers (24 and 25), whereas a single
product was isolated from the reactions of 1,2,3-triazole
and 1,2,4-triazole. Performing the alkylation with 1,2,4-
triazole under melt conditions (Method C)15 gave access
to the 1,2,4-triazol-4-yl isomer (22).
Compound 35 was prepared in seven steps from
commercially available 2,4,5-trichloronitrobenzene (36)
(Scheme 2). First, a vicarious nucleophilic aromatic
substitution reaction16 with tert-butyl chloroacetate gave
37, which underwent concurrent replacement of the
ortho chlorine and hydrolysis and decarboxylation of the
ester to give the nitrotoluene 38 on treatment with
concentrated aqueous ammonia at high temperature.
The moderate yield for this reaction is partly due to the
vigorous conditions and the formation of side products
arising from displacement of the para chlorine, but this
two-step sequence17 was preferred to the seven-step
literature route.18 Next, reduction of the nitro group and
condensation with oxalic acid19 gave 39. Treatment with
thionyl chloride gave the 2,3-dichloroquinoxaline which
reacted smoothly with sodium methoxide to give 40.
Finally, bromination with N-bromosuccinimide afforded
35.
While good in vitro binding affinity was obtained, the
new hydroxyquinolone series lacked activity in vivo.
Drug metabolism studies showed that compounds typi-
cally possessed very short plasma half-lives in the rat
(t1/2 < 0.2 h for the two compounds illustrated). Glucu-
ronidation of the 3-hydroxyl was evident in rat liver
microsome preparations in vitro, and we also speculate
that active hepatic uptake might also have contributed
to the rapid clearance rate. After many unsuccessful
efforts to improve pharmacokinetic parameters of the
hydroxyquinolinones, we decided to investigate qui-
noxalinediones, since studies with DCQX showed it to
be stable toward glucuronidation by rat liver mi-
crosomes in vitro. Molecular modeling, based on the
receptor model shown in Figure 1, suggested that the
H-bond donor interaction could be reached easily from
the 5-position. In addition, the greater potency of ACEA-
1021 compared to DCQX (Table 1) could be rationalized
in terms of a hydrogen bond acceptor interaction by the
5-nitro group. We therefore prepared a range of novel
derivatives, all of which were designed to target this
interaction. In this paper, we report SARs for a series
of compounds (1-17) (Chart 2) in which the hydrogen
bond acceptor is an aromatic heterocycle.
Ch em istr y
Quinoxalinediones are generally very insoluble so that
their handling and purification are difficult. We thus
sought to protect the dione functionality as the 2,3-
Two 1,2,3-triazole derivatives (26 and 27) bearing
amine-containing side chains were prepared. Thus,
reaction of 35 with methyl 1,2,3-triazole-4-carboxylate20

1,4-Dihydro-(1H,4H)-quinoxaline-2,3-diones
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 12 1953
Ta ble 1. NMDA Receptor Binding Affinity, Functional Potency in Vitro, and Inhibition of NMDA Induced Hyperlocomotion in the
Mouse by Quinoxalinedione Derivatives
a
Displacement of [3H]-L-689,560 from rat cortical membranes. Results are the average of two determinations, and a 2-fold difference
b
between compounds should not be regarded as significant. Inhibition of NMDA-induced depolarizations in rat cortical wedges. Results
are the average of two determinations, and a 2-fold difference between compounds should not be regarded as significant. c Ability to
d
prevent NMDA-induced hyperlocomotion in the mouse following iv administration. NT ) not tested
(Scheme 3) followed by reduction of 41 to the aldehyde
42 and reductive amination with diethylamine or N-
methylpiperazine gave 26 and 27, respectively.
Compound 35 was also a useful intermediate for the
synthesis of carbon-linked heterocyclic derivatives 28-
30 (Scheme 4). We found that the benzylic zinc reagent
could be formed under mild conditions by treating 35
with activated zinc dust in THF at room temperature.21
Cross-coupling with chloracetyl chloride22 (at room
temperature) or 2-bromopyridine23 (reflux) with tetraki-
stri(2-furylphosphine)palladium(0) catalysis24 gave chlo-
roketone 43 and the pyridine 29, respectively. Reaction
of 43 with 2-amino-5,6-dihydro-4H-1,3-thiazine25 fol-
lowed by Raney nickel desulfurization of 44 was used

1954 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 12
Fray et al.
Sch em e 1a,b
Sch em e 2a
Reagents and conditions: a: ClCH2CO2But, KOBut, THF, -40
°C, 30 min; b: concentrated aqueous NH3, 2-methoxyethanol,
autoclave, 150 °C, 72 h; c: Na2S2O4, KHCO3, H2O/MeOH, 20 °C,
30 min; d: (CO2H)2, 4 M HCl(aq), reflux, 6 h; e: SOCl2, catalyst
DMF, reflux, 3 h; f: NaOMe, THF/MeOH, 20 °C, 1 h; g: N-
bromosuccinimide, catalyst AIBN, CH3CCl3, sunlamp, reflux, 4 h;
Method B: Het-H, K2CO3, CH3CONMe2 or THF, 50 °C, 18 h.
a
a
b
Reagents and conditions: 2 M HCl(aq), dioxane, reflux. For
Sch em e 3a
details of R and Het, see Table 1. c Solubility in water, pH 7.
Solubility in water as hydrochloride salt, pH 4-5. e No combus-
d
tion analysis obtained due to the high level of nitrogen present.
Compound characterized by spectroscopic means: 1H NMR (300
MHz, DMSO-d6) 4.23 (3H, s), 4.62 (2H, s), 7.30 (1H, s), 1.66 (1H,
s), 12.1 (1H, s), m/z (thermospray) 344 (MNH4+). fSolubility in 0.1
M aqueous D-arginine, pH 9.
to ensure formation of the N-propylimidazole 28 regi-
oselectively. Preparation of the tetrazole 30 proceeded
via the nitrile 45, 1,3-dipolar cycloaddition with azidot-
ributylstannane and N-methylation of 46. Compound
30 was obtained together with its 1-methyl isomer, and
the two isomers were separated by column chromatog-
raphy. Measurement of nuclear Overhauser enhance-
ments between the quinoxaline 5-position methylene
hydrogens and the tetrazole N-methyl hydrogens per-
mitted a clear structural assignment for the two iso-
mers.
We also wished to investigate the effect of branching
next to the heterocycle (Scheme 5). We prepared the
necessary bromopropyl intermediate 51 in eight steps
from 5-nitroquinoxalinedione 47.26 Thus, chlorination
with thionyl chloride, reduction of the 5-nitro group, and
nucleophilic displacement of the 2- and 3-position chlo-
rines gave the 2,3-dimethoxy derivative 48. A Sandm-
eyer reaction gave the iodide 49,27 which underwent
cross-coupling with tributylvinylstannane28 and subse-
quent ozonolysis and addition of ethylmagnesium bro-
mide to give the alcohol 50. Finally, bromination29 of
the alcohol 50 afforded bromide 51, whose proton NMR
spectrum was of interest since two diastereomers were
evident. This was presumably due to hindered rotation
of the 5-substituent bond and the stereogenic center
contained within that substituent. Presumably, the
6-position chlorine restricts rotation of the 5-position
a
Reagents and conditions: a: methyl 1,2,3-triazole-4-carboxy-
late, K2CO3, THF, reflux, 18h, then separation of the isomers by
chromatography; b: DIBAL-H, CH2Cl2, -78 °C, 2.25 h; c: amine,
NaBH(OAc)3, HOAc, THF, 20 °C, 2 h.
substituent when the latter is sufficiently bulky. Nu-
cleophilic substitution of the bromide was problematic
presumably due to steric hindrance and competing
elimination, but satisfactory yields of products 31-34
were obtained using a melt of the heterocycle (Method
C).

1,4-Dihydro-(1H,4H)-quinoxaline-2,3-diones
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 12 1955
Sch em e 4a
a
Reagents and conditions: a: zinc dust, THF, 20 °C, ultrasound, then ClCH2COCl, Pd2(dba)3, tri(2-furyl)phosphine, THF, 20 °C, 45
min; b: 2-amino-5,6-dihydro-4H-1,3-thiazine, NBun Br, K2CO3, DMF, 80 °C, 2 h; c: Raney nickel, EtOH, reflux, 2 days. d: zinc dust,
4
THF, 20 °C, ultrasound, then 2-bromopyridine, Pd2(dba)3, tri(2-furyl)phosphine, THF, 65 °C, 60 min; e: NaCN, i-PrOH/H2O, reflux, 1 h;
f: azidotributylstannane, toluene, reflux, 42 h; g: CH3I, K2CO3, DMF, 20 °C, 2 h, then separation of tetrazole isomers by chromatography.
Resu lts a n d Discu ssion
triazole 5 was less potent than the imidazole 1. Intro-
duction of a methyl group adjacent to the methylene (as
in 6 and 8) was not well-tolerated in contrast to
attachment to the 3-position relative to the methylene.
Thus, the 3-methyltriazole 7 was equipotent with the
imidazole 1.
Compounds 1-17 were first evaluated as glycine site
ligands at the NMDA receptor by measuring their
ability to displace the specific glycine antagonist [3H]-
L-689,560 from thoroughly washed rat cortical mem-
branes30 (see Table 1). Functional activity in vitro was
measured from compounds’ ability to inhibit NMDA-
induced depolarizations of rat cortical wedges.31 Com-
parative data, generated in our laboratories, for three
literature glycine antagonists, DCQX,26 ACEA 1021
(47),26 and L-687,414,32 are also shown in Table 1.
The in vitro SARs of N-linked heterocyclic derivatives
1-8 will be discussed first. The imidazole derivative 1
had good binding affinity (IC50 ) 23 nM), suggesting
that the desired hydrogen bonding interaction with the
receptor had indeed been made. The affinity of the
pyrazole 2 was 5-fold greater, which could be due to
moving the sp2 nitrogen atom into a better position to
make the hydrogen bond. The 1,2,4-triazole 3 was only
as potent as the imidazole, which can be rationalized
because its dominant H-bond acceptor is the 3-position
nitrogen. Enhancing the H-bond acceptor power of the
heterocycle as in the 1,2,3-triazole increased binding
affinity significantly (IC50 ) 1 nM). The isomeric 1,2,4-
Compounds 1, 4, 5, and 7 were found to be functional
glycine antagonists in vitro with 4 being 3-fold more
potent than ACEA 1021. The falloff in potency between
the binding and the cortical wedge assays is probably
due to the high concentration of endogenous glycine
present in the cortical wedge preparation which must
be overcome by the antagonist and high protein binding
of the test compounds. However, it is notable that the
falloff in potency is only about 13-fold for L-687,414
which has very low plasma protein binding and high
water-solubility. The potency of several of the com-
pounds 1-8 could not be measured in the cortical wedge
assay due to low solubility. We therefore sought to
append side-chains to the 4-position of the heterocycle
to enhance solubility. Compounds 9 and 10 are repre-
sentative of a much larger group of compounds we made
in which the heterocycle (1,2,3-triazole, 1,2,3-benzot-
riazole, pyrazole), the amine, and its distance from the

1956 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 12
Fray et al.
Sch em e 5a
a
Reagents and conditions: a: SOCl2, catalyst DMF, reflux, 3
F igu r e 2. Schematic model of the glycine binding site with
compound 15 (top) and 7-chloro-5-iodokynurenic acid docked
into place.
h; b: SnCl2‚2H2O, EtOAc, reflux, 4 h; c: NaOMe, MeOH, reflux,
30 min. d: HCl, NaNO2, acetone/water, 0 °C, then KI, 10 °C, 30
min; e: tributylvinylstannane, LiCl, PdCl2‚2PPh3, DMF, 100 °C,
1.5 h; f: ozone, CHCl3, -60 °C, then PPh3, 16 h; g: EtMgBr, THF,
20 °C, 30 min; h: CBr4, PPh3, CH2Cl2, 20 °C, 18 h.
triazole 15 being the best (IC50 ) 0.75 nM). In addition,
compounds 14 and 17 were very potent functional
antagonists. The increase in potency relative to the
unbranched analogues 1, 3, and 5 may result from an
increased conformational restriction which holds the
heterocycle in a better orientation relative to the recep-
tor, or possibly because the ethyl group helps to fill the
lipophilic pocket occupied by the 6-chlorine atom (see
Figure 2). It is known in the kynurenic acid series that
replacement of the 5-chlorine by iodo or ethyl results
in improved binding affinity.35 In addition aqueous
solubility was improved, for example 14 (as hydrochlo-
ride salt in water, pH 4) and 17 (in aqueous arginine,
pH 9) could be dissolved at g 5 mg/mL.
Our series of compounds bears some similarity to a
group of substituted 5-(aminomethyl)quinoxalinediones
such as 52 and 53 (Chart 2)36 in which the amino,
pyridyl, or carboxylate groups might be able to function
as a hydrogen bond acceptors. Whereas compound 52
is reported to have little affinity for the NMDA (glycine
site) (0% inhibition of [3H]-5,7-dichlorokynurenic acid
at 1 µM), introduction of a carboxylic acid group, as in
53, increased potency significantly (binding IC50 ) 30
nM). As part of this work we prepared benzotriazole 54
(Chart 2), which has similar potency to the 1,2,3-triazole
4, and possesses a more marked structural resemblance
to 53. Compounds 52 and 53 also have affinity for the
AMPA receptor ([3H]-AMPA binding IC50 1-2 µM).36 We
have measured the AMPA receptor binding affinity
(displacement of [3H]-AMPA from rat cortical mem-
branes) for a number of our compounds. Of these, only
compounds 3 (IC50 ) 0.25 µM) and 17 (IC50 ) 2.3 µM)
had measurable affinity (IC50 > 10 µM for compounds
1, 2, 4, 7, 9, and 11).
heterocycle were varied. In general, the potency did not
vary much with structure, and the amine side chains
were well-tolerated by the receptor. However, functional
potency and solubility improved significantly. Thus,
compound 10 possessed IC50 ) 7.6 nM for binding and
EC50 ) 88 nM in the cortical wedge assay, data which
compare favorably with L-689,560 (IC50 ) 4 nM, EC50
) 230 nM),33 one of the most potent known glycine
antagonists in vitro, and its solubility in water (as
hydrochloride salt, pH 5) was >10 mg/mL.
The in vitro SARs for the carbon-linked heterocycles
11-13 will be discussed next. The imidazole 11 was
designed to retain the sp2 nitrogen atom adjacent to the
methylene linker. Imidazole is a powerful H-bond ac-
ceptor,34 and its moderate basicity was expected to help
increase aqueous solubility. Table 1 shows that 11
possesses good binding affinity and functional potency,
and aqueous solubility was excellent (>10 mg/mL as
hydrochloride salt, pH 4). In contrast, the 2-pyridyl and
tetrazolyl analogues 12 and 13 would be expected to
provide similar H-bond acceptor capabilities, but they
were significantly less potent than 11. In addition, they
were less soluble (<1 mg/mL).
Finally, the in vitro SARs for the branched alkyl
dervatives 14-17 will be discussed. We believed that
the poor aqueous solubility of the quinoxalinediones is
due to high crystal lattice strength, as evidenced by
their very high melting points. Thus, introducing groups
that lie above and below the plane of the quinoxaline
might be expected to disrupt the crystal packing some-
what. Examination of molecular models suggested
introduction of an ethyl group next to the heterocycle
might perform this function. Compounds 14-17 pos-
sessed very high binding affinities, with the 1,2,3-
Selected compounds were evaluated in vivo in the
mouse using a modification of the published method
(Table 1).37 In this assay, administration of NMDA (300

1,4-Dihydro-(1H,4H)-quinoxaline-2,3-diones
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 12 1957
mg/kg sc) to mice causes a series of well-characterized
behavioral changes, of which the first is hyperlocomo-
tion. The time to onset and duration of hyperlocomotion
was recorded (up to a maximum of 15 min) for groups
of animals receiving test compounds (as solutions
intravenously via a tail vein 15 min prior to NMDA
administration) or saline (as controls). Inhibition of
hyperlocomotion was recorded either as the number of
animals reaching 15 min post NMDA challenge at the
relevant dose, or an ED50 was determined as the derived
dose at which 50% of animals reach 15 min post NMDA
administration before the end of the hyperlocomotion
phase.
same species could have been undertaken, but we have
not pursued this line of enquiry.
In this paper, we describe a novel series of 5-substi-
tuted quinoxalinediones possessing potent activity in
vitro as glycine antagonists. However, obtaining good
in vivo activity was elusive. The majority of glycine
antagonists would appear to suffer from one or more
problems which handicap their utility in vivo, namely
acidity (low pKa), high plasma protein binding, high
hydrogen bond count, and rapid clearance. We consid-
ered that these could be overcome in the quinoxalinedi-
ones, for although the hydrogen bond count is high, we
and others13,26 had demonstrated that good potency in
vitro may be achieved in other chemical series with
moderate molecular weight and relatively low acidity
(pKa ) 6-7.5). Furthermore, compounds with low
lipophilicity or a basic center had suitable aqueous
solubility and could cross the blood-brain barrier. De-
spite a poor understanding of the in vivo performance
of 17, we nevertheless decided to persevere with the
quinoxalinediones and examine other structural series
that might address the deficiencies of the compounds
described here. These results will be detailed in a
following paper (Part 2).39
Only two compounds, the imidazole 11 and the
R-ethyl 1,2,4-triazole 17 were able to prevent NMDA-
induced hyperlocomotion. Raising the dose of 17 beyond
20 mg/kg iv did not increase the proportion of mice
protected. The stringency of the test is demonstrated
by the fact that ACEA 1021 also failed to inhibit the
NMDA challenge completely.
Rat pharmacokinetic parameters were measured for
11 and 17 to see if they threw light on the relatively
poor in vivo performance. Compound 11 possessed
moderate clearance (37 mL/min/Kg) and moderate
volume of distribution (1.0 L/Kg) leading to a short
plasma half-life (0.3 h). Penetration into brain tissue
following a 2 mg/kg iv dose was rapid, resulting in total
drug concentrations in plasma, brain tissue, and CSF
of 4.6 µM, 0.38 µM, and 0.26 µM at 0.1 h post dose and
2.4 µM, 0.39 µM, and <0.1 µM at 0.25 h post dose,
respectively. The brain/plasma concentration ratio of 11
was in the range 0.08-0.21, consistent with its lipophi-
licity (logD ) 2.4). However, the CSF/plasma concentra-
tion ratio (0.06) was measurable at a single time point
and was higher than expected considering the high
plasma protein binding (98%) of 11.
Exp er im en ta l Section
Biology. Liga n d Bin d in g Affin ity. The method of Grim-
wood et al.30 was followed with minor adjustments. Thoroughly
washed membrane protein from rat brain (cortex and hippoc-
ampus) was incubated for 90 min at 4 °C with test compounds
(dissolved in Tris-acetate buffer (pH 7.4) containing 1% DMSO)
and [3H]-L-689,560. Mixtures were filtered using a Brandel
cell harvester, and the filters were added to scintillation fluid
for quantification of remaining radioactivity. Displacement of
the radioligand using a range of concentrations of test com-
pounds was used to derive IC50 values. The binding affinity of
5,7-dichlorokynurenic acid was measured every time the assay
was performed and gave IC50 ) 210 ( 100 nM.
In comparison with compound 11, compound 17 (2
mg/kg iv) possessed lower clearance (15 mL/min/Kg) and
slightly greater volume of distribution (1.6 L/Kg) leading
to a plasma half-life of 1.2 h. Penetration into brain
tissue was also rapid, resulting in total drug concentra-
tions in plasma, brain tissue, and CSF of 7.5 µM, 0.58
µM, and 0.52 µM at 0.1 h post dose and 4.6 µM, 0.22
µM, and 0.26 µM at 0.25 h post dose, respectively. The
brain/plasma and CSF/plasma concentration ratios were
0.05-0.08 and 0.06-0.07, respectively (the two ratios
relate to the two time points). The close similarity of
the ratios is consistent with the measured logD ) 0.2
(octanol, pH 7.4) (i.e. almost equal partitioning between
brain cell membranes and extracellular fluid). The
magnitude of the ratios is consistent with the measured
rat plasma protein binding of 94%, since theoretical
models of brain penetration usually propose that only
an unbound drug may cross the blood-brain barrier.38
NMDA-In du ced Depolar ization s of Rat Cor tical Wedges.
The ability of compounds to inhibit NMDA-induced depolar-
ization of rat cortical tissue was determined using the cortical
wedge technique described by Harrison and Simmonds.31
Compounds were dissolved in 1% DMSO/Krebs buffer and
applied in increasing concentrations (up to 10 µM) for 15 min
prior to a standard challenge of 12.5 µM NMDA to evoke
depolarization. The percentage inhibition from control re-
sponses was used to calculate an IC50. Assays were performed
in duplicate and five-point dose-response curves measured.
Spontaneous activity was blocked by adding 0.1 µM tetrodot-
oxin to the Krebs buffer.
NMDA-In d u ced Hyp er locom otion in Mice. In vivo
potency was determined using a modification of the literature
method37 in which a larger dose of NMDA was given subcu-
taneously rather than intracerebroventricularly, and the hy-
perlocomotion phase (which precedes clonic convulsions) was
used as an endpoint. Mice (CD1, Charles River, 25-30 g) were
administered NMDA (300 mg/kg sc), and the time to onset and
duration of hyperlocomotion were recorded (up to a maximum
of 15 min). Animals were euthanased at the end of the
hyperlocomotion phase or at 15 min post NMDA administra-
tion, whichever was earliest. Test compounds were dissolved
in buffered saline (pH 7-9) or 0.1 M aqueous D-arginine and
given intravenously (up to 30 mg/Kg) via a tail vein 15 min
prior to NMDA administration, and usually 9-10 mice were
used per dose. Inhibition of hyperlocomotion was recorded
either as the number of animals reaching 15 min post NMDA
challenge at the relevant dose, or an ED50 was determined as
the derived dose at which 50% of animals reach 15 min post
NMDA administration before the end of the hyperlocomotion
phase.
Thus, the relatively poor in vivo activity with 11 in
the mouse probably stems from its short plasma half-
life, resulting in low brain concentrations at the time
of the NMDA challenge, rather than an inability to cross
the blood brain barrier. In contrast, the low in vivo
potency of 17 is less well understood. In principle,
pharmacokinetic experiments that compare free drug
concentrations in brain tissue and CSF with functional
potency at the NMDA receptor (as a surrogate for the
efficacious concentration) following various doses in the

1958 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 12
Fray et al.
1
Ch em istr y. Melting points were determined using open
glass capillary tubes and a Gallenkamp melting point ap-
paratus and are uncorrected. Spectroscopic data were recorded
on a Perkin-Elmer 983 (IR), Finnigan Mat. TSQ 7000 or a
Fisons Instruments Trio 1000 (MS), and Varian Unity Inova-
300 or Bruker AC300 (NMR) instruments and are consistent
with the assigned structures. Flash chromatography refers to
column chromatography on silica gel (Kieselgel 60, 230-400
mesh, from E. Merck, Darmstadt). Kieselgel 60 F254 plates from
E. Merck were used for TLC, and compounds were visualized
using UV light, 5% aqueous potassium permanganate or
chloroplatinic acid/potassium iodide solution. Where analyses
are indicated only by the symbols of the elements, results
obtained are within ( 0.4% of the theoretical values. In cases
where compounds were analyzed as hydrates, the presence of
water was evident in the enhanced peak due to water in the
NMR spectra. The purity of compounds was carefully assessed
using analytical TLC and proton NMR (300 MHz), and the
latter was used to calculate the amount of solvent in solvated
samples. In multistep sequences, the purity and structure of
intermediates were verified spectroscopically by NMR.
25: colorless solid, yield 23%; H NMR (300 MHz, CDCl3)
2.61 (3H, s), 4.10 (3H, s), 4.15 (3H, s), 5.90 (2H, s), 7.71 (1H,
s), 7.97 (1H, s). LRMS m/z (thermospray) 353 (M+).
5-Br om om et h yl-6,7-d ich lor o-2,3-d im et h oxyq u in oxa -
lin e (35). A solution of 2,4,5-trichloronitrobenzene (36) (103
g, 0.46 mol) and tert-butyl chloroacetate (79 mL, 0.55 mol) in
dry tetrahydrofuran (400 mL) was added dropwise over 30 min
to a solution of potassium tert-butoxide (128 g, 1.14 mol) in
dry tetrahydrofuran (800 mL) with stirring under nitrogen,
keeping the temperature at -40 °C. After the addition was
complete, the resulting dark blue solution was stirred for a
further 30 min. The mixture was poured into 0.5 M hydro-
chloric acid (2 L), and the product was extracted into ethyl
acetate (2.5 and 1 L). The combined organic solutions were
dried (MgSO4) and evaporated onto silica gel (70-200 µ, 200
g). The silica gel was applied to the top of a silica gel
chromatography column (800 g), and the product was eluted
using a hexane/ethyl acetate gradient. Product-containing
fractions were combined and evaporated to give a yellow solid,
which was triturated with hexane to give 37 (91.8 g, 59%) as
a white solid. 1H NMR (300 MHz, CDCl3) 1.42 (9H, s), 3.73
(2H, s), 7.60 (1H, s). LRMS m/z (thermospray) 357 (MNH4+).
Anal. (C12H12Cl3NO4) C,H,N.
ACEA-1021,26 L-687,414,40 methyl 1,2,3-triazole-4-carboxy-
late,20 and 2-amino-5,6-dihydro-4H-1,3-thiazine25 were pre-
pared by methods in the literature.
A mixture of compound 37 (123 g, 0.361 mol) and saturated
aqueous ammonia (300 mL) in 2-methoxyethanol (360 mL) was
heated in an autoclave at 150 °C for 72 h. The resulting
viscous, black mixture was diluted with water (1 L) and ethyl
acetate (1 L) and filtered through Arbocel filter aid. The dark
red filtrate was separated, and the aqueous layer was ex-
tracted with ethyl acetate (2 × 1 L). The combined organic
solutions were washed with brine (1 L), dried (MgSO4), and
evaporated onto silica gel (70-200 µ, 200 g). The silica gel was
applied to the top of a chromatography column containing
silica gel (40-60 µ, 800 g). Elution with hexane/ethyl acetate
(98:2-92:8) gave 38 as a bright orange solid (39.7 g), which
was contaminated with 5-amino-3,6-dichloro-2-nitrotoluene
(14%). This mixture was carried onto the next step without
Meth od A. Hyd r olysis of 2,3-Dim eth oxyqu in oxa lin es
18-34 To Give 1-17. The 2,3-dimethoxyquinoxaline was
added to a mixture of 2 M hydrochloric acid and dioxane (1:1,
40 mL/g) and heated at reflux until TLC analysis indicated
complete consumption of starting material (2-24 h). The
solvent was removed under reduced pressure, and the residue
was suspended in water (or ether or acetone for the compounds
forming hydrochloride salts) and filtered off. The solid was
dried in vacuo.
Meth od B. P r ep a r a tion of 2,3-Dim eth oxyqu in oxa lin es
18-21 a n d 23-25. (Compound 22 was prepared by Method
C, below.) A mixture of 35 (1 equiv), heterocycle (2 equiv), and
anhydrous potassium carbonate (2 equiv) in dry N,N-dimethy-
lacetamide (4 mL/equiv) or THF (for compounds 21, 24, and
25) was heated at 50 °C with stirring for 18 h. The mixture
was diluted with ethyl acetate, washed with water, and
concentrated under reduced pressure. The residue was purified
by flash chromatography to afford 2,3-dimethoxyquinoxalines
18-25. The reaction of 1,2,4-triazole with 35 yielded a mixture
of 24 and 25 which were separated by flash chromatography
(gradient elution with ethyl acetate/hexane). The structures
of 24 and 25 were confirmed by Rotating Frame Overhauser
Enhancement Spectroscopy, which confirmed the proximity of
the triazole methyl group to the methylene in 25.
1
further purification. H NMR (300 MHz, CDCl3) 2.48 (3H, s),
4.80 (2H, s), 6.82 (1H, s).
A solution of sodium dithionite (94 g, 0.54 mol) in water (1
L) was added to a stirred mixture of 38 (39.7 g, 0.18 mol) and
potassium bicarbonate (94 g, 0.94 mmol) in methanol (1 L) at
room temperature. After 30 min, the mixture was concentrated
under reduced pressure, and the resulting suspension was
extracted with ethyl acetate (total of 700 mL). The extracts
were dried (MgSO4) and concentrated under reduced pressure
to give 2,3-diamino-5,6-dichlorotoluene (26.1 g, 38% over two
1
steps) as a brown solid. H NMR (300 MHz, CDCl3) 2.28 (3H,
P h ysica l a n d Sp ectr oscop ic Da ta for 2,3-Dim eth ox-
yqu in oxa lin es 18-25. 18: colorless solid, yield 68%; 1H NMR
(300 MHz, CDCl3) 4.15 (3H, s), 4.22 (3H, s), 5.80 (2H, s), 6.97
(1H, s), 7.09 (1H, s), 7.74 (1H, s), 7.95 (1H, s). LRMS m/z
(thermospray) 338 (M+).
s), 3.36 (2H, br s), 3.42 (2H, br s), 6.72 (1H, s). LRMS m/z
(thermospray) 191 (M+).
A mixture of 2,3-diamino-5,6-dichlorotoluene (21.6 g, 0.137
mol) and oxalic acid (18.45 g, 0.206 mol) in hydrochloric acid
(4 M, 900 mL) was heated at reflux for 6 h, cooled, and filtered.
The dark brown solid was suspended in diethyl ether, filtered,
and washed with more ether to give 39 (22.06 g, 66%). 1H NMR
(300 MHz, DMSO-d6) 2.40 (3H, s), 7.14 (1H, s), 11.37 (1H, s),
11.94 (1H, s). LRMS m/z (thermospray) 262 (MNH4+).
1
19: colorless solid, yield 69%; H NMR (300 MHz, CDCl3)
4.16 (3H, s), 4.17 (3H, s), 6.03 (2H, s), 6.20 (1H, m), 7.41 (1H,
d, J 2 Hz)), 7.49 (1H, d, J 2 Hz), 7.94 (1H, s). LRMS m/z
(thermospray) 338 (M+).
1
20: colorless solid, yield 58%; H NMR (300 MHz, CDCl3)
A mixture of 39 (22.06 g, 90 mmol), thionyl chloride (300
mL), and dimethylformamide (1 mL) was heated at reflux for
3 h, cooled, and poured slowly into iced water. The resulting
dark yellow precipitate was filtered off to give 5-methyl-2,3,6,7-
tetrachloroquinoxaline (24.42 g, 96%). 1H NMR (300 MHz,
CDCl3) 2.85 (3H, s), 8.02 (1H, s). LRMS m/z (thermospray) 280
(MH+). Anal. (C9H4Cl4N2) C,H,N.
4.15 (6H, s), 6.05 (2H, s), 7.85 (1H, s), 7.95 (1H, s), 8.10 (1H,
s). LRMS m/z (thermospray) 340 (MH+).
1
21: colorless solid, yield 36%; H NMR (300 MHz, CDCl3)
4.15 (6H, s), 6.28 (2H, s), 7.50 (1H, s), 7.63 (1H, s), 7.97 (1H,
s). LRMS m/z (thermospray) 339 (M+).
1
22: colorless solid, yield 37%; H NMR (300 MHz, CDCl3)
4.16 (3H, s), 4.22 (3H, s), 5.85 (2H, s), 7.97 (1H, s), 8.32 (2H,
A solution of sodium methoxide (38 mL, 25% solution in
methanol, 175 mmol) was added over 10 min to a solution of
5-methyl-2,3,6,7-tetrachloroquinoxaline (21 g, 74 mmol) in dry
tetrahydrofuran (200 mL) at 20 °C. There was a mildly
exothermic reaction followed by formation of a precipitate.
After 1 h the mixture was diluted with ethyl acetate (3 L),
washed with water (1 L), dried (MgSO4), and concentrated
under reduced pressure to give 40 (20.3 g, 100%). 1H NMR
(300 MHz, CDCl3) 2.75 (3H, s), 4.15 (3H, s), 4.18 (3H, s), 7.78
s). LRMS m/z (thermospray) 340 (MH+).
23: colorless solid, yield 68%, mp 180 °C, 1H NMR (300
MHz, CDCl3) 2.42 (3H, s), 4.10 (3H, s), 4.16 (3H, s), 5.65 (2H,
s), 6.55 (1H, s), 6.80 (1H, s), 7.96 (1H, s). LRMS m/z (thermo-
spray) 353 (MH+). Anal. (C15H14Cl2N4O2) C,H,N.
1
24: colorless solid, yield 33%; H NMR (300 MHz, CDCl3)
2.35 (3H, s), 4.15 (6H, s), 5.95 (2H, s), 7.90 (1H, s), 7.97 (1H,
s). LRMS m/z (thermospray) 354 (MH+).

1,4-Dihydro-(1H,4H)-quinoxaline-2,3-diones
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 12 1959
(1H, s). LRMS m/z (thermospray) 273 (MH+). Anal. (C11H10
Cl2N2O2) C,H,N.
-
yqu in oxa lin e (43). Zinc dust (2.234 g, 34.1 mmol) was
suspended in anhydrous THF (50 mL) under nitrogen. 1,2-
Dibromoethane (273 µL) was added, and the mixture was
heated to reflux for 1 min. After being cooled, chlorotrimeth-
ylsilane (341 µL) was added. A solution of 35 (6.00 g, 17.1
mmol) in anhydrous THF (100 mL) was added with sonication
at room temperature over 45 min. Sonication was continued
for 1.5 h. In a separate flask, tris(benzylideneacetone)dipal-
ladium (0) (117 mg, 0.25 mmol) was added to a solution of tri-
(2-furyl)phosphine (239 mg, 2.0 mmol) in anhydrous THF (3
mL) under nitrogen with stirring. The solution of the organo-
zinc reagent was added by syringe to the palladium catalyst
solution, and finally chloracetyl chloride (2.06 mL, 25.6 mmol)
was added. The mixture was stirred for 45 min at room
temperature, poured into saturated aqueous ammonium chlo-
ride (300 mL), and extracted with dichloromethane (3 × 330
mL). The extracts were washed with brine, dried (MgSO4), and
concentrated under reduced pressure. The residue was purified
by flash chromatography (eluting with hexane:ethyl acetate
A mixture of 40 (22.0 g, 80.5 mmol), N-bromosuccinimide
(17.2 g, 96.6 mmol), and R,R-azoisobutyronitrile (1.3 g, 8.0
mmol) was heated at reflux in 1,1,1-trichloroethane (400 mL)
for 4 h under irradiation from a 500 W sunlamp. The mixture
was cooled, silica gel (50 g, 60-230 µ) was added, and the
solvent was removed under reduced pressure. The residue was
applied to the top of a silica gel chromatography column, and
the product was eluted using a hexane/ethyl acetate gradient.
The product was triturated with hexane to give 35 (25.3 g,
1
87%) as a fluffy white solid. H NMR (300 MHz, CDCl3) 4.15
(3H, s), 4.22 (3H, s), 5.20 (2H, s), 7.89 (1H, s). Anal. (C11H9-
BrCl2N2O2) C,H,N.
P r ep a r a tion of Com p ou n d s 26 a n d 27. A mixture of 35
(2.11 g, 6.00 mmol), methyl 1,2,3-triazole-4-carboxylate (1.07
g, 8.4 mmol), and anhydrous potassium carbonate (1.66 g, 12
mmol) in dry tetrahydrofuran (25 mL) was heated at reflux
for 18 h, cooled, and poured into water, and the solid was
filtered off. Purification of the solid by flash chromatography
(gradient elution with hexane/dichloromethane) gave 41 (1.278
g, 54%), Rf ) 0.3 (silica gel, hexane:ethyl acetate ) 1:1), as
the slowest eluted of the three product isomers. 1H NMR (300
MHz, CDCl3) 3.90 (3H, s), 4.11 (3H, s), 4.15 (3H, s), 6.29 (2H,
s), 7.93 (1H, s), 8.00 (1H, s). m/z (thermospray) 398 (MH+).
The structure was verified using 2D ROESY NMR, which
confirmed that the triazole proton was adjacent to the meth-
ylene to which the triazole is attached.
A solution of diisobutylaluminum hydride (1 M in dichlo-
romethane, 9 mL, 9 mmol) was added dropwise to a stirred
suspension of 41 (1.087 g, 2.73 mmol) under nitrogen at -78
°C. After 2.25 h, methanol (5 mL) was added, followed by
saturated aqueous ammonium chloride (25 mL) 5 min later.
The mixture was allowed to warm to room temperature and
filtered through Arbocel filter aid, washing the filter cake with
dichloromethane. The organic solution was dried (MgSO4) and
concentrated under reduced pressure to give 42 (981 mg, 98%),
as a white solid. 1H NMR (300 MHz, CDCl3) 4.13 (3H, s), 4.15
(3H, s), 6.30 (2H, s), 7.99 (1H, s), 8.0 (1H, s), 10.08 (1H, s).
m/z (thermospray) 370 (MH+).
1
) 9:1) to give an off-white solid (4.215 g, 75%). H NMR (300
MHz, CDCl3) 4.10 (3H, s), 4.15 (3H, s), 4.24 (2H, s), 4.63 (2H,
s), 7.91 (1H, s). m/z (thermospray) 349 (MH+). νmax. (KBr disk)
2990, 2950, 1730, 1607, 1510, 1407, 1332, 1246, 1228, 988, 888
cm-1
.
6,7-Dich lor o-[(6,7-dih ydr o-5H-im idazo[2,1-b][1,3]th iazin -
2-yl)m eth yl]-2,3-d im eth oxyqu in oxa lin e (44). Anhydrous
potassium carbonate (373 mg, 2.7 mmol) was added to a stirred
solution of 2-amino-5,6-dihydro-4H-1,3-thiazine hydrobromide
(496 mg, 2.7 mmol) at room temperature. After 45 min, 43
(524 mg, 1.5 mmol) and tetra-n-butylammonium bromide (484
mg, 1.5 mmol) were added, and the resulting mixture was
heated at 80 °C for 2 h. The mixture was cooled, diluted with
ethyl acetate, and washed with saturated aqueous sodium
bicarbonate and water. The organic solution was dried (Mg-
SO4) and concentrated under reduced pressure. The crude
product was purified by flash chromatography (gradient elu-
tion with dichloromethane/ethyl acetate) to give an off-white
solid (353 mg, 57%). 1H NMR (300 MHz, CDCl3) 2.25 (2H, m),
3.10 (2H, m), 3.85 (2H, m) 4.11 (3H, s), 4.15 (3H, s), 4.60 (2H,
s), 6.16 (1H, s), 7.83 (1H, s). m/z (thermospray) 411 (MH+).
A mixture of 42 (200 mg, 0.54 mmol), diethylamine (68 µL,
0.65 mmol), and glacial acetic acid (31 µL, 0.54 mmol) in
anhydrous THF (3 mL) was treated with sodium triacetoxy-
borohydride (172 mg, 0.81 mmol) at room temperature with
stirring. After 2 h, the solution was diluted with dichlo-
romethane, washed with saturated aqueous sodium bicarbon-
ate, dried (MgSO4), and concentrated under reduced pressure.
The residue was purified by flash chromatography (gradient
elution with ethyl acetate/methanol) to give 26 (175 mg, 74%),
6,7-Dich lor o-2,3-d im eth oxy-[5-(2-p yr id yl)m eth yl]qu i-
n oxa lin e (29). Zinc dust (131 mg, 2.0 mmol) was suspended
in anhydrous THF (3 mL) under nitrogen. 1,2-Dibromoethane
(16 µL) was added, and the mixture was heated to reflux for
1 min. After being cooled, chlorotrimethylsilane (40 µL) was
added. A solution of 35 (352 mg, 1.0 mmol) in anhydrous THF
(10 mL) was added with sonication at room temperature over
20 min. The mixture was stirred for 15 min and then allowed
to stand. In a separate flask, tris(benzylideneacetone)dipal-
ladium(0) (7 mg, 0.0076 mmol) was added to a solution of tri-
(2-furyl)phosphine (14 mg, 0.061 mmol) in anhydrous THF (1
mL) under nitrogen with stirring. The solution of the organo-
zinc reagent was added by syringe to the palladium catalyst
solution, and finally 2-bromopyridine (114 µL, 1.2 mmol) was
added. The mixture was stirred at reflux for 60 min, cooled,
poured into saturated aqueous sodium bicarbonate (30 mL),
and extracted with dichloromethane (3 × 25 mL). The extracts
were washed with brine, dried (MgSO4), and concentrated
under reduced pressure to give an off-white solid (205 mg,
1
as a colorless solid. H NMR (300 MHz, CDCl3) 1.02 (6H, t, J
7 Hz), 2.49 (4H, q, J 7 Hz), 3.69 (2H, s), 4.15 (6H, s), 6.23 (2H,
s), 7.96 (1H, s). m/z (thermospray) 425 (MH+).
Similarly, reaction of 42 with N-methylpiperazine gave 27
(124 mg, 55%), as an off-white solid. 1H NMR (300 MHz,
CDCl3) 2.35 (3H, s), 2.60 (8H, br s), 3.65 (2H, s), 4.15 (6H, s),
6.23 (2H, s), 7.44 (1H, br s), 7.97 (1H, s). m/z (thermospray)
452 (MH+).
6,7-Dich lor o-[5-(1-p r op ylim id a zol-4-yl)m e t h yl]-2,3-
d im eth oxyqu in oxa lin e (28). A 50% aqueous suspension of
Raney nickel (0.6 mL) was added to a solution of 44 (350 mg,
0.85 mmol) in ethanol (6 mL), and the mixture was heated
under reflux for 16 h. Further amounts of Raney nickel
suspension were added as follows (mL/total reaction time): 0.3/
16 h, 0.6/30 h. After 36 h total reaction time, starting material
had been consumed, so the mixture was cooled and filtered
through Arbocel filter aid, washing with dichloromethane. The
filtrate was dried (MgSO4) and concentrated under reduced
pressure. The residue was purified by flash chromatography
(gradient elution with ethyl acetate/dichloromethane) to give
1
59%), mp 126-8 °C. H NMR (300 MHz, CDCl3) 4.00 (3H, s),
4.13 (3H, s), 6.88 (1H, d J 8 Hz), 7.06 (1H, dd, J 5 and 7 Hz),
7.47 (1H, dt, J 2 and 8 Hz), 7.88 (1H, s), 8.51 (1H, appar. d J
5 Hz).
6,7-Dich lor o-2,3-d im eth oxy-[5-(2-m eth yltetr a zol-5-yl)-
m eth yl]qu in oxa lin e (30). Sodium cyanide (2.06 g, 42 mmol)
was added to a suspension of 35 (10.0 g, 28 mmol) in
2-propanol/water (4:1, 300 mL), and the mixture was heated
at reflux for 1h. After being cooled, the mixture was diluted
with water (400 mL) and extracted with ethyl acetate (3 ×
400 mL). The combined extracts were dried (MgSO4) and
concentrated under reduced pressure. The crude product was
triturated with ether (100 mL), filtered, and dried to give 45
1
a white solid (228 mg, 70%). H NMR (300 MHz, CDCl3) 0.87
(3H, t, J 7 Hz), 2.72 (2H, sextet, J 7 Hz), 3.74 (2H, t, J 7 Hz)
4.15 (6H, s), 4.68 (2H, s), 6.40 (1H, s), 7.33 (1H, s), 7.82 (1H,
s). m/z (thermospray) 381 (MH+).
1
[5-(3-Ch lor o-2-oxopr op-1-yl)]-6,7-dich lor o-2,3-dim eth ox-
(6.89 g, 81%) as a colorless solid. H NMR (300 MHz, CDCl3)

1960 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 12
4.15 (3H, s), 4.20 (3H, s), 4.38 (2H, s), 7.92 (1H, s). m/z
Fray et al.
1
to give 5-nitro-2,3,6,7-tetrachloroquinoxaline (78 g, 73%). H
(thermospray) 298 (MH+), νmax. (KBr) 2270 cm-1
.
NMR (300 MHz, CDCl3) 8.6 (1H, s).
Tin(II) chloride dihydrate (346.3 g, 1.54 mol) was added to
a solution of 5-nitro-2,3,6,7-tetrachloroquinoxaline (96.2 g, 0.31
mol) in ethyl acetate (1.8 L). The mixture was heated under
reflux for 4 h, cooled, and poured cautiously into an excess of
aqueous saturated sodium bicarbonate. The mixture was
filtered through Celite, washing well with ethyl acetate. The
filter cake was macerated with more ethyl acetate, and the
solid material was filtered off. The combined ethyl acetate
solutions were dried (MgSO4) and concentrated under reduced
pressure to give 5-amino-2,3,6,7-tetrachloroquinoxaline (73.4
g, 84%) as a yellow solid. 1H NMR (300 MHz, CDCl3) 5.45 (2H,
br s), 7.47 (1H, s). m/z (thermospray) 385 (MH+).
A solution of sodium methoxide (25% solution in methanol,
274 mL, 1.28 mol) was added to a suspension of 5-amino-
2,3,6,7-tetrachloroquinoxaline (72.4 g, 0.256 mol) in dry metha-
nol (1 L), and the resulting mixture was heated at reflux for
30 min. The mixture was cooled and concentrated under
reduced pressure, and the residue partitioned between water
and ethyl acetate (total of 8 L). The organic solution was dried
(MgSO4) and concentrated under reduced pressure. The crude
product was purified by trituration with methanol, followed
by dissolution in dichloromethane (2 L) and filtration. The
filtrate was concentrated under reduced pressure to give 48
as a yellow solid (55.0 g, 79%). 1H NMR (300 MHz, CDCl3)
4.13 (3H, s), 4.14 (3H, s), 5.07 (2H, br s), 7.26 (1H, s). m/z
(thermospray) 274 (MH+).
Compound 48 (38.12 g, 0.14 mol) was dissolved in acetone
(2 L) and cooled to 0 °C. While being agitated using a
mechanical stirrer, the solution was treated first with 2 M
hydrochloric acid (396 mL, 0.79 mol) and then dropwise with
1 M aqueous sodium nitrite (208 mL, 0.28 mol) while main-
taining the temperature of the mixture at 0 °C. After the
additions were complete, the mixture was stirred for 15 min
and then treated with 5 M aqueous potassium iodide (278 mL,
1.39 mol) maintaining the temperature below 5 °C. The
mixture was then allowed to warm to 10 °C over 30 min. The
acetone was removed under reduced pressure, and the residue
was partitioned between water and ethyl acetate (total of 4
L). The organic solution was washed with 10% aqueous sodium
bisulfite and saturated aqueous sodium bicarbonate, dried
(MgSO4), and concentrated under reduced pressure. Purifica-
tion by flash chromatography (eluting with toluene) gave 49
(16.9 g, 32%). 1H NMR (300 MHz, CDCl3) 4.17 (3H, s), 4.24
(3H, s), 7.91 (1H, s).
A mixture of compound 49 (3.0 g, 7.8 mmol), tributylvinyl-
stannane (4.94 g, 15.6 mmol), lithium chloride (991 mg, 23.4
mmol), and bis(triphenylphosphine)palladium(II) chloride (600
mg, 1.56 mmol) in dry dimethylformamide (100 mL) was
heated at 100 °C for 1.5 h. The mixture was cooled and filtered,
and the filtrate was concentrated under reduced pressure. The
residue was purified by flash chromatography (elution with
toluene) to give 6,7-dichloro-2,3-dimethoxy-5-ethenylquinoxa-
line (1.76 g, 79%), as a white solid. 1H NMR (300 MHz, CDCl3)
4.11 (3H, s), 4.14 (3H, s), 5.84 (1H, d, J 12 Hz), 6.33 (1H, d, J
18 Hz), 7.18 (1H, dd, J 12 and 18 Hz), 7.77 (1H, s). m/z
(thermospray) 285 (MH+).
A solution of azidotributyltin (800 mg, 2.41 mmol) in toluene
(5 mL) was added to a suspension of 45 (600 mg, 2.0 mmol) in
toluene (20 mL), and the mixture was heated at reflux under
nitrogen for 17 h. More azidotributyltin (400 mg, 1.2 mmol)
was added, and heating was continued for a further 25 h. The
mixture was cooled, concentrated under reduced pressure, and
partitioned between water and 5% methanol in dichlo-
romethane. The combined extracts were dried (MgSO4) and
concentrated under reduced pressure. Purification by repeated
flash chromatography gave two fractions: the desired tetrazole
46 (132 mg) and its tributyltin derivative. The latter was
dissolved in dichloromethane, washed with 1 M hydrochloric
acid, dried (MgSO4), and concentrated under reduced pressure.
The residues were redissolved in 10% methanol/dichlo-
romethane and washed with 5% aqueous potassium fluoride.
The organic solution was dried (MgSO4) and concentrated
under reduced pressure. Finally, trituration with acetonitrile
gave additional 46 (89 mg). The total yield was 221 mg (32%).
1H NMR (300 MHz, DMSO-d6) 3.93 (3H, s), 4.05 (3H, s), 4.9
(2H, s), 8.06 (1H, s), m/z (thermospray) 341 (MH+).
A mixture of 46 (175 mg, 0.5 mmol), iodomethane (34 µL,
79 mg, 0.56 mmol), and anhydrous potassium carbonate (77
mg, 0.56 mmol) in anhydrous DMF (5 mL) was stirred at room
temperature for 2 h. The mixture was concentrated under
reduced pressure, diluted with water (5 mL), and extracted
with ethyl acetate (10 mL). The organic phase was dried
(MgSO4) and concentrated under reduced pressure, and the
residue was purified by flash chromatography (elution with
dichloromethane then dichloromethane:methanol ) 99:1) to
give 30 (76 mg 41%) as the faster-running of the two tetrazole
isomers. 1H NMR (300 MHz, DMSO-d6) 4.07 (3H, s), 4.16 (3H,
s), 4.22 (3H, s), 4.98 (2H, s), 7.87 (1H, s), m/z (thermospray)
355 (MH+).
Meth od C: P r ep a r a tion of Com p ou n d s 22 a n d 31-34.
A mixture of 35 or 51 (1 equiv) and the heterocycle (ca. 20
equiv) was heated without solvent for 30 min at 100 °C (for
reaction with imidazole or 1,2,3-triazole) or 140 °C (for reaction
with 1,2,4-triazole) and then cooled. The mixture was parti-
tioned between brine and dichloromethane. The combined
organic extracts were washed with water, dried (MgSO4), and
concentrated under reduced pressure. Crude products were
purified by flash chromatography (gradient elution with
dichloromethane/methanol). The reaction of 1,2,4-triazole with
51 afforded a mixture of 33 and 34 which was separated by
flash chromatography (gradient elution with dichloromethane/
methanol).
P h ysica l a n d Sp ectr oscop ic Da ta for 2,3-Dim eth ox-
1
yqu in oxa lin es 31-34. 31: colorless oil, yield 70%, H NMR
(300 MHz, CDCl3) 0.92 (3H, t), 2.76 (1H, m), 2.91 (1H, m), 4.08
(3H, s), 4.12 (3H, s), 6.34 (1H, m), 6.98 (1H, s), 7.01 (1H, s),
7.78 (1H, s), 7.93 (1H, s). LRMS m/z (thermospray) 367 (MH+).
32: colorless oil, yield 32%, 1H NMR (300 MHz, CDCl3) 0.99
(3H, t), 2.98 (2H, m), 3.98 (3H, s), 4.14 (3H, s), 6.79 (1H, m),
7.60 (1H, s), 7.63 (1H, s), 7.97 (1H, s). LRMS m/z (thermospray)
368 (MH+).
1
33: colorless solid, yield 19%, H NMR (300 MHz, CDCl3)
A mixture of ozone and oxygen was bubbled gently through
a stirred solution of 6,7-dichloro-2,3-dimethoxy-5-ethenylqui-
noxaline (1.76 g, 6.2 mmol) in chloroform (200 mL) at -60 °C
until a blue color persisted. The solution was purged with
nitrogen, and then triphenylphosphine (3.23 g, 12.3 mmol) was
added. The mixture was allowed to warm to room temperature
and stirred for a further 16 h. The chloroform was removed
under reduced pressure, and the residue was purified by flash
chromatography (eluting with toluene) to give 6,7-dichloro-
2,3-dimethoxy-5-formylquinoxaline (1.60 g, 90%) as a fluffy
white solid. 1H NMR (300 MHz, CDCl3) 4.16 (6H, s), 8.08 (1H,
s), 11.06 (1H, s). m/z (thermospray) 287 (MH+).
1.01 (3H, t), 2.86 (2H, m), 4.02 (3H, s), 4.15 (3H, s), 6.61 (1H,
m), 7.84 (1H, s), 7.97 (1H, s), 8.32 (1H, s). LRMS m/z
(thermospray) 368 (MH+).
1
34: colorless foam, yield 32%, H NMR (300 MHz, CDCl3)
0.98 (3H, t), 2.74 (1H, m), 2.89 (1H, m), 4.05 (3H, s), 4.16 (3H,
s), 6.40 (1H, m), 7.98 (1H, s), 8.31 (2H, s). LRMS m/z
(thermospray) 368 (MH+).
5-(1-Br om op r op -1-yl)-6,7-d ich lor o-2,3-d im et h oxyq u i-
n oxa lin e (51). A mixture of 6,7-dichloro-5-nitroquinoxalin-
2,3-dione (47) (84 g, 0.34 mol), thionyl chloride (840 mL), and
dimethylformamide (0.5 mL) was heated at reflux for 3 h,
cooled, and concentrated under reduced pressure. Ethyl acetate
(300 mL) was added and removed under reduced pressure,
followed by petroleum ether (bp 100-120 °C). The solid residue
was recrystallized from hot petroleum ether (bp 100-120 °C)
Ethylmagnesium bromide (9.08 mL, 3 M in diethyl ether,
2.3 mmol) was added to a suspension of 6,7-dichloro-2,3-
dimethoxy-5-formylquinoxaline (3.91 g, 13.62 mmol) in dry
tetrahydrofuran (200 mL) under nitrogen at room tempera-

1,4-Dihydro-(1H,4H)-quinoxaline-2,3-diones
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 12 1961
(12) Cai, S. X.; Kher, S. M.; Zhou, Z.-L.; Ilyin, V.; Espitia, S. A.; Tran,
M.; Hawkinson, J . E.; Woodward, R. M.; Weber, E.; Keana, J .
F. W. Structure-Activity Relationships of Alkyl- and Alkoxy-
Substituted 1,4-Dihydroquinoxaline-2,3-diones: Potent and Sys-
temically Active Antagonists for the Glycine Site of the NMDA
Receptor. J . Med. Chem. 1997, 40, 730-738, and references
therein.
(13) Fray, M. J .; Bull, D. J .; Carr, C. L.; Stobie, A. Structure Activity
Relationships of 3-Hydroxy-2(1H)-Quinolinones as N-Methyl-D-
Aspartate (Glycine Site) Antagonists. Med. Chem. Res. 1996, 6,
581-592.
(14) We had anticipated that the SARs of 3-hydroxy-2(1H)-quinoli-
nones would parallel the quinoxalinediones, but in fact they were
more akin to the kynurenic acids, see ref 10.
(15) Abenhaim, D.; Diez-Barra, E.; Hoz, A.; Loupy, A.; Sanchez-
Migallon, A. Selective alkylations of 1,2,4-triazole and benzot-
riazole in the absence of solvent. Heterocycles 1994, 38, 793-
802.
ture. After 30 min, saturated ammonium chloride (50 mL) was
added, and the product was extracted into ethyl acetate (2 ×
100 mL). The combined extracts were dried (MgSO4) and
concentrated under reduced pressure. The residues was puri-
fied by flash chromatography (eluting with dichloromethane)
to give 50 (1.86 g, 43%), as a pale yellow solid. H NMR (300
CDCl3) 1.05 (3H, t, J 7 Hz), 1.99 (2H, m), 4.11 (3H, s), 4.12
(3H, s), 5.40 (1H, m), 5.65 (1H, d, J 11 Hz), 7.81 (1H, s).
1
Carbon tetrabromide (466 mg, 1.41 mmol) was added in
portions to a solution 50 (223 mg, 0.70 mmol) and triph-
enylphosphine (369 mg, 1.41 mmol) in dry dichloromethane
(15 mL) at 20 °C. The mixture was stirred for 18 h, and the
solvent was removed under reduced pressure. Purification by
flash chromatography eluting with hexane/dichloromethane
(1:1) gave the title compound (202 mg, 46%), as a white solid,
mp 134-6 °C. 1H NMR (300 MHz, CDCl3) diastereomers
evident: 0.96 (3H, t, J 7 Hz, CH3 both isomers), 2.60 (2H, m,
CH2 minor), 2.80 (2H, m, CH2 major), 4.13 (3H, s, CH3O, both
isomers), 4.18 (3H, s, CH3O, minor), 4.20 (3H, s, CH3O, major),
5.87 (1H, t, J 7 Hz, CH major rotamer), 6.68 (1H, t, J 7 Hz,
CH minor), 7.87 (1H, s, aromatic H both isomers). m/z
(thermospray) 379 (MH+).
(16) Mudryk, B.; Makosza, M. Reactions of organic anions. 156. A
simple and efficient synthesis of alkyl (2-nitroaryl)acetates and
alkyl 2-(2-nitroaryl)propanoates via vicarious nucleophilic sub-
stitution of hydrogen in nitroarenes by carbanions of alkyl
chloroacetates or alkyl 2-chloropropanoates. Synthesis 1988,
1007-9.
(17) Bull, D. J .; Fray, M. J .; MacKenny, M. C.; Malloy, K. A. Facile
decarboxylation of nitrobenzeneacetic acids as a convenient route
to alkylnitrobenzenes. Synlett 1996, 647-648.
(18) European Patent Application E.P. 385,850.
(19) Kazimierczuk, Z.; Pfleiderer, W. Nucleosides. XXXVIII. Synthesis
and properties of 1-(â-D-ribofuranosyl)-2-quinoxalinones and
-2,3-quinoxalinediones. Liebigs Ann. Chem. 1982, 754-61.
(20) Thurber, T. C.; Townsend, L. B.; Ring contractions of 5-diazou-
racils. I. Conversions of 5-diazouracils into 1,2,3-triazoles by
hydrolysis and methanolysis. J . Org. Chem. 1976, 41, 1041-
51.
(21) Berk, S. C.; Knochel, P.; Yeh, M.-C. P. General approach to
highly functionalized benzylic organometallics of zinc and copper.
J . Org. Chem. 1988, 53, 5789-91.
(22) Grey, R. A. A palladium-catalyzed synthesis of ketones from acid
chlorides and organozinc compounds. J . Org. Chem. 1984, 49,
2288-9.
(23) Walker, M. A.; Kaplita, K. P.; Chen, T.; King, D. H. Synthesis
of all three regioisomers of pyridylalanine. Synlett 1997, 169-
170.
Ack n ow led gm en t. The authors wish to thank Drs.
G.T.G. Swayne and S.G. J ezequel for advice and ac-
knowledge the able assistance of Mrs. N.S. Crouch and
Miss S.-A. Thompson for ligand binding assays, Mr. M.
C. MacKenny for preparing some of the compounds, Dr.
J . Urenjak and Miss M. A. Parry for in vitro functional
assays, Mr. G. Fox and Dr. R. A. O’Donnell for in vivo
assays, and Mr. R. Webster for pharmacokinetic mea-
surements. We also thank the staff of the Physical
Sciences Department for analytical data and measuring
spectroscopic data.
(24) Tetrakistri(phenylphosphine)palladium(0) catalysis gave lower
yields and significant amounts of 38 as a side product.
(25) Scho¨berl, A.; Magosch, K.-H. Cyclisation Reactions with 2-Ami-
nopentthiazoline. A Contribution to the Reactivity of R-Amino-
N-Heterocycles. Liebigs Ann. Chem. 1970, 742, 85-97.
(26) Keana, J . F. W.; Kher, S. M.; Cai, S. X.; Dinsmore, C. M.; Glenn,
A. G.; Guastella, J .; Huang, J . C.; Lu, Y.; Ilyin, V.; Lu, Y.;
Mouser, P. L.; Woodward, R. M.; Weber, E. Synthesis and
Structure-Activity Relationships of Substituted 1,4-Dihydro-
quinoxaline-2,3-diones (QXs): Antagonists of the Glycine Site
on the NMDA Receptor and of Non-NMDA Glutamate Receptors.
J . Med. Chem. 1995, 38, 4367-4379.
(27) Clive, D. L. J .; Angoh, A. G.; Bennett, S. M. Radical Spirocycli-
sation: Synthesis of an Appropriately Oxygenated Spiro Com-
pounds Related to the Anti-tumour Antibiotic Fredericamycin
A. J . Org. Chem. 1987, 52, 1339.
(28) McKean, D. R.; Parrinello, G.; Renaldo, A. F.; Stille, J . K.
Synthesis of Functionalised Styrenes via Palladium-Catalysed
Coupling of Aryl Bromides with Vinyl Tin Reagents. J . Org.
Chem. 1987, 52, 422.
(29) Lan, A. J . Y.; Heukeroth, R. O.; Mariano, P. S. Electron-transfer
Induced Photocyclisation Reactions of Arene-Iminium Salt
Systems. Effects of Cation Diradical Deprotonation and Desi-
lylation on the Nature and Efficiencies of Reaction Pathways
Followed. J . Am. Chem. Soc. 1987, 109, 2738-45.
(30) Grimwood, S.; Moseley, A. M.; Carling, R. W.; Leeson, P. D.;
Foster, A. C. Characterisation of the Binding of [3H]-L-689, 560
and Antagonist for the Glycine Site on the N-Methyl-D-aspartate
Receptor, to Rat Brain Membranes. Mol. Pharmacol. 1992, 41,
923-30.
Refer en ces
(1) Tranquillini, M. E.; Reggiani, A.; Glycine Antagonists and
Stroke. Exp. Opin. Invest. Drugs 1999, 8, 1837-48.
(2) Dannhardt, G.; Kohl, B. K. The glycine site on the NMDA
receptor: structure-activity relationships and possible thera-
peutic applications. Curr. Med. Chem. 1998, 5, 253-263.
(3) Parson, C. G.; Danyscz, W.; Hesselink, M.; Hartmann, S.; Lorenz,
B.; Wollenburg, C.; Quack, G.; Modulation of NMDA receptors
by glycine. Introduction to some basic aspects and recent
developments. Amino Acids 1998, 14, 207-216.
(4) Trullas, R.; Neuroprotective Effects of Functional Antagonists
at the Glycine Site of the NMDA Receptor. Methods Find. Exp.
Clin. Pharmacol. 1997, 19 (Suppl. A), 55-56.
(5) The NMDA Receptor, 2nd ed.; Watkins, J . C., Collingridge, G.
L., Eds.; IRL Press: Oxford, England, 1994.
(6) Wood, P. L.; Hawkinson, J . E. N-Methyl-D-Aspartate Antagonists
for Stroke and Head Trauma. Expert Opin. Invest. Drugs 1997,
6, 389-397. Data obtained with glycine site antagonists so far
suggest they do not cause PCP-like psychotomimetic effects or
vacuolisation of cortical neurones which have been observed with
glutamate and ion-channel site antagonists.
(7) Lees, K. R.; Cerestat and other NMDA antagonists in ischaemic
stroke. Neurology 1997, 49 (5, Suppl. 4), S66-S69.
(8) Albers, G. W.; Clark, W. M.; Atkinson, R. P.; Madden, K.; Data,
J . L.; Whitehouse, M. J . Dose escalation study of the NMDA
glycine-site antagonist licostinel in acute ischemic stroke. Stroke
1999, 30, 508-13.
(9) Dyker, A. G.; Lees, K. R. Safety and tolerability of GV150526 (a
glycine site antagonist at the N-methyl-D-aspartate receptor) in
patients with acute stroke. Stroke 1999, 30, 986-92.
(10) Kulagowski, J . J . Glycine-site NMDA Receptor Antagonists: an
update. Expert Opin. Ther. Pat. 1996, 6, 1069-1079.
(11) Di Fabio, R.; Capelli, A. M.; Conti, N.; Cugola, A.; Donati, D.;
Feriani, A.; Gastaldi, P.; Gavirighi, G.; Hewkin, C. T.; Micheli,
F.; Missio, A.; Mugnaini, M.; Pecunioso, A.; Quaglia, A. M.; Ratti,
E.; Rossi, L.; Tedesco, G.; Trist, D. G.; Reggiani, A. Substituted
Indole-2-carboxylates as in Vivo Potent Antagonists Acting at
the Strychnine-Insensitive Glycine Binding Site. J . Med. Chem.
1997, 40, 841-850.
(31) Harrison, N. L.; Simmonds, M. A. Quantitative Studies on some
Antagonists of N-Methyl D-Aspartate in Slices of Rat Cerebral
Cortex. Brit. J . Pharmacol. 1985, 84, 381-91.
(32) Leeson, P. D.; Williams, B. J .; Baker, R.; Ladduwahetty, T.;
Moore, K. W.; Rowley, M. Effects of five-membered ring confor-
mation on bioreceptor recognition: identification of 3R-amino-
1-hydroxy-4R-methylpyrrolidin-2-one (L-687,414) as a potent
glycine/N-methyl-D-aspartate receptor antagonist. J . Chem. Soc.,
Chem. Commun. 1990, 1578-80. Foster, A. C.; Kemp, J . A.;
Tricklebank, M. D.; Leeson, P. D.; Carling, R. W.; Williams, B.
J .; Grimwood, S.; Donald A. E.; Priestley, T.; et al. Selective
antagonists for the glycine site of the N-methyl-D-aspartate
receptor. Mol. Neuropharmacol. 1992, 2, 97-102.

1962 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 12
Fray et al.
(33) Leeson, P. D.; Carling, R. W.; Moore, K. W.; Moseley, A. M.;
Smith, J . D.; Stevenson, G.; Chan, T.; Baker, R.; Foster, A. C.;
Grimwood, S.; Kemp, J . A.; Marshall, G. R.; Hoogsteen, K.;
4-Amido-2-carboxytetrahydroquinolines. Structure-activity re-
lationships for antagonism at the glycine site of the NMDA
receptor. J . Med. Chem. 1992, 35, 1954-68.
(34) Abraham, M. H.; Duce, P. P.; Prior, D. V.; Barratt, D. G.; Morris,
J . J .; Taylor, P. J . Hydrogen Bonding. Part 9. Solute proton donor
and Proton Acceptor Scales for use in Drug Design. J . Chem.
Soc., Perkin Trans. 2, 1989, 1355-75.
(35) Leeson, P. D.; Baker, R.; Carling, R. W.; Curtis, N. R.; Moore,
K. W.; Williams, B. J .; Foster, A. C.; Donald, A. E.; Kemp, J . A.;
Marshall, G. R.; Kynurenic Acid Derivatives. Structure-activity
Relationships for Excitatory Amino Acid Antagonism and Iden-
tification of Potent and Selective Antagonists at the Glycine Site
on the N-Methyl-D-Aspartate Receptor. J . Med. Chem. 1991, 34,
1243.
(36) Auberson, Y. P.; Acklin, P.; Allgeier, H.; Biollaz, M.; Bischoff,
S.; Ofner, S.; Veenstra, S. J . 5-Aminomethylquinoxaline-2,3-
diones. Part II: N-Aryl Derivatives as Novel NMDA/Glycine
and AMPA Antagonists. Bioorg. Med. Chem. Lett. 1998, 8, 71-
74.
(37) Moreau, J . L.; Pieri, L.; Prud’hon, B. Convulsions induced by
centrally administered NMDA in mice: effects of NMDA an-
tagonists, benzodiazepines, minor tranquilizers and anticonvul-
sants. Brit. J . Pharmacol. 1989, 98, 1050-1054.
(38) For a discussion, see: J ezequel, S. G. Central Nervous System
Penetration of Drugs: Importance of Physicochemical Properties.
Prog. Drug. Metab. 1992, 13, 141-178.
(39) Fray, M. J .; Bull, D. J .; Carr, C. L.; Mowbray, C. E.; Stobie, A.
Structure-Activity Relationships of 1,4-Dihydro-(1H, 4H)-qui-
noxaline-2,3-diones as N-Methyl-D-Aspartate (Glycine Site) Re-
ceptor Antagonists. 2. Discovery of UK-240,455, a Water-soluble
N-Substituted 5-Methanesulfonamido Derivative with Potent In
Vivo Activity. Manuscript in preparation.
(40) Fray, M. J .; Bull, D. J .; J ames, K. A short synthesis of the glycine
antagonist (3R,4R)-3-amino-1-hydroxy-4-methylpyrrolidin-2-one
(L-687,414). Synlett. 1992, 709-10.
J M001124P