Synthesis of Willardiine and 6-azawillardiine analogs: Pharmacological characterization on cloned homomeric human AMPA and kainate receptor subtypes

Article abstract of DOI:10.1021/jm9702387

Both willardiine and azawillardiine analogs (18-28) have been reported to be potent and selective agonists for either AMPA or kainate receptors. We report here the novel synthesis and pharmacological characterization of a range of willardiine (18-23) and 6-azawillardiine (24-28) analogs on cells individually expressing human homomeric hGluR1, hGluR2, hGluR4, or hGluR5 receptors. Reaction of the sodium salts of substituted uracils (7-12) or 6- azauracils (13-16) with (S)-3-[(tert-butoxycarbonyl)amino]oxetan-2-one (17) in dry DMF, subsequent deprotection in TFA, and purification by ion-exchange chromatography gave mainly the willardiine analog in which alkylation took place on N1 of the uracil ring. We have investigated the subtype selectivity of these compounds by examining their binding affinity for homomeric hGluR1, -2, -4, or -5 (and hGluR6 in the case of 5-iodowillardiine (22)). From this study we have demonstrated that 22 has high affinity for hGluR5 and, compared to kainate, displays excellent selectivity for this receptor over both the AMPA receptor subtypes and the homomeric kainate receptor, hGluR6. 5- Fluorowillardiine (19) has higher affinity than AMPA for both homomeric hGluR1 and hGluR2 and compared to AMPA displays greater selectivity for AMPA receptor subtypes over the kainate receptor, hGluR5. Some structural features required for optimal activity at homomeric AMPA or kainate receptor subtypes have also been identified. It would appear that quite large lipophilic substituents at the 5-position of the uracil ring not only are accommodated by hGluR5 receptors but also lead to enhanced affinity for these receptors. In contrast to this, for optimal binding affinity to hGluR1, -2, or -4, smaller, electron-withdrawing substituents are required. For optimal activity at hGluR4 receptors a 6-azasubstituted willardiine is favored. The subtype- selective compounds described here are likely to be useful tools to probe the distribution and the physiological roles of the various glutamate receptor subunits in the central nervous system.

Full text of DOI:10.1021/jm9702387

J . Med. Chem. 1997, 40, 3645-3650
3645
Syn th esis of Willa r d iin e a n d 6-Aza w illa r d iin e An a logs: P h a r m a cologica l
Ch a r a cter iza tion on Clon ed Hom om er ic Hu m a n AMP A a n d Ka in a te Recep tor
Su btyp es
David E. J ane,*,‡ Ken Hoo,† Raj Kamboj,† Michele Deverill,† David Bleakman,§ and Allan Mandelzys†
Department of Pharmacology, School of Medical Sciences, University of Bristol, U.K., Lilly Research Centre, U.K., and
Allelix Biopharmaceuticals, Inc., Mississauga, Ontario, Canada
Received April 8, 1997X
Both willardiine and azawillardiine analogs (18-28) have been reported to be potent and
selective agonists for either AMPA or kainate receptors. We report here the novel synthesis
and pharmacological characterization of a range of willardiine (18-23) and 6-azawillardiine
(24-28) analogs on cells individually expressing human homomeric hGluR1, hGluR2, hGluR4,
or hGluR5 receptors. Reaction of the sodium salts of substituted uracils (7-12) or 6-azauracils
(13-16) with (S)-3-[(tert-butoxycarbonyl)amino]oxetan-2-one (17) in dry DMF, subsequent
deprotection in TFA, and purification by ion-exchange chromatography gave mainly the
willardiine analog in which alkylation took place on N1 of the uracil ring. We have investigated
the subtype selectivity of these compounds by examining their binding affinity for homomeric
hGluR1, -2, -4, or -5 (and hGluR6 in the case of 5-iodowillardiine (22)). From this study we
have demonstrated that 22 has high affinity for hGluR5 and, compared to kainate, displays
excellent selectivity for this receptor over both the AMPA receptor subtypes and the homomeric
kainate receptor, hGluR6. 5-Fluorowillardiine (19) has higher affinity than AMPA for both
homomeric hGluR1 and hGluR2 and compared to AMPA displays greater selectivity for AMPA
receptor subtypes over the kainate receptor, hGluR5. Some structural features required for
optimal activity at homomeric AMPA or kainate receptor subtypes have also been identified.
It would appear that quite large lipophilic substituents at the 5-position of the uracil ring not
only are accommodated by hGluR5 receptors but also lead to enhanced affinity for these
receptors. In contrast to this, for optimal binding affinity to hGluR1, -2, or -4, smaller, electron-
withdrawing substituents are required. For optimal activity at hGluR4 receptors a 6-aza-
substituted willardiine is favored. The subtype-selective compounds described here are likely
to be useful tools to probe the distribution and the physiological roles of the various glutamate
receptor subunits in the central nervous system.
Ch a r t 1. Structures of Selective Ionotropic Glutamate
Receptor Agonists
In tr od u ction
It is widely accepted that (S)-glutamic acid (L-Glu, 1)
is the endogenous transmitter at a range of excitatory
amino acid (EAA) receptors in the central nervous
system (CNS). EAA receptors can be subdivided into
two main classes, ionotropic receptors,1 which mediate
fast synaptic transmission through ligand-gated ion
channels, and metabotropic glutamate receptors,2 which
are G-protein-coupled to second-messenger systems and
are thought to be responsible for modulation of synaptic
transmission. At present there are three main classes
of ionotropic glutamate receptors named after the
selective agonists that activate them: the N-methyl-D-
aspartate (NMDA, 2),3 (S)-2-amino-3-(5-methyl-3-hy-
droxyisoxazol-4-yl)propanoic acid (AMPA, 3), and kainic
acid (4) receptors,4,5 (see Chart 1 for agonist structures).
Molecular biology has revealed that these pharmaco-
logically distinct receptor subtypes are encoded by
separate, albeit related, gene families.4,5 Human AMPA
receptors are encoded by the hGluR1-4 genes. These
subunits can form functional homomeric ion channels
that can be activated by both AMPA (3) and kainate
(4). Human kainate receptors, hGluR5 and hGluR6, are
activated by kainate and more potently by domoate (5),
while AMPA has very little activity. Other kainate
receptor assemblages (homomeric hGluR7, KA1, and
KA2) bind kainate with high affinity but do not form
functional ion channels.4,5
* To whom correspondence should be addressed.
† Allelix Biopharmaceuticals.
‡ University of Bristol.
§ Lilly Research Centre.
X Abstract published in Advance ACS Abstracts, October 1, 1997.
S0022-2623(97)00238-0 CCC: $14.00 © 1997 American Chemical Society

3646 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 22
J ane et al.
Sch em e 1
We have reported that analogs of the naturally
occurring substance, (S)-willardiine (18), have potent
and selective actions on AMPA and kainate receptors,
the corresponding R enantiomers being inactive.6-14
Thus, (S)-5-fluorowillardiine (19) is more potent and
selective than AMPA as an AMPA receptor agonist in
both cultured hippocampal neurons8,9 and motoneurons
in the neonatal rat spinal cord.6 Conversely, (S)-5-
iodowillardiine (22) is more potent and selective as a
kainate receptor agonist than kainate in both cultured
dorsal root ganglion (DRG) cells9 and dorsal roots of the
neonatal rat spinal cord.7,13 In the case of the actions
of willardiine analogs on hippocampal neurons and DRG
cells, an extensive structure-activity study was under-
taken.9 It was observed that for potent and selective
activity at AMPA receptors, electron-withdrawing groups
at the 5-position of the uracil ring were required and
that such groups should be as small as possible, because
large groups were less well accommodated by the
receptor. For optimal activity at kainate receptors
larger electron-withdrawing groups with a high degree
of lipophilicity were required. As a result of this study
we decided to synthesize a range of 6-azawillardiine
analogs (24-28) in an attempt to improve the potency
and selectivity at both AMPA and kainate receptors. It
was reasoned that this 6-aza substitution would lower
the pKa of the uracil ring, and therefore more of the
ionized form would be available at physiological pH. In
addition, because this substitution would not result in
an increase in bulk around the uracil ring, it would be
likely that such a substitution would not be detrimental
to activity at either AMPA or kainate receptors. We
have recently reported that these (S)-6-azawillardiine
analogs are more potent agonists at AMPA receptors
in the spinal cord than the corresponding willardiines
as determined electrophysiologically in the neonatal rat
spinal cord preparation14 and by displacement of (S)-
[3H]AMPA in rat brain.10-12 Since both willardiine and
azawillardiine analogs were potent and selective ligands
for either AMPA or kainate receptors, we decided to
investigate whether these compounds had selective
action on homomeric non-NMDA receptors. Such sub-
type-selective compounds would likely be useful tools
to probe the distribution and the physiological roles of
the various glutamate receptor subunits in the CNS. We
report here the novel synthesis and pharmacological
characterization of a range of willardiine (18-23) and
6-azawillardiine (24-28) analogs on cells individually
expressing human homomeric hGluR1, hGluR2, hGluR4,
or hGluR5 receptors (and hGluR6 in the case of 5-io-
dowillardiine (22)). In addition, the structural features
for selective action at such homomeric glutamate recep-
tors are discussed.
the synthesis of a range of 5-halo-substituted 1-(formyl-
methyl)pyrimidine-2,4-diones using the procedures pre-
viously described for the synthesis of the unsubstituted
aldehyde16,17 but obtained only poor yields of impure
aldehydes. Arnold and co-workers have reported that
(S)-3-[(tert-butoxycarbonyl)amino]oxetan-2-one (17)
(readily prepared from (S)-serine) can act as a chiral
electrophilic alanine cation equivalent, which reacts
with nucleophiles to give optically pure â-substituted
alanines.18-21 We have applied this reaction to the
synthesis of a range of willardiine analogs.
Reaction of the sodium salts of substituted uracils (7-
12) or 6-azauracils (13-16) with (S)-3-[(tert-butoxycar-
bonyl)amino]oxetan-2-one (17) in dry DMF (Scheme 1)
gave mainly the willardiine analog in which alkylation
took place on N1 of the uracil ring; none of the product
alkylated on N3 was observed (due to N1 being more
acidic than N3 (i.e., pK < pK )22). However, in some
cases small amounts of the N1,N3-dialkylated com-
pound were formed, but this could be minimized by
using excess of the uracil and by adding the (S)-3-[(tert-
butoxycarbonyl)amino]oxetan-2-one (17) portionwise.
Removal of the N-Boc protecting group by treatment
with trifluoroacetic acid gave the crude willardiines,
which were purified by ion-exchange chromatography.
Contamination by the starting uracils was removed
by applying a suspension of the crude product in water
to a bed of Dowex 50WX8-400 H+ resin. Elution of this
column with water/ethanol (50:50) completely removed
the contaminating uracil. Willardiine analogs were
then isolated by elution of the column with 10% aqueous
pyridine solution. Crystallization of the crude amino
acid from water gave the corresponding willardiine
analog free from contamination by either serine or the
N1,N3-dialkylated compound.
aN1
aN3
Ch em istr y
Racemic willardiine (6) has been previously synthe-
sized in a 62% yield utilizing the Strecker amino acid
synthesis with 1-(formylmethyl)pyrimidine-2,4-dione as
the substrate.15 Yields in the first two stages involving
the synthesis of 1-(formylmethyl)pyrimidine-2,4-dione
have been improved.16-17 However, this approach to the
synthesis of willardiines has a number of disadvan-
tages: it involves a multistep synthesis, cannot be
applied to a range of substituted analogs, and leads to
racemic mixtures, which require resolution in order to
obtain the individual enantiomers. We have attempted
This method therefore is the method of choice for
preparing a range of willardiine analogs with a variety
of substituents at the 5-position of the uracil ring. The
5-bromo-substituted compound (28) was obtained by
treatment of 24 with a solution of bromine in aqueous

Synthesis of Willardiine and 6-Azawillardiine Analogs
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 22 3647
Ta ble 1. Data for Displacement of Either [3H]AMPA or
[3H]Kainate Binding from Homomeric hGluRl, hGluR2,
hGluR4, or hGluR5 Receptors by Willardiine Analogs (18-28)
cal study on neonatal rat motoneurons.14 Furthermore,
on the cloned receptors, the 6-aza analogs (24-27) are
more potent than the corresponding non-aza analogs
(18, 20-22), with the 5-chloro-6-aza analog (25) being
9 times more potent on hGluR1 and 3 times more potent
on hGluR2 than 5-chlorowillardiine (20).
Ki (nM)a
compd
hGluR1
hGluR2
hGluR4
hGluR5
1
3
4
1360 ( 257
103 ( 13
940 ( 93
107 ( 16
868 ( 219
155 ( 10
701 ( 46
2144 ( 416
177 ( 22
Hansch analysis of the electrophysiological data for
19-22 obtained in hippocampal cells revealed a strong
relationship between size and electron-withdrawing
ability of the 5-substituent and agonist potency for
AMPA receptors, with small, electron-withdrawing sub-
stituents being required for optimal activity.9 As previ-
ously reported,12,14 6-aza substitution results in an
increase in agonist potency at AMPA receptors, which
may be explained by the resultant lowering of the pKa
of the uracil ring upon 6-aza substitution.23 The binding
profile for 18-28 on homomeric hGluR4 receptors is
markedly different from that observed for homomeric
hGluR1 or hGluR2 receptors. There is an approxi-
mately 2500-fold range in binding affinities for this
series of compounds, suggesting that the homomeric
hGluR4 receptor binding site is more sensitive to
changes in substitution pattern on the uracil ring than
either homomeric hGluR1 or hGluR2 receptors. The
rank order of potency on homomeric hGluR4 receptors
is 5-Cl-6-azawil g 5-Br-6-azawil g 5-I-6-azawil > 5-Me-
6-azawil > 6-azawil g 5-NO2-wil g 5-F-wil g 5-Cl-wil
) 5-Br-wil > 5-I-wil > wil. For optimal binding affinity
to hGluR4 small electron-withdrawing substituents are
required, with 5-chloro-6-azawillardiine (25) being 25
times more potent than the 5-methyl-6-aza analog (27)
even though the 5-substituent is of similar size. For
the 5-halo-substituted analogs (20-22, 25, 26, 28) the
rank order of potency is Cl > Br > I, in parallel to that
found for hGluR1 and -2. The 6-aza analogs (24-28)
have higher affinity for homomeric hGluR4 receptors
than the corresponding non-aza analogs (18-23), with
the 5-chloro-6-aza analog (25) being the most potent (Ki
) 3.6 ( 1.4 nM).
In the case of 5-iodowillardiine (22), 6-aza substitution
leads to a 50-fold increase in affinity for hGluR4. The
improvement in potency on hGluR4 receptors upon
6-aza substitution is not likely to be due solely to a
decrease in the pKa of the uracil ring, as the 5-methyl-
6-aza analog (27) is more potent than 5-nitrowillardiine
(23) even though it is likely that 23 has a lower pKa.
Alternative explanations may be that the lone pair of
the nitrogen atom at the 6-position of the uracil ring
may be involved in extra hydrogen bonding with the
receptor, or perhaps a π-π interaction between the
uracil ring and an aromatic group on the receptor is
involved and this interaction is enhanced upon 6-aza
substitution. There is an approximately 120000-fold
range in affinity for the willardiine analogs (18-28) on
homomeric hGluR5 receptors, demonstrating the re-
markable sensitivity of these receptors to small changes
in chemical structure. For example, 5-iodowillardiine
(22) is approximately 7600 times more potent on hGluR5
than 5-fluorowillardiine (19). The rank order of potency
of willardiine analogs (18-28) on homomeric hGluR5
receptors is 5-I-wil > 5-Br-6-azawil > 5-Br-wil > 5-I-6-
azawil > 5-Cl-6-azawil g 5-Cl-wil > 5-NO2-wil > 5-Me-
6-azawil > 5-F-wil g 6-azawil . wil. For the 5-halo-
substituted compounds (19-22) the rank order of
potency on hGluR5 receptors (5-I > 5-Br > 5-Cl > 5-F)
7450 ( 2020 12200 ( 2740 1710 ( 170
18
19
20
21
22
23
24
25
26
27
28
386 ( 92
14.7 ( 1.3
65 ( 11
898 ( 86
25.1 ( 5.2
53.1 ( 4.4
101 ( 19
176 ( 29
115 ( 22
461 ( 57
16.7 ( 6.6
101 ( 37
372 ( 61
19.3 ( 8.3
8850 ( 923 28900 ( 7350
305 ( 107
451 ( 65
457 ( 49
972 ( 155
207 ( 59
189 ( 22
3.6 ( 1.4
19.9 ( 8.2
89 ( 8
1820 ( 149
57.1 ( 23.5
9.1 ( 2.4
0.24 ( 0.06
393 ( 217
1980 ( 456
30.9 ( 5.4
19.8 ( 6.0
1150 ( 172
2.9 ( 0.3
92 ( 25
163 ( 42
279 ( 102
163 ( 14
7.1 ( 2.2
73.9 ( 4.8
240 ( 39
29.0 ( 6.8
7.9 ( 3.8
a
Versus [3H]AMPA (hGluRl, -2, or -4) or [3H]kainate (hGluR5)
binding at homomeric non-NMDA receptor subtypes. Values
represent a minimum of three determinations ( SEM.
acetic acid (Scheme 1) followed by purification by ion-
exchange resin chromatography using Dowex 50WX8-
400 resin.
Discu ssion
The pharmacological characterization of willardiine
analogs (19-28) as potent agonists at either AMPA or
kainate receptors has been previously reported;6-14
however, these compounds have not been characterized
on cloned homomeric AMPA or kainate receptor sub-
types. We have investigated the subtype selectivity of
these compounds by examining their binding affinity for
homomeric hGluR1, -2, -4, or -5. Examination of the
binding data in Table 1 reveals that in general substitu-
tion at the 5-position of the uracil ring of willardiine
with either halogen or nitro substituents or 6-aza
substitution increases affinity for all the homomeric
receptors (hGluR1, -2, -4, or -5). As discussed in detail
below, 5-halo substitution leads to some selectivity
between the various cloned homomeric subtypes; how-
ever, in the case of 5-nitro substitution very little
selectivity is observed. In comparison, glutamate (1)
shows very little selectivity between the various sub-
types, while AMPA (3) is 14-20-fold more selective for
AMPA receptor subtypes (but does not show selectivity
between hGluR1, -2, or -4) over hGluR5. Kainate (4)
shows 10-40-fold selectivity for hGluR5 over the AMPA
receptor subtypes (hGluR1, -2, and -4).
The binding profiles for willardiine and its analogs
(18-28) are broadly similar for homomeric hGluR1 and
hGluR2 receptors with the rank order of potencies
being: (hGluR1) 5-Cl-6-azawil > 5-F-wil > 5-Br-6-
azawil > 5-Cl-wil g 5-I-6-azawil g 5-Br-wil > 5-I-wil )
6-azawil g 5-NO2-wil g 5-Me-6-azawil g wil; (hGluR2)
5-Cl-6-azawil g 5-Br-6-azawil g 5-F-wil > 5-Cl-wil >
5-I-6-azawil ) 5-Br-wil g 5-NO2-wil > 5-I-wil > 5-Me-
6-azawil g 6-azawil > wil. For 5-halowillardiines (19-
22) the rank order of potency on both homomeric
hGluR1 and hGluR2 receptors is 5-F > 5-Cl > 5-Br >
5-I, a trend that is also observed with the 5-halo-6-aza
analogs (25, 26, 28). This order of potency is in
agreement with that observed for 19-22 in electro-
physiological studies on hippocampal cells8,9 and neo-
natal rat spinal cord.6 A similar rank order of potency
for 25, 26, and 28 was observed in an electrophysiologi-

3648 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 22
J ane et al.
oxetan-2-one (17) was synthesized by the reported proce-
dure.18,19 The 6-azauracils (14, 15) were synthesized using a
literature method.24 All other reagents were obtained from
Aldrich Chemical Co. All uracils (7-16) were dried in vacuo
(0.1 mmHg) over P2O5 for 3 days before use.
is the reverse of that observed for AMPA receptors
(hGluR1 and hGluR2 receptors), which is in agreement
with earlier reports on the electrophysiological charac-
terization of 19-22 in dorsal root ganglion cells9 and
immature rat dorsal roots.7,13 The 6-aza analogs (25,
26, 28) are no more potent than the 5-halo-substituted
willardiines (20-22), and in fact 6-aza substitution of
(S)-5-iodowillardiine (22) lowers affinity for hGluR5 30-
fold, again in agreement with an earlier report on the
electrophysiological characterization of 25, 26, and 28.13
The most potent ligand reported to date for homomeric
hGluR5 receptors, (S)-5-iodowillardiine (22) (Ki ) 0.24
( 0.06 nM), is approximately 700-fold more selective
for hGluR5 receptors over either hGluR1 or hGluR2
receptors and 4000-fold more selective for hGluR5 over
hGluR4 receptors. In addition, (S)-5-iodowillardiine
(22) (100 µM) did not significantly displace [3H]kainate
binding from homomeric hGluR6 receptors and thus
displays >400000-fold selectivity for hGluR5 over hGluR6
receptors. The selectivity of 22 for kainate receptors is
in agreement with work on the electrophysiological
characterization, where this compound was found to be
137 times more potent on kainate receptors in dorsal
root ganglion cells (EC50 ) 0.14 µM) than on AMPA
receptors in hippocampal neurons (EC50 ) 19.2 µM).9
In contrast to this, and in agreement with previous
work,6-12 5-fluorowillardiine (19) is highly selective for
AMPA receptors (hGluR1, hGluR2, and hGluR4) over
kainate receptors (hGluR5), being over 120-fold more
potent on hGluR1 than hGluR5.
In conclusion, it would appear that quite large lipo-
philic substituents at the 5-position of the uracil ring
not only are accommodated by hGluR5 receptors but
also lead to enhanced affinity for these receptors. In
contrast to this, for optimal binding affinity to hGluR1,
-2, or -4, smaller, electron-withdrawing substituents are
required. For optimal activity at hGluR4 receptors a
6-aza substituent is favored. We have identified that
22 has high affinity for hGluR5 and, compared to
kainate, displays excellent selectivity for this receptor
over both the AMPA receptor subtypes and the homo-
meric kainate receptor, hGluR6. 5-Fluorowillardiine
(19) has higher affinity than AMPA for both homomeric
hGluR1 and hGluR2 and compared to AMPA displays
greater selectivity for AMPA receptor subtypes over the
kainate receptor, hGluR5. Thus, from this study some
structural features required for optimal activity at
homomeric AMPA or kainate receptor subtypes have
been identified. It is hoped that these can be used to
design analogs with greater subtype selectivity in the
future and that such compounds will be useful tools to
probe not only the involvement of receptor subtypes in
physiological processes but also their distribution in the
CNS.
(S)-2-(2-Am in o-2-ca r b oxyet h yl)-1,2,4-t r ia zin e-3,5-d i-
on e (24). To a vigorously stirred solution of 1,2,4-triazine-
3,5-dione (13) (5.45 g, 48.2 mmol) in dry dimethylformamide
(DMF; 150 mL) was added a 60% suspension of sodium hydride
in mineral oil (1.9 g, 48 mmol). The mixture was stirred at
room temperature overnight. The next day, (S)-3-[(tert-bu-
toxycarbonyl)amino]oxetan-2-one (17) (3.00 g, 16.1 mmol) was
added portionwise over 1 h. Stirring was continued overnight.
The next day, the mixture was evaporated under reduced
pressure (0.5 mmHg). TFA (40 mL) was added to the residue
with ice-cooling, and the mixture stirred at room temperature
overnight. Excess TFA was removed under reduced pressure,
and the residue was applied to a column of Dowex 50WX8-
400 resin (0.5 mmol of cation/mL of resin). The column was
eluted with 50% aqueous ethanol to remove the excess
unbound uracil and then 1 M aqueous pyridine. The ninhy-
drin-positive fractions of the 1 M aqueous pyridine eluate were
combined and evaporated to dryness under reduced pressure.
Crystallization of the residue from water yielded (S)-2-(2-
amino-2-carboxyethyl)-1,2,4-triazine-3,5-dione (24) (0.94 g,
29%) as a white solid: mp 225-226 °C dec; [R]25 +22.38° (c
D
0.36, 6 M HCl); 1H NMR (300.4 MHz, DCl/D2O, pH 1) δ 4.42-
4.58 (m, 3H), 7.59 (s, 1H); 13C NMR (300.4 MHz, DCl/D2O, pH
1) δ 41.8, 53.9, 138.1, 153.1, 160.4, 171.6. Anal. (C6H8N4O4)
C, H, N.
(S)-2-(2-Am in o-2-ca r boxyeth yl)-6-ch lor o-1,2,4-tr ia zin e-
3,5-d ion e (25) was obtained similarly from 6-chloro-1,2,4-
triazine-3,5-dione (14) (6.94 g, 47.0 mmol), a 60% suspension
of sodium hydride in mineral oil (0.9 g, 24 mmol) in dry DMF
(100 mL), and (S)-3-[(tert-butoxycarbonyl)amino]oxetan-2-one
(17) (3.00 g, 16.1 mmol). Deprotection in TFA (40 mL) followed
by purification by ion-exchange chromatography and crystal-
lization from water as described for 24 gave (S)-2-(2-amino-
2-carboxyethyl)-6-chloro-1,2,4-triazine-3,5-dione (25) (2.26 g,
52%) as an off-white solid: mp 200.8-201.4 °C dec; [R]25
D
1
+18.97° (c 0.485, 6 M HCl); H NMR (300.4 MHz, DCl/D2O,
pH 1) δ 4.04 (dd, 1H, J ) 7.7, 7.7 Hz), 4.37-4.51 (ABX system,
2H, J ) 14.5, 4.1, 6.7 Hz); 13C NMR (300.4 MHz, DCl/D2O,
pH 1) δ 44.2, 55.4, 139.4, 152.7, 157.3, 172.8. Anal. (C6H7N4O4-
Cl‚0.5H2O) C, H, N.
(S)-2-(2-Am in o-2-ca r b oxyet h yl)-6-iod o-1,2,4-t r ia zin e-
3,5-d ion e (26) was obtained similarly from 6-iodo-1,2,4-
triazine-3,5-dione (15) (3.20 g, 13.4 mmol), a 60% suspension
of sodium hydride in mineral oil (0.5 g, 13 mmol) in dry DMF
(75 mL), and (S)-3-[(tert-butoxycarbonyl)amino]oxetan-2-one
(17) (1.25 g, 6.7 mmol). Deprotection in TFA (20 mL) followed
by purification by ion-exchange chromatography and crystal-
lization from water as described for 24 gave (S)-2-(2-amino-
2-carboxyethyl)-6-iodo-1,2,4-triazine-3,5-dione (26) (0.58 g,
27%) as yellow crystals: mp 222-223 °C dec; [R]25 +18.09°
D
(c 0.29, 6 M HCl); 1H NMR (300.4 MHz, DCl/D2O, pH 1) δ
4.44-4.63 (m, 3H); 13C NMR (300.4 MHz, DCl/D2O, pH 1) δ
43.4, 53.8, 111.4, 152.3, 158.3, 171.4. Anal. (C6H7N4O4I) C,
H, N.
(S)-2-(2-Am in o-2-ca r boxyeth yl)-6-m eth yl-1,2,4-tr ia zin e-
3,5-d ion e (27) was obtained similarly from 6-methyl-1,2,4-
triazine-3,5-dione (16) (10.0 g, 88.4 mmol), a 60% suspension
of sodium hydride in mineral oil (1.2 g, 30 mmol) in dry DMF
(200 mL), and (S)-3-[(tert-butoxycarbonyl)amino]oxetan-2-one
(17) (5.51 g, 29.5 mmol). Deprotection in TFA (80 mL) followed
by purification by ion-exchange chromatography and crystal-
lization from water as described for 24 gave (S)-2-(2-amino-
2-carboxyethyl)-6-methyl-1,2,4-triazine-3,5-dione (27) (0.89 g,
Exp er im en ta l Section
Ch em istr y: Gen er a l P r oced u r es. Melting points were
taken on a digital electrical melting point apparatus. Proton
and carbon NMR spectra were measured with a 300.4 MHz
J eol spectrometer in D2O and 20% aqueous DCl, using the
sodium salt of 3-(trimethylsilyl)propanesulfonic acid as an
internal standard. Thin layer chromatography was performed
on 60 Å Whatman precoated silica gel plates. Ion-exchange
chromatography was carried out using Dowex 50WX8-400
resin (obtained from Aldrich Chemical Co., U.K.). Elemental
analyses were performed at the Microanalytical Laboratory
of the University of Bristol. (S)-3-[(tert-Butoxycarbonyl)amino]-
10%) as white crystals: mp 204-205 °C dec; [R]25 +20.95° (c
D
0.525, 6 M HCl); 1H NMR (300.4 MHz, DCl/D2O, pH 1) δ 2.22
(s, 3H), 4.44-4.58 (m, 3H); 13C NMR (300.4 MHz, DCl/D2O,
pH 1) δ 18.6, 42.1, 53.9, 147.0, 153.3, 160.8, 171.6. Anal.
(C7H10N4O4) C, H, N.
(S)-1-(2-Am in o-2-car boxyeth yl)pyr im idin e-2,4-dion e (18)
was obtained similarly from pyrimidine-2,4-dione (7) (5.39 g,
48.1 mmol), a 60% suspension of sodium hydride in mineral

Synthesis of Willardiine and 6-Azawillardiine Analogs
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 22 3649
oil (1.8 g, 48 mmol) in dry DMF (100 mL), and (S)-3-[(tert-
butoxycarbonyl)amino]oxetan-2-one (17) (3.00 g, 16.1 mmol).
Deprotection in TFA (40 mL) followed by purification by ion-
exchange chromatography and crystallization from water as
described for 24 gave (S)-1-(2-amino-2-carboxyethyl)pyrimi-
dine-2,4-dione (18) (2.24 g, 63%) as a white solid: mp 207-
hydride in mineral oil (1.3 g, 32 mmol) in dry DMF (50 mL),
and (S)-3-[(tert-butoxycarbonyl)amino]oxetan-2-one (17) (2.00
g, 10.7 mmol). Deprotection in TFA (40 mL) followed by
purification by ion-exchange chromatography and crystalliza-
tion from water as described for 24 gave (S)-1-(2-amino-2-
carboxyethyl)-5-nitropyrimidine-2,4-dione (23) (1.41 g, 54%) as
208 °C dec (lit.15 mp 206-211 °C); [R]25 -24.4° (c 1.16, 6 M
a yellow powder: mp 222-223 °C dec; [R]25 +15.4° (c 1.62, 6
D
D
HCl) (lit.15 [R]20D -20° (c 2.0, 1 M HCl)); 1H NMR (300.4 MHz,
DCl/D2O, pH 1) δ 4.33-4.54 (ABX system, 2H, J ) 15.4, 4.5,
6.5 Hz), 4.61 (dd, 1H, J ) 4.6, 4.6 Hz), 5.88 (d, 1H, J ) 7.9
Hz), 7.70 (d, 1H, J ) 7.9 Hz); 13C NMR (300.4 MHz, DCl/D2O,
pH 1) δ 50.5, 54.6, 104.6, 149.4, 155.3, 168.8, 171.0. Anal.
(C7H9N3O4‚1.25H2O) C, H, N.
M HCl); 1H NMR (300.4 MHz, DCl/D2O, pH 1) δ 4.42-4.53
(ABX system, 2H, J ) 7.8, 3.9, 5.5 Hz), 4.58 (dd, 1H, J ) 4.7,
1.6 Hz), 9.26 (s, 1H); 13C NMR (300.4 MHz, DCl/D2O, pH 1) δ
52.3, 55.6, 130.4, 154.6, 156.4, 161.6, 172.4. Anal. (C7H8N4O6)
C, H, N.
(S)-2-(2-Am in o-2-ca r boxyeth yl)-6-br om o-1,2,4-tr ia zin e-
3,5-d ion e (28). To a stirred solution of (S)-2-(2-amino-2-
carboxyethyl)-1,2,4-triazine-3,5-dione (24) (0.10 g, 0.5 mmol)
in acetic acid (20 mL) and water (5 mL) was added a 1 M
solution of bromine in acetic acid (1.5 mL, 1.5 mmol). The
mixture was stirred in a stoppered container overnight. The
next day, the solution was evaporated under reduced pressure.
The residue was dissolved in water (5 mL) and applied to a
column of Dowex 50WX8-400 resin. The column was eluted
with water and then 1 M aqueous pyridine. Ninhydrin-
positive fractions of the 1 M aqueous pyridine eluate were
evaporated under reduced pressure. Crystallization of the
residue from water yielded (S)-2-(2-amino-2-carboxyethyl)-6-
(S)-1-(2-Am in o-2-ca r b oxyet h yl)-5-flu or op yr im id in e-
2,4-d ion e (19) was obtained similarly from 5-fluoropyrimi-
dine-2,4-dione (8) (3.13 g, 24.1 mmol), a 60% suspension of
sodium hydride in mineral oil (1.0 g, 24 mmol) in dry DMF
(80 mL), and (S)-3-[(tert-butoxycarbonyl)amino]oxetan-2-one
(17) (1.5 g, 8.0 mmol). Deprotection in TFA (30 mL) followed
by purification by ion-exchange chromatography and crystal-
lization from water as described for 24 gave (S)-1-(2-amino-
2-carboxyethyl)-5-fluoropyrimidine-2,4-dione (19) (0.78 g, 43%)
as white crystals: mp 221-222 °C dec; [R]25 -28.2° (c 1.24,
D
1
6 M HCl); H NMR (300.4 MHz, DCl/D2O, pH 1) δ 4.3-4.51
(ABX system, 2H, J ) 15.2, 4.6, 6.3 Hz), 4.59 (dd, 1H, J ) 4.6,
4.6 Hz), 7.93 (d, 1H, J ) 6.1 Hz); 13C NMR (300.4 MHz, DCl/
D2O, pH 1) δ 50.8, 54.6, 133.4 (d, J ) 139 Hz), 140.3, 143.5,
154.4, 168.4. Anal. (C7H8N3O4F‚0.4H2O) C, H, N.
bromo-1,2,4-triazine-3,5-dione (28) (0.09 g, 63%) as a white
1
solid: mp 208-210 °C dec; [R]25 +7.5° (c 0.08, 6 M HCl); H
D
(S)-1-(2-Am in o-2-ca r b oxyet h yl)-5-ch lor op yr im id in e-
2,4-d ion e (20) was obtained similarly from 5-chloropyrimi-
dine-2,4-dione (9) (7.05 g, 48.1 mmol), a 60% suspension of
sodium hydride in mineral oil (1.9 g, 48 mmol) in dry DMF
(60 mL), and (S)-3-[(tert-butoxycarbonyl)amino]oxetan-2-one
(17) (3.00 g, 16.1 mmol). Deprotection in TFA (60 mL) followed
by purification by ion-exchange chromatography and crystal-
lization from water as described for 24 gave (S)-1-(2-amino-
2-carboxyethyl)-5-chloropyrimidine-2,4-dione (20) (2.30 g, 61%)
as an off-white solid: mp 228-229 °C dec; [R]25D +2.9° (c 1.39,
NMR (300.4 MHz, DCl/D2O, pH 1) δ 4.18 (dd, 1H, J ) 7.6, 7.5
Hz), 4.4-4.54 (ABX system, 2H, J ) 14.5, 5, 7.3 Hz); 13C NMR
(300.4 MHz, DCl/D2O, pH 1) δ 44.2, 55.3, 132.7, 152.4, 157.8,
172.9. Anal. (C6H7N4O4Br) C, H, N.
Meth od ology for Bin d in g Stu d ies. Cell membranes
were prepared from frozen HEK293 cells expressing either
recombinant AMPA27,28 or kainate receptors25,26 by resuspend-
ing the cells in ice-cold distilled water, sonicating, and
centrifuging at 50000g for 10 min. The membrane pellets were
then washed in >100× volumes of 50 mM Tris-HCl buffer, pH
7.5, and centrifuged to remove endogenous glutamate. Bind-
ing reactions were performed at 4 °C for 60 min in a total
volume of 250 µL containing 50 µL of membrane suspension
(100-150 µg of protein). For kainate receptor binding (hGluR525
or hGluR626), the reaction mixture consisted of 150 µL of 50
mM Tris-HCl, pH 7.5, 25 µL of [3H]kainate (DuPont NEN),
and 25 µL of unlabeled competitor (10-12-10-3). The final [3H]-
kainate concentrations used in the competitive inhibition
experiments were 20 nM. For AMPA receptor binding
(hGluRl,27 hGluR2,27 or hGluR428) 20 nM [3H]AMPA (DuPont
NEN) was used for each receptor subtype and 100 mM KSCN
was added to the Tris-HCl buffer. Following the 60 min
incubation, the membranes were centrifuged at 50000g for 20
min to separate bound from free ligand, and the pellets were
washed three times in cold assay buffer. Nonspecific binding
was determined by incubation in the presence 10 µM glutamate.
All data was analyzed by GRAFIT 2.0 software.
1
6 M HCl); H NMR (300.4 MHz, DCl/D2O, pH 1) δ 4.35-4.54
(ABX system, 2H, J ) 15.2, 4.6, 6.5 Hz), 4.61 (dd, 1H, J ) 4.6,
4.5 Hz), 8.05 (s, 1H); 13C NMR (300.4 MHz, DCl/D2O, pH 1) δ
50.6, 54.5, 111.3, 145.9, 154.0, 164.1, 170.5. Anal. (C7H8N3O4-
Cl) C, H, N.
(S)-1-(2-Am in o-2-ca r b oxyet h yl)-5-b r om op yr im id in e-
2,4-d ion e (21) was obtained similarly from 5-bromopyrimi-
dine-2,4-dione (10) (9.19 g, 48.1 mmol), a 60% suspension of
sodium hydride in mineral oil (1.9 g, 48 mmol) in dry DMF
(60 mL), and (S)-3-[(tert-butoxycarbonyl)amino]oxetan-2-one
(17) (3.00 g, 16.1 mmol). Deprotection in TFA (60 mL),
followed by purification by ion-exchange chromatography and
crystallization from water as described for 24 gave (S)-1-(2-
amino-2-carboxyethyl)-5-bromopyrimidine-2,4-dione (21) (3.00
g, 67%) as a white solid: mp 227-228 °C dec; [R]25 +9.1° (c
D
1.76, 6 M HCl); 1H NMR (300.4 MHz, DCl/D2O, pH 1) δ 4.35-
4.53 (ABX system, 2H, J ) 15.2, 4.6, 6.4 Hz), 4.59 (dd, 1H, J
) 4.6, 4.6 Hz), 8.14 (s, 1H); 13C NMR (300.4 MHz, DCl/D2O,
pH 1) δ 50.7, 54.6, 99.2, 148.6, 154.6, 164.8, 171.2. Anal.
(C7H8N3O4Br) C, H, N.
Ack n ow led gm en t. The authors would like to thank
Andrew Pellizzari for assisting in the preparation of
recombinant cell stocks for the ligand binding assays
and Prof. J . C. Watkins and Dr. Richard Baker for
helpful comments on this manuscript. This work was
partly funded by the Medical Research Council, U.K.
(S)-1-(2-Am in o-2-ca r boxyeth yl)-5-iod op yr im id in e-2,4-
d ion e (22) was obtained similarly from 5-iodopyrimidine-2,4-
dione (11) (7.64 g, 32.1 mmol), a 60% suspension of sodium
hydride in mineral oil (1.3 g, 32 mmol) in dry DMF (40 mL),
and (S)-3-[(tert-butoxycarbonyl)amino]oxetan-2-one (17) (2.00
g, 10.7 mmol). Deprotection in TFA (40 mL) followed by
purification by ion-exchange chromatography and crystalliza-
tion from water as described for 24 gave (S)-1-(2-amino-2-
carboxyethyl)-5-iodopyrimidine-2,4-dione (22) (2.02 g, 58%) as
Refer en ces
(1) Watkins, J . C.; Krogsgaard-Larsen, P.; Honore´, T. Structure-
activity relationships in the development of excitatory amino acid
receptor agonists and competitive antagonists. Trends Pharma-
col. Sci. 1990, 11, 25-33.
(2) Pin, J . P.; Duvoisin, R. The metabotropic glutamate receptors -
structure and functions. Neuropharmacology 1995, 34, 1-26.
(3) J ane, D. E.; Olverman, H. J .; Watkins, J . C. Agonists and
competitive antagonists: structure-activity and molecular mod-
elling studies. In The NMDA Receptor, 2nd ed.; Collingridge, G.
L., Watkins, J . C., Eds.; Oxford University Press: Oxford, U.K.,
1994.
a white solid: mp 222-224 °C dec; [R]25 +15.1° (c 1.19, 6 M
D
HCl); 1H NMR (300.4 MHz, DCl/D2O, pH 1) δ 4.33-4.51 (ABX
system, 2H, J ) 15.2, 4.6, 6.4 Hz), 4.56 (dd, 1H, J ) 4.6, 4.8
Hz), 8.19 (s, 1H); 13C NMR (300.4 MHz, DCl/D2O, pH 1) δ 50.6,
54.7, 72.0, 153.5, 154.9, 165.9, 170.6. Anal. (C7H8N3O4I) C,
H, N.
(S)-1-(2-Am in o-2-ca r boxyeth yl)-5-n itr op yr im id in e-2,4-
d ion e (23) was obtained similarly from 5-nitropyrimidine-2,4-
dione (12) (5.04 g, 32.1 mmol), a 60% suspension of sodium

3650 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 22
J ane et al.
(4) Fletcher, E. J .; Lodge, D. New developments in the molecular
pharmacology of R-amino-3-hydroxy-5-methyl-4-isoxazolepropi-
onate and kainate receptors. Pharmacol. Ther. 1996, 70, 65-
89.
(5) Bettler, B.; Mulle, C. Neurotransmitter receptors. 2. AMPA and
kainate receptors. Neuropharmacology 1995, 34, 123-139.
(6) J ane, D. E.; Pook, P. C.-K.; Sunter, D. C.; Udvarhelyi, P. M.;
Watkins, J . C. New willardiine analogs with potent stereose-
lective actions on mammalian spinal neurones. Br. Pharmacol.
1991, 104, 333P.
(7) Blake, J . F.; J ane, D. E.; Watkins, J . C. Action of willardiine
analogs on immature rat dorsal roots. Br. J . Pharmacol. 1991,
104, 334P.
(8) Patneau, D. K.; Mayer, M. L.; J ane, D. E.; Watkins, J . C.
Activation and desensitization of AMPA/kainate receptors by
novel derivatives of willardiine. J . Neurosci. 1992, 12, 595-606.
(9) Wong, L. A.; Mayer, M. L.; J ane, D. E.; Watkins, J . C. Willar-
diines differentiate agonist binding sites for kainate-versus
AMPA-preferring glutamate receptors in DRG and hippocampal
neurones. J . Neurosci. 1994, 14, 3881-3897.
(10) Hawkins, L. M.; Beaver, K. M.; J ane, D. E.; Taylor, P. M.; Sunter,
D. C.; Roberts, P. J . Binding of the new radioligand (S)-[3H]-
AMPA to rat-brain synaptic membranes - effects of a series of
structural analogs of the non-NMDA receptor agonist willardi-
ine. Neuropharmacology 1995, 34, 405-410.
(11) Hawkins, L. M.; Beaver, K. M.; J ane, D. E.; Taylor, P. M.; Sunter,
D. C.; Roberts, P. J . Characterization of the pharmacology and
regional distribution of (S)-[3H]-5-fluorowillardiine binding in rat
brain. Br. J . Pharmacol. 1995, 116, 2033-2039.
(12) Hawkins, L. M.; J ane, D. E.; Roberts, P. J . The effect of 6-aza
substitution on the affinity of willardiine derivatives for the
AMPA receptor. Br. J . Pharmacol. Proc. Suppl. 1995, 117, 332P.
(13) Thompson, G. A.; J ane, D. E.; Watkins, J . C. Depolarising effects
of certain derivatives of (S)-willardiine upon in vitro neonatal
rat dorsal roots. Br. J . Pharmacol. Proc. Suppl. 1995, 117, 331P.
(14) Pinkney, J . M.; Ahluwalia, P.; Kingston, W. P.; Pook, P. C.-K.;
J ane, D. E.; Watkins, J . C. An electrophysiological study of
6-azawillardiine derivatives on mammalian spinal neurons. Br.
J . Pharmacol. 1996, 118, 50P.
(17) Doel, M. T.; J ones, A. S.; Taylor, N. An approach to the synthesis
of peptide analogues of oligonucleotides (nucleopeptides). Tet-
rahedron Lett. 1969, 2285-2288.
(18) Arnold, L. D.; Kalantar, T. H.; Vederas, J . C. Conversion of serine
to stereochemically pure â-substituted R-amino acids via â-lac-
tones. J . Am. Chem. Soc. 1985, 107, 7105-7109.
(19) Ramer, S. E.; Moore, R. N.; Vederas, J . C. Mechanism of
formation of serine-â-lactones. Can. J . Chem. 1986, 64, 706-
713.
(20) Arnold, L. D.; Drover, J . C. G.; Vederas, J . C. Conversion of
serine-â-lactones to chiral R-amino acids by copper-containing
organolithium and organomagnesium reagents. J . Am. Chem.
Soc. 1987, 109, 4649-4659.
(21) Arnold, L. D.; May, R. G.; Vederas, J . C. Synthesis of optically
pure R-amino acids via salts of R-amino-â-propiolactone. J . Am.
Chem. Soc. 1988, 110, 2237-2241.
(22) Nollet, A. J . H.; Pandit, U. K. Unconventional nucleotide analogs
- III 4-(N1-Pyrimidyl)-2-aminobutyric acids. Tetrahedron 1968,
25, 5989-5994.
(23) J onas, J .; Gut, J . Nucleic acid components and their analogs.
XVI Dissociation constants of uracil, 6-azauracil, 5-azauracil and
related compounds. Collect. Czech. Chem. Commun. 1962, 27,
716-723.
(24) Chang, P. K. New 5-substituted 6-azauracils. J . Org. Chem.
1961, 1118-1120.
(25) Korczak, B.; Nutt, S. L.; Fletcher, E. J .; Hoo, K. H.; Elliott, C.
E.; Rampersad, V.; McWhinnie, E. A.; Kamboj, R. K. cDNA
cloning and functional properties of human glutamate receptor
EAA3 (GluR5) in homomeric and heteromeric configuration.
Receptors Channels 1995, 3, 41-49.
(26) Hoo, K. H.; Nutt, S. L.; Fletcher, E. J .; Elliott, C. E.; Korczak,
B.; Deverill, R. M.; Rampersad, V.; Fantaske, R. P.; Kamboj, R.
K. Functional expression and pharmacological characterization
of the human EAA4 (GluR6) glutamate receptor: A kainate
selective channel subunit. Receptors Channels 1994, 2, 327-337.
(27) Kamboj, R.; Elliott, C.; Nutt, S. L. Amino-hydroxy-methyl-
isoxazole-propionate binding human glutamate receptors. Eur.
Pat. Appl. 93304500.7, 1993.
(28) Fletcher, E. J .; Nutt, S. L.; Hoo, K. H.; Elliott, C. E.; Korczak,
B.; McWhinnie, E. A., Kamboj, R. K. Cloning, expression and
pharmacological characterization of a human glutamate recep-
tor: hGluR4. Receptors Channels 1995, 3, 21-31.
(15) Dewar, J . H.; Shaw, G. Purines, pyrimidines and imidazoles.
Part XVII. A synthesis of willardiine. J . Chem. Soc. 1962, 1,
583-585.
(16) Martinez, A. P.; Lee, W. W. An improved synthesis of willardiine
and 1-(2′,2′-diethoxyethyl)uracil. J . Org. Chem. 1965, 317-318.
J M9702387