AMPA receptor agonists: Synthesis, protolytic properties, and pharmacology of 3-isothiazolol bioisosteres of glutamic acid

Article abstract of DOI:10.1021/jm9607212

A number of 3-isothiazolol bioisosteres of glutamic acid (1) and analogs of the AMPA receptor agonist, (RS)-2-amino-3-(3-hydroxy-5-methylisoxazol-4- yl)propionic acid (AMPA, 2a), including (RS)-2-amino-3-(3-hydroxy-5- methylisothiazol-4-yl)propionic acid (thio-AMPA, 2b), were synthesized. Comparative in vitro pharmacological studies on this series of 3-isothiazolol and the corresponding 3-isoxazolol amino acids were performed using a series of receptor binding assays (IC₅₀ values) and the electrophysiological rat cortical slice model (EC₅₀ values). Whereas 2a (IC₅₀ = 0.04 ± 0.005 μM, EC₅₀ = 3.5 ± 0.2 μM) is markedly more potent than the tert-butyl analog ATPA (3a) (IC₅₀ = 2.1 ± 0.16 μM, EC₅₀ = 34 ± 2.4 μM) in [³H]AMPA binding and electrophysiological studies, 2b (IC₅₀ = 1.8 ± 0.13 μM, EC₅₀ = 15.0 ± 2.4 μM) was approximately equipotent with thio-ATPA (3b) (IC₅₀ = 0.63 ± 0.07 μM, EC₅₀ = 14 ± 1.3 μM). (RS)-2-Amino-3-(3- hydroxyisoxazol-5-yl)propionic acid (HIBO, 4a) was approximately equipotent with its thio analog 4b, whereas 4-Br-HIBO (5a) (IC₅₀ = 0.65 ± 0.12 μM, EC₅₀ = 22 ± 0.6 μM) turned out to be much more potent than the corresponding 3-isothiazolol 5b (IC₅₀ = 17 ± 2.2 μM, EC₅₀ = 500 ± 23 μM), 2b (ED₅₀ = 130 μmol/kg) was more potent than 2a (220 μmol/kg) as a convulsant after subcutaneous administration in mice. The protolytic properties of 2a,b-4a,b were determined using ¹³C NMR spectroscopy. For each pair of compounds, the α-amino acid groups showed similar protolytic properties, whereas the 3-isoxazolol moieties typically showed pK(a) values 2 units lower than those of the 3-isothiazolols. Accordingly, calculations of ionic species distributions revealed pronounced differences between 3- isoxazolol and 3-isothiazolol amino acids. No simple correlation between activity as AMPA agonists in vitro and pK(a) values of these compounds was apparent. On the other hand, the relative potencies of AMPA (2a) and thio- AMPA (2b) in vitro and in vivo may reflect that these compounds predominantly penetrate the blood-brain barrier as net uncharged diprotonated ionic species.

Full text of DOI:10.1021/jm9607212

+
+
520
J . Med. Chem. 1997, 40, 520-527
AMP A Recep tor Agon ists: Syn th esis, P r otolytic P r op er ties, a n d P h a r m a cology
of 3-Isoth ia zolol Bioisoster es of Glu ta m ic Acid
Lisa Matzen, Anne Engesgaard, Bjarke Ebert, Michael Didriksen,† Bente Frølund,
Povl Krogsgaard-Larsen,* and J erzy W. J aroszewski
PharmaBiotec Research Center, Department of Medicinal Chemistry, The Royal Danish School of Pharmacy,
2 Universitetsparken, DK-2100 Copenhagen, Denmark, and Department of Pharmacological Research,
H. Lundbeck A/ S, DK-2500 Valby-Copenhagen, Denmark
Received October 15, 1996X
A number of 3-isothiazolol bioisosteres of glutamic acid (1) and analogs of the AMPA receptor
agonist, (RS)-2-amino-3-(3-hydroxy-5-methylisoxazol-4-yl)propionic acid (AMPA, 2a ), including
(RS)-2-amino-3-(3-hydroxy-5-methylisothiazol-4-yl)propionic acid (thio-AMPA, 2b), were syn-
thesized. Comparative in vitro pharmacological studies on this series of 3-isothiazolol and the
corresponding 3-isoxazolol amino acids were performed using a series of receptor binding assays
(IC50 values) and the electrophysiological rat cortical slice model (EC50 values). Whereas 2a
(IC50 ) 0.04 ( 0.005 µM, EC50 ) 3.5 ( 0.2 µM) is markedly more potent than the tert-butyl
analog ATPA (3a ) (IC50 ) 2.1 ( 0.16 µM, EC50 ) 34 ( 2.4 µM) in [3H]AMPA binding and
electrophysiological studies, 2b (IC50 ) 1.8 ( 0.13 µM, EC50 ) 15.0 ( 2.4 µM) was approximately
equipotent with thio-ATPA (3b) (IC50 ) 0.63 ( 0.07 µM, EC50 ) 14 ( 1.3 µM). (RS)-2-Amino-
3-(3-hydroxyisoxazol-5-yl)propionic acid (HIBO, 4a ) was approximately equipotent with its thio
analog 4b, whereas 4-Br-HIBO (5a ) (IC50 ) 0.65 ( 0.12 µM, EC50 ) 22 ( 0.6 µM) turned out
to be much more potent than the corresponding 3-isothiazolol 5b (IC50 ) 17 ( 2.2 µM, EC50
)
500 ( 23 µM). 2b (ED50 ) 130 µmol/kg) was more potent than 2a (220 µmol/kg) as a convulsant
after subcutaneous administration in mice. The protolytic properties of 2a ,b-4a ,b were
determined using 13C NMR spectroscopy. For each pair of compounds, the R-amino acid groups
showed similar protolytic properties, whereas the 3-isoxazolol moieties typically showed pKa
values 2 units lower than those of the 3-isothiazolols. Accordingly, calculations of ionic species
distributions revealed pronounced differences between 3-isoxazolol and 3-isothiazolol amino
acids. No simple correlation between activity as AMPA agonists in vitro and pKa values of
these compounds was apparent. On the other hand, the relative potencies of AMPA (2a ) and
thio-AMPA (2b) in vitro and in vivo may reflect that these compounds predominantly penetrate
the blood-brain barrier as net uncharged diprotonated ionic species.
In tr od u ction
Glutamic acid (1) (Figure 1), which is the main
excitatory neurotransmitter in the central nervous
system (CNS), and other excitatory amino acids (EAAs)
operate through four different classes of receptors. In
addition to the three heterogeneous classes of ionotropic
EAA receptors (iGluRs), named N-methyl-D-aspartic
acid (NMDA), (RS)-2-amino-3-(3-hydroxy-5-methyl-
isoxazol-4-yl)propionic acid (AMPA, 2a ), and kainic acid
receptors,1-3 a heterogeneous class of metabotropic EAA
receptors (mGluRs) has been shown to have important
functions in the central excitatory neurotransmission
processes.4 It is now generally agreed that iGluRs as
well as mGluRs play important roles in the healthy as
well as in the diseased CNS and that all subtypes of
these receptors are potential targets for therapeutic
intervention in a number of diseases.5,6
EAA receptors are involved in the mechanisms of
long-term potentiation, which is believed to play an
important role in learning and memory functions, and
the deficits of these functions in the Alzheimer patients
may, to some extent, be caused by hypoactivity at
iGluRs and/or mGluRs in the brain.7-10 There is also
a growing evidence of an implication of EAA receptors
F igu r e 1. Structures of glutamic acid (1), the previously
described AMPA receptor agonists AMPA (2a ), ATPA (3a ),
HIBO (4a ), and Br-HIBO (5a ), and the new compounds thio-
AMPA (2b), thio-ATPA (3b), 4b, and 5b.
in schizophrenia.11,12 As in Alzheimer’s disease, the role
of these receptors in the etiology and the clinical
manifestations of schizophrenia is still very incom-
pletely understood, but there is evidence to suggest that
* Corresponding author. Phone: (+45) 35 37 08 50, ext. 511. Fax:
(+45) 35 37 22 09.
† H. Lundbeck A/S.
X Abstract published in Advance ACS Abstracts, J anuary 15, 1997.
S0022-2623(96)00721-2 CCC: $14.00 © 1997 American Chemical Society

+
+
AMPA Receptor Agonists
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 4 521
Sch em e 1a
Sch em e 2a
a
a
(i) H2S, HCl; (ii) I2, K2CO3; (iii) 1,3,5-trioxane, 62% aqueous
(i) MeOH, HCl; (ii) MeI, K2CO3; (iii) NaBH4; (iv) SOCl2; (v)
HBr, MeOH; (iv) AcNHCH(COOMe)2, NaH; (v) 1 M aqueous
CF3COOH.
AcNHCH(COOMe)2, NaH; (vi) Br2; (vii) 4 M HCl; (viii) 48%
aqueous HBr.
hypoactivity at EAA receptors also is a factor of impor-
tance in the latter CNS disorder.12-14
conditions resulted in the formation of 4-bromo-5-
methylisothiazol-3-ol in varying quantities. With iodine
as oxidizing agent, 7c was converted into 8c without
formation of iodinated byproducts, and the reaction
conditions for the conversion of 6c into 8c were suc-
cessfully applied to the synthesis of 8d from 6d ,
although the yield of the intermediate 7d (20%) was
lower than that of 7c (36%). In both cases, the desired
reactions were accompanied by the formation of sub-
stantial amounts of decomposition products. Steric
hindrance by the tert-butyl group may explain the
particularly low yield of 7d . Although the oxa analog
of 9c has been prepared in almost quantitative yield
from 5-methylisoxazol-3-ol,38 the same reaction only
provided 9c,d in yields of 27% and 39%, respectively.
In attempts to increase these yields, prolonged reaction
times and higher reaction temperatures led to extensive
decomposition, whereas reduction of reaction time and/
or temperature resulted in incomplete bromomethyl-
ation. Deprotection under acidic conditions of interme-
diates 10c,d provided thio-AMPA (2b) and thio-ATPA
(3b) in yields of 28% and 31%, respectively.
Compounds 4b and 5b were synthesized according to
Scheme 2. Methylation of 12, prepared by methanolysis
of 11,39 provided 13 (68%) and the corresponding N-
methylated product (not shown) in 26% yield. Com-
pound 13 was reduced40 to 14, which was further
converted into fully protected compound 16. The depro-
tection of 16, which required concentrated hydrobromic
acid, was accompanied with marked decomposition of
the product and provided 4b‚HBr in a yield of only 25%.
Loss of bromine was observed after treatment of the
brominated analog of 16, compound 17, with concen-
trated hydrobromic acid. Deprotection of 17 without
significant loss of bromine was achieved by reflux in 4
M HCl for a period of less than 4 h, but, again,
pronounced decomposition resulted in a low yield (20%)
of 5b‚HCl.
In Alzheimer’s disease as well as in schizophrenia,
EAA receptor agonists, partial agonists, or functional
partial agonists may have therapeutic interest.15,16
While we have previously described the principles for
the establishment of functional partial agonism at
iGluRs,10,15-18 attempts to develop truly partial agonists
at these classes of EAA receptors have so far been
unsuccessful.15 Thus, structure-activity studies on a
comprehensive series of analogs of AMPA (2a ) contain-
ing alkyl groups of different sizes and polarities show
either full AMPA agonism or inactivity.19-25 Further-
more, although there is circumstantial evidence to
suggest that the tert-butyl analog of AMPA, (RS)-2-
amino-3-(5-tert-butyl-3-hydroxyisoxazol-4-yl)propionic acid
(ATPA, 3a ), shows the capacity to enter the CNS after
systemic administration,26,27 principles for the design
of EAA receptor ligands capable of penetrating the
blood-brain barrier (BBB) have not been developed yet.
In the 4-aminobutyric acidA (GABAA) agonist field, the
ability to penetrate the BBB is largely determined by
the protolytic properties of the zwitterionic com-
pounds.28-30 Although the 3-isothiazolol unit is mark-
edly less acidic than the 3-isoxazolol unit, both of these
heterocyclic systems are effective bioisosteres of the
carboxyl group in this CNS transmitter system.29,31
These aspects prompted us to synthesize the 3-iso-
thiazolol analogs 2b-5b of the previously described
3-isoxazolol AMPA receptor agonists 2a -5a 19,32-36 and
to describe the protolytic and pharmacological properties
of these new compounds. The in vitro pharmacological
studies comprise investigations of the relationship
between AMPA receptor affinity and agonist potency
and efficacy. The potencies of AMPA (2a ) and thio-
AMPA (2b) as convulsants after systemic administra-
tion in mice were compared.
Resu lts
Acid ity Con sta n ts a n d Ion ic Sp ecies Distr ibu -
tion . The use of 13C NMR for determination of pro-
tolytic properties is based upon the measurement of
chemical shift changes as a function of pH.41 The pKa
values for 2a ,b, 3a ,b, 4a ,b, and 5a ,b (Table 1) were
determined by nonlinear fitting of the NMR titration
Ch em istr y. Thio-AMPA (2b) and thio-ATPA (3b)
were synthesized as outlined in Scheme 1. Compound
8c has previously been synthesized from 6c via 7c, the
latter being converted into 8c by oxidation with bro-
mine.37 In preliminary experiments, these reaction

+
+
522 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 4
Ta ble 1. pKa Values, Molar Fraction of Diprotonated Ionic Species (x2), Receptor Binding, and Electrophysiological Data
Matzen et al.
IC50 (µM),
EC50 (µM),
compound
pKa values
x2, pH 7.4
[3H]AMPAa
electrophysiologya
AMPA (2a )
thio-AMPA (2b)
ATPA (3a )
thio-ATPA (3b)
HIBO (4a )
4b
Br-HIBO (5a )
5b
glutamic acid (1)
1.94 ( 0.05,h 5.12 ( 0.03,f 10.09 ( 0.01j
1.84 ( 0.05,j 7.00 ( 0.01,f 10.27 ( 0.01j
2.27 ( 0.13,j 4.76 ( 0.09,g 10.17 ( 0.02j
2.31 ( 0.05,h 7.03 ( 0.08,h 10.56 ( 0.07j
1.90 ( 0.04,i 5.07 ( 0.16,f 8.81 ( 0.02h
1.82 ( 0.05,i 6.69 ( 0.03,g 9.01 ( 0.05g
1.75 ( 0.06,h 3.49 ( 0.03,f 8.70 ( 0.02h
1.64 ( 0.04,h 5.35 ( 0.04,f 8.77 ( 0.03h
2.23,k 4.37,k 9.53k
0.01
0.28
0.01
0.30
0.01
0.16
0.01
0.01
0.01
0.04 ( 0.005b
1.8 ( 0.13
2.1 ( 0.16c
0.63 ( 0.07
1.5 ( 0.8e
5.8 ( 1.4
3.5 ( 0.2b
15.0 ( 2.4
34 ( 2.4d
14 ( 1.3
370 ( 10e
380 ( 22
22 ( 0.6
0.65 ( 0.12
17 ( 2.2
500 ( 23
a
b
d
Mean ( SEM. Reference 22. c Reference 21. Reference 14. e Reference 36. f -jBased on carbon atoms C-3, C-4, R-CH, CH2, and
k
COOH, respectively. Reference 43.
F igu r e 3. Structures of the triprotonated (2a 3, 2b3), dipro-
tonated (2a 2, 2b2), monoprotonated (2a 1, 2b 1), and nonpro-
tonated (2a 0, 2b0) ionic species of AMPA (2a ) and thio-AMPA
(2b).
atom to give 5a ,b, respectively, decreased the pKa value
of the hydroxyl group by about 1.5 pKa units as
expected. Except for 5a , the pKa values of the two
heterocyclic bioisosteres were higher than the pKa value
of the distal carboxyl group in glutamic acid.43
Based upon the pKa values obtained, ionic species
distribution curves (Figure 4) were calculated for the
amino acids 2a ,b, 3a ,b, 4a ,b, and 5a ,b using eqs 2-6,
where x0, x1, x2, and x3 are the molar fractions of non-,
mono-, di-, and triprotonated species present.
F igu r e 2. 13C NMR titration curves for AMPA (2a ) and thio-
AMPA (2b) in D2O (25 °C, ionic strength 1 M): (A) for C-3
carbon atoms and (B) for methine (CH) carbon atoms.
curve, defined by eq 1, to the experimental data for
appropriate carbon atoms (Figure 2). In eq 1 δi is the
observed chemical shift of the ith site, δ0i , δi1, δi2, and
δ3i are the chemical shifts of the ith site in non-, mono-,
di-, and triprotonated ionic species, and K1, K2, and K3
are the equilibrium constants for the successive depro-
tonations of fully protonated forms (Figure 3).
x0 ) K2K3/[H+]2F
(2)
x1 ) K2/[H+]F
x1 ) 1/F
(3)
(4)
δi ) (δi0K1K2K3 + δ1i K1K2[H+] + δ2i K1[H+]2 +
δ3i [H+]3)/([H+]3 + K1[H+]2 + K1K2[H+] + K1K2K3) (1)
x3 ) [H+]2/K1F
(5)
where F ) [H+]/K1 + 1 + K2/[H+] + K2K3/[H+]2 (6)
The NMR titration curves were obtained at relatively
high salt concentration (1 M) to maintain a constant
ionic strength during the experiment. The pKavalues
of the amino and the caboxyl groups were closely similar
for the 3-isoxazolol and the 3-isothiazolol analogs,
whereas the pKa values of the 3-hydroxyl group were
1.62-2.27 pKa units lower in the 3-isoxazolol series than
in the 3-isothiazolol series (Table 1). Introduction of the
R-amino acid moiety and alkyl substituents lowered the
pKa values of the hydroxyl group by approximately 0.5
pKa unit compared to the unsubstituted 3-isoxazolol and
3-isothiazolol units.42 Replacement of the hydrogen
atom in the 4-position of HIBO and 4b with a bromine
The molar fractions at pH ) 7.4 of the net uncharged,
diprotonated ionic species, x2, in which the amino group
as well as the 3-hydroxyl group are protonated and the
carboxyl group is ionized, are shown in Table 1. The
3-isoxazolol analogs 2a -5a as well as the 3-isothiazolol
analog 5b exist mainly as the negatively charged,
monoprotonated ionic species at physiological pH, the
hydroxyl and carboxyl groups being completely depro-
tonated. By contrast, the calculation of the molar
fraction of the diprotonated species present in the
solution of 2b-4b gave values in the range of 0.16-
0.30 (Table 1), suggesting that the net uncharged,

+
+
AMPA Receptor Agonists
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 4 523
F igu r e 5. Dose-response curves as determined in the rat
cortical wedge preparation for AMPA (2a ) and analogs 2b and
3a ,b. Values are mean values ( SEM relative to the maximal
AMPA response. Data were fitted to the equation: % response
) MAX × [Ago]n/(EC50 + [Ago]n), where MAX is the maximal
n
response relative to the AMPA plateau response, [Ago] is the
agonist concentration in µM, and n is the Hill slope, deter-
mined to be close to 2 for all compounds; 100% response is
determined as the maximal response for AMPA (for details,
see ref 16).
F igu r e 4. Ionic species distribution diagrams: (A) thio-AMPA
(2b) and (B) AMPA (2a ).
diprotonated ionic species are significant at physiologi-
cal conditions.
In Vitr o P h a r m a cology. The affinities of the new
compounds for AMPA, NMDA, and kainic receptor sites
were determined using [3H]AMPA,44 [3H][3-(2-carboxy-
piperazin-4-yl)propyl]phosphonic acid ([3H]CPP),45 and
[3H]kainic acid46 as radioligands (Table 1). None of the
compounds showed detectable affinity for [3H]CPP- or
[3H]kainic acid-binding sites (IC50 > 100 µM). While
AMPA (2a ) is markedly more potent than ATPA (3a )
in [3H]AMPA binding, the affinities of thio-AMPA (2b)
and thio-ATPA (3b) were comparable. Introduction of
a bromine atom into the 4-position of the isoxazole ring
of HIBO (4a ) to give Br-HIBO (5a ) is accompanied by
an increase in affinity for AMPA binding sites. On the
other hand, 5b is proportionally weaker than 4b, the
3-isothiazolol analog of 4a (Table 1). With the exception
of HIBO (4a ) and 4b, a relatively good correlation was
observed between AMPA receptor affinity of the com-
pounds and their agonistic effects (Figure 5) sensitive
to the AMPA receptor antagonist 6-nitro-7-sulfamoyl-
benzo[f]quinoxaline-2,3-dione (NBQX).47 Compound 4b
and, in particular, 4a showed weaker AMPA agonist
effects than predicted from their AMPA receptor affini-
ties. Interestingly, (RS)-2-amino-3-(3-hydroxyisoxazol-
4-yl)propionic acid (demethyl-AMPA), which also has an
unsubstituted position in the heterocyclic ring, is, like
4a ,b, a potent inhibitor of [3H]AMPA binding but a very
weak AMPA receptor agonist.25 It has been hypoth-
esized that the recognition site of the AMPA receptor
contains a lipophilic cavity and that occupancy of this
putative cavity by appropriately sized alkyl groups in
the isoxazole ring of AMPA contributes to the stabiliza-
tion of the agonist conformation of the AMPA receptor.25
A similar effect may explain the relatively low agonist
potency of 4a ,b.
Ta ble 2. Convulsive Potencies of AMPA (2a ) and Thio-AMPA
(2b) after Subcutaneous Administration in Mice
compound
ED50 (µmol/kg)a
AMPA (2a )
thio-AMPA (2b)
220
130
a
Reference 27.
In Vivo P h a r m a cology. The convulsant activity of
AMPA (2a ) and thio-AMPA (2b) was studied in mice
after subcutaneous administration. The results are
shown in Table 2. Both AMPA (2a ) and thio-AMPA (2b)
induced convulsions, thio-AMPA (2b) being more potent
than AMPA (2a ).
Investigation of urine collected 30 min after the
administration of 20 mg/kg 2b by 1H NMR strongly
suggested that the compound is mainly excreted in
unchanged form. Thus, the 400 MHz 1H NMR spectrum
of urine from the 2b-treated mice contained a singlet
at δ 2.40 (CH3) not present in the urine from control
mice (Figure 6). In addition, the eight lines of the AB
part of the ABX system of the amino acid side chain
(the methylene group) were readily recognizable around
δ 3.0. The identity of the signals was confirmed by
spiking with authentic 2b.
Discu ssion
The synthesis and pharmacological characterization
of the 3-isothiazolol amino acids 2b-5b, which are
analogs of the previously described 3-isoxazolol amino
acids AMPA (2a ),32,33 ATPA (3a ),19,34 HIBO (4a ),32,36 and
Br-HIBO (5a ),32,33 respectively (Table 1), served several
purposes, including comparative in vitro pharmacologi-
cal studies of these structurally related compounds. Like

+
+
524 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 4
Matzen et al.
AMPA agonists. In general, the pKa values of the
carboxyl and amino groups in the 3-isothiazolol series
were closely similar to those of the 3-isoxazolol analogs,
although some minor differences were observed. As
expected from the higher electronegativity of the oxygen
atom compared to the sulfur atom, the pKa values of
the 3-isoxazolol groups of 2a -5a were consistently lower
by 1.5-2 pKa units. From the point of view of systemic
pharmacological activity, it is interesting to determine
the ionic species of the test compounds (Table 1) present
at physiological pH. The net uncharged, diprotonated
species are believed to be capable of penetrating the
BBB. Calculation of species distribution showed that
the proportion of net uncharged, diprotonated species
of thio-AMPA (2b), thio-ATPA (3b), and thio-HIBO (4b)
becomes significant at physiological pH (Figure 4).
Contrary to this, 2a -5a and also 5b mainly exist in the
negatively charged, monoprotonated forms at physi-
ological pH. The significance of the ionic species
distribution for the systemic activity was elucidated by
the comparison of the convulsive activities of 2a ,b after
subcutaneous administration in mice. Although being
intrinsically less potent, 2b showed higher convulsant
activity in mice than 2a (Table 2).
The 1H NMR spectra of urine collected from 2b-
treated mice revealed the presence of unchanged com-
pound (Figure 6) and apparently no metabolites con-
taining the 5-methyl group. To our knowledge, this is
the first report about the metabolic fate of a glutamic
acid analog containing the isothiazole moiety.
In conclusion, no simple correlation between pKa
values of 3-isoxazolol or 3-isothiazolol analogs of glutam-
ic acid (1) and their potency as AMPA receptor agonists
was apparent. All of the compounds studied, 2a -5a
and 2b-5b, showed full AMPA agonism of widely
different potency. Although thio-AMPA (2b) is mark-
edly weaker than AMPA (2a ) as an AMPA agonist, 2b
is the more potent convulsant after subcutaneous
administration in mice. On the basis of analyses of ionic
species distribution of 2a ,b at physiological pH, we
suggest that these compounds penetrate the BBB in the
net uncharged, diprotonated form.
F igu r e 6. Detection of thio-AMPA (2b) in urine: (A) 400 MHz
1H NMR spectrum of urine containing 10% D2O, collected 30
min after subcutaneous administration of 20 mg/kg thio-AMPA
(2b) to mice, (B) expansion of spectrum A showing signals of
thio-AMPA (2b) (starred), and (C) corresponding region of the
spectrum of urine collected from untreated mice.
2a -5a , all the new compounds were shown to be
selective AMPA receptor agonists, exhibiting no detect-
able affinity for NMDA or kainic acid receptor sites
(Table 1). However, no simple correlation between
structure and biological activity of these compounds was
apparent. Thus, whereas 2b was markedly weaker
than 2a as an AMPA agonist, the corresponding tert-
butyl analogs 3b,a showed the reverse relative potency,
3b being some 3 times more potent than 3a . Introduc-
tion of bromine atoms into the 4-position of the rings of
the approximately equipotent compounds 4a ,b had
opposite effects, 5a being more potent than 4a and 5b
less potent than 4b. These structure-activity relation-
ships indicate that a number of factors, including steric
bulk of substituents, protolytic properties, and probably
electrostatic parameters, have to be considered in stud-
ies of the structural requirements for binding to and
activation of AMPA receptors.
One of our objectives of replacing the ring oxygen
atoms of 2a -5a by sulfur actually was to convert these
full AMPA agonists into partial agonists. Since all of
the new compounds 2b-5b were shown to produce full
AMPA receptor agonism, as exemplified in Figure 5, this
attempt to reduce the relative efficacy of AMPA receptor
agonists obviously was unsuccessful. Another attempt
to convert the full AMPA receptor agonist AMPA (2a )
into partial agonists by substitution of a hydrogen atom
or a variety of alkyl groups for the methyl group in
AMPA was equally unsuccessful.25 Without exceptions,
these analogs of AMPA were either full AMPA agonists
or inactive.
Exp er im en ta l Section
Ch em istr y. Gen er a l P r oced u r es. Melting points were
determined in capillary tubes and are uncorrected. Elemental
analyses were performed by Mrs. K. Linthoe, Department of
General and Organic Chemistry, University of Copenhagen,
Mr. G. Cornali, Microanalytical Laboratory, Leo Pharmaceuti-
cal Products, Denmark, or Analytical Research Department,
H. Lundbeck A/S, Denmark, and are within 0.4% of the
calculated values, unless otherwise stated. 1H and 13C NMR
spectra were recorded on a Bruker AC 200 or AMX 400 WB
spectrometer, using CDCl3, D2O (with 1,4-dioxane, δ ) 3.69,
as an internal standard), H2O-D2O (9:1), or DMSO-d6 as
solvent. Drying of organic phases was performed using
MgSO4. Column chromatography (CC) and TLC were per-
formed on silica gel 60 (70-230 mesh; Merck) and silica gel
F254 plates (Merck), respectively. The compounds were visual-
ized on TLC plates using UV light and KMnO4 spraying
reagent. Compounds containing amino groups were visualized
using a ninhydrin-spraying reagent, whereas compounds
containing the 3-isothiazolol unit were visualized using a
FeCl3-spraying reagent.
4,4-Dim eth yl-3-oxop en ta n a m id e (6d ). Aqueous NH3
(25%, 645 mL) was added dropwise to ethyl 4,4-dimethyl-3-
oxopentanoate48 (64.3 g, 374 mmol) with ice cooling. The
mixture was stirred for 72 h at room temperature. Evapora-
tion and CC [eluent: toluene-EtOAc (9:1)] followed by re-
A further goal of this project was to shed light on the
relationship between protolytic properties and the abil-
ity to penetrate the BBB of this structural class of

+
+
AMPA Receptor Agonists
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 4 525
crystallization (EtOAc-light petroleum) afforded 6d (33.7 g,
63%): mp 83-84 °C. 1H NMR (DMSO-d6): δ 7.37 (s, 1H), 7.00
(s, 1H), 3.38 (s, 2H), 1.07 (s, 9H). Anal. (C7H13NO2) C, H, N.
(RS)-2-Am in o-3-(3-h yd r oxy-5-m et h ylisot h ia zol-4-yl)-
p r op ion ic Acid Mon oh yd r a te (Th io-AMP A, 2b). A solu-
tion of 10c (2.2 g, 6.1 mmol) in 1 M aqueous CF3CO2H (45
mL, 45 mmol) was refluxed for 12 h and evaporated to dryness.
The residue was dissolved in H2O (10 mL) and subjected to
ion exchange chromatography (IRA-400) using 1 M AcOH as
the eluent. Evaporation of the ninhydrin-positive fractions
and recrystallization (EtOH-H2O) gave 2b (350 mg, 28%): mp
180-184 °C dec. 1H NMR (D2O): δ 4.14 (t, 1H, J ) 6.4 Hz),
2.92 (d, 2H, J ) 6.5 Hz), 2.26 (s, 3H). 13C NMR [D2O-H2O
(1:9), pH ) 12.5]: δ 185.2 (CO2H), 180.3 (C-3), 159.0 (C-5),
123.7 (C-4), 58.5 (CH), 34.2 (CH2), 15.2 (CH3). 13C NMR [D2O-
H2O (1:9), pH ) 1.08]: δ 174.2 (C-3), 174.1 (CO2H), 160.4 (C-
5), 118.6 (C-4), 55.0 (CH), 28.3 (CH2), 15.1 (CH3). Anal.
(C7H10N2O3S‚1H2O) C, H, N, S.
(RS)-2-Am in o-3-(5-ter t-bu tyl-3-h yd r oxyisoth ia zol-4-yl)-
p r op ion ic Acid Hem ih yd r a te (Th io-ATP A, 3b). Com-
pound 3b was synthesized as described for 2b using 10d (370
mg, 0.9 mmol) in 1 M aqueous CF3CO2H (10 mL, 10 mmol).
Recrystallization (H2O) gave 3b (70 mg, 31%): mp 216 °C dec.
1H NMR (D2O/CF3CO2D): δ 4.07 (dd, 1H, J ) 5.0, 8.3 Hz),
3.25 (dd, 1H, J ) 5.2, 16.6 Hz), 3.01 (dd, 1H, J ) 8.7, 16.6
Hz), 1.36 (s, 9H). 13C NMR [D2O-H2O (1:9), pH ) 12.3]: δ
185.0 (CO2H), 181.4 (C-3), 173.6 (C-5), 122.1 (C-4), 58.4 (CH),
37.0 (quart C), 34.9 (CH2), 32.5 [(CH3)3]. 13C NMR [D2O-H2O
(1:9), pH ) 0.84]: δ 173.8 (CO2), 173.7 (C-3), 174.6 (C-5), 117.6
(C-4), 54.8 (CH), 37.5 (quart C), 31.5 [(CH3)3], 29.7 (CH2). Anal.
(C10H16N2O3S‚0.5H2O) C, N; H: calcd, 6.76; found, 6.25. S:
calcd, 12.65; found, 13.12.
Meth yl 3-Hydr oxyisoth ia zole-5-ca r boxyla te (12). 3-Hy-
droxyisothiazole-5-carboxamide (11)39 (14.2 g, 99 mmol) was
added to an ice-cooled solution of methanolic HCl prepared
from MeOH (460 mL) and acetyl chloride (92 mL). Stirring
for 12 h at 50 °C, evaporation, CC [eluent: toluene-2-propanol
(10:1)], and recrystallization (toluene) gave 12 (7.1 g, 45%):
mp 170-172 °C. 1H NMR (CDCl3/DMSO-d6): δ 7.06 (s, 1H),
3.90 (s, 3H), 10.10 (s, 1H). Anal. (C5H5NO3S) C, H, N, S.
Meth yl 3-Meth oxyisoth iazole-5-car boxylate (13). Meth-
yl iodide (16.3 mL, 72 mmol) was added dropwise to a
suspension of 12 (2.7 g, 17 mmol) and K2CO3 (5.9 g, 43 mmol)
in acetone (160 mL). Stirring for 24 h in the dark, evaporation,
and CC (eluent: EtOAc) gave the O-alkylated compound 13
and the N-alkylated compound methyl 2-methyl-3-oxoisothi-
azoline-5-carboxylate. Fractions containing this undesired
reaction product were discarded. Sublimation (45 °C, 0.5
mmHg) of the O-alkylated compound gave 13 (2.0 g, 68%): mp
54-55 °C. 1H NMR (CDCl3): δ 7.11 (s, 1H), 4.03 (s, 3H), 3.98
(s, 3H). Anal. (C6H7NO3S) C, H, N, S.
4,4-Dim eth yl-3-th ioxop en ta n a m id e (7d ). Dry EtOH
(175 mL) was saturated with HCl and H2S by passing HCl
gas through it for 0.5 h followed by H2S gas for 0.5 h at -6 °C.
6d (15.0 g, 105 mmol) was added, and H2S gas was passed
through the solution for an additional 15 h, keeping the
temperature at 0 °C. Evaporation and CC [eluent: toluene-
1
EtOAc (2:1)] resulted in 7d (3.3 g, 20%): mp 120-121 °C. H
NMR (DMSO-d6): δ 10.90 (s, 1H), 7.77 (s, 1H), 7.56 (s, 1H),
5.99 (s, 1H), 1.16 (s, 9H). Anal. (C7H13NOS) H, N, S; C: calcd,
52.80; found, 53.27.
5-ter t-Bu tyl-3-isoth ia zolol (8d ). A solution of I2 (6.1 g,
24 mmol) in EtOH (35 mL) was added dropwise to a mixture
of 7d (2.9 g, 18 mmol) and K2CO3 (9.6 g, 69 mmol) in EtOH
(35 mL) with ice cooling. After stirring for 48 h at room
temperature, H2O (100 mL) was added and pH was adjusted
to 3 using 1 M H2SO4. Extraction of the aqueous phase with
Et2O (4 × 200 mL), drying of the combined organic extracts,
evaporation, CC [eluent: toluene-EtOAc (2:1), 1% AcOH], and
recrystallization (toluene) gave 8d (2.2 g, 77%): mp 125-127
°C. 1H NMR (CDCl3): δ 10.21 (s, 1H), 6.31 (s, 1H), 1.36 (s,
9H). Anal. (C7H11NOS) C, H, N, S.
4-(Br om om et h yl)-2-(m et h oxym et h yl)-5-m et h ylisot h i-
a zolin -3-on e (9c). 5-Methyl-3-isothiazolol (8c)37 (6.0 g, 52
mmol) and 1,3,5-trioxane (7.0 g, 78 mmol) were stirred in 62%
aqueous HBr (60 mL) at 60 °C for 3 h. The mixture was
extracted with CH2Cl2 (5 × 50 mL). MeOH (75 mL) was added
to the combined organic extracts, and the mixture was stirred
for 1 h at room temperature. CH2Cl2 (100 mL) was added,
and the mixture was washed with H2O (3 × 200 mL), dried,
and evaporated. CC [eluent: toluene-EtOAc (1:1)] gave crude
9c (3.5 g, 27%) as a yellow oil. 1H NMR (CDCl3): δ 5.31 (s,
2H), 4.40 (s, 2H), 3.36 (s, 3H), 2.45 (s, 3H). 13C NMR (CDCl3):
δ 168.2 (C-3), 155.1 (C-5), 120.0 (C-4), 74.4 (NCH2O), 56.6
(OCH3), 21.8 (CH2Br), 13.2 (CH3).
4-(Br om om eth yl)-5-ter t-bu tyl-2-(m eth oxym eth yl)isoth i-
a zolin -3-on e (9d ). Compound 9d was synthesized as de-
scribed for 9c using 5-tert-butyl-3-isothiazolol (8d ) (3.0 g, 19
mmol), 1,3,5-trioxane (2.6 g, 28 mmol), and 62% aqueous HBr
(20 mL) at 80 °C for 6.5 h. Purification by CC [eluent:
toluene-EtOAc (2:1)] gave crude 9d (2.2 g, 39%) as a yellow
oil. 1H NMR (CDCl3): δ 5.18 (s, 2H), 4.48 (s, 2H), 3.38 (s, 3H),
1.47 (s, 9H). 13C NMR (CDCl3): δ 179.2 (C-3), 177.0 (C-5),
117.9 (C-4), 74.4 (NCH2O), 56.8 (OCH3), 35.8 (quart C), 29.5
[(CH3)3], 24.0 (CH2Br).
Meth yl 2-Aceta m id o-2-(m eth oxyca r bon yl)-3-[2-(m eth -
oxym eth yl)-5-m eth yl-3-oxoisoth ia zolin -4-yl]p r op ion a te
(10c). A 60% suspension of NaH in mineral oil (560 mg, 14
mmol) was added to a solution of dimethyl acetamidomalonate
(2.7 g, 14 mmol) in dry DMF (35 mL). After stirring for 30
min, a solution of 9c (3.5 g, 14 mmol) in dry DMF (10 mL)
was added, and stirring was continued at room temperature
for 12 h. AcOH (1.3 mL) was added, and the reaction mixture
was evaporated to dryness. The residue was dissolved in
CH2Cl2 (200 mL) and washed with H2O (100 mL). The
aqueous phase was extracted with CH2Cl2 (2 × 100 mL), and
the combined organic extracts were dried and evaporated. CC
[eluent: toluene-EtOAc (1:1)] followed by recrystallization
(Et2O-light petroleum) gave 10c (3.5 g, 70%): mp 129-130
°C. 1H NMR (CDCl3): δ 7.25 (s, 1H), 5.11 (s, 2H), 3.81 (s, 6H),
3.42 (s, 2H), 3.32 (s, 3H), 2.30 (s, 3H), 2.01 (s, 3H). Anal.
(C14H20N2O7S) C, H, N, S.
Meth yl 2-Aceta m id o-2-(m eth oxyca r bon yl)-3-[5-ter t-bu -
t yl-2-(m e t h oxym e t h yl)-3-oxoisot h ia zolin -4-yl]p r op i-
on a te (10d ). Compound 10d was synthesized as described
for 10c using a 60% suspension of NaH in mineral oil (671
mg, 17 mmol), dimethyl acetamidomalonate (2.9 g, 15 mmol),
and 9d (4.5 g, 15 mmol) in dry DMF (100 mL). Purification
by CC [eluent: toluene-EtOAc (1:1)] followed by recrystalli-
zation (Et2O-light petroleum) gave 10d (4.4 g, 71%): mp 120-
122 °C. 1H NMR (CDCl3): δ 7.66 (s, 1H), 5.12 (s, 2H), 3.80 (s,
6H), 3.61 (s, 2H), 3.34 (s, 3H), 1.96 (s, 3H), 1.39 (s, 9H). Anal.
(C17H26N2O7S) C, H, N, S.
5-(Hyd r oxym eth yl)-3-m eth oxyisoth ia zole (14). Sodium
borohydride (642 mg, 18 mmol) was added to a solution of 13
(2.2 g, 13 mmol) in dry THF (70 mL). MeOH (13 mL) was
added dropwise to the refluxing reaction mixture over 30 min.
After cooling to room temperature, 1 M HCl (30 mL) was
added. Extraction with Et2O (3 × 100 mL), drying, evapora-
tion, and CC [eluent: toluene-EtOAc (3:1)] resulted in a
yellow oil. Kugelrohr distillation gave 14 (1.6 g, 87%): bp 150
°C, 1 mmHg.1H NMR (CDCl3): δ 6.42 (s, 1H), 4.83 (s, 2H),
4.15 (s, 1H), 3.95 (s, 3H). Anal. (C5H7NO2S) C, H, S; N: calcd,
9.66; found, 9.14.
5-(Ch lor om eth yl)-3-m eth oxyisoth ia zole (15). Thionyl
chloride (10 mL) was added dropwise to ice-cooled 14 (1.5 g,
10.3 mmol). The mixture was refluxed for 1 h, evaporated,
and poured into H2O (35 mL). The solution was extracted with
CH2Cl2 (2 × 50 mL). CC [eluent: toluene-EtOAc (8:1)]
followed by Kugelrohr distillation gave 15 (1.3 g, 77%): bp 150
°C, 0.5 mmHg. 1H NMR (CDCl3): δ 6.57 (br s, 1H), 4.70 (d,
2H, J ) 0.8 Hz), 3.99 (s, 3H). Anal. (C5H6ClNOS) C, H, N, S;
Cl: calcd, 21.70; found, 22.13.
Meth yl 2-Aceta m id o-2-(m eth oxyca r bon yl)-3-(3-m eth -
oxyisoth ia zol-5-yl)p r op ion a te (16). A 60% suspension of
NaH in mineral oil (1.2 g, 31 mmol) was added to a solution
of dimethyl acetamidomalonate (5.2 g, 28 mmol) in dry DMF
(50 mL). After stirring for 30 min, a solution of 15 (4.5 g, 28
mmol) in dry DMF (25 mL) was added. Stirring at room
temperature was continued for 12 h. The mixture was
acidified with AcOH (2.5 mL) and evaporated to dryness. The

+
+
526 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 4
Matzen et al.
residue was dissolved in CH2Cl2 (150 mL), washed with H2O
(50 mL), dried, and evaporated. CC [eluent: toluene-EtOAc
(2:1)] followed by recrystallization (Et2O-light petroleum) gave
16 (4.4 g, 50%): mp 104-105 °C. 1H NMR (CDCl3): δ 6.80 (s,
1H), 6.26 (s, 1H), 3.96 (s, 3H), 3.89 (s, 2H), 3.80 (s, 6H), 2.09
(s, 3H). Anal. (C12H16N2O6S) C, H, N, S.
Meth yl 2-Aceta m id o-2-(m eth oxyca r bon yl)-3-(4-br om o-
3-m eth oxyisoth ia zol-5-yl)p r op ion a te (17). Bromine (68
mL, 1.3 mol) was added dropwise to 16 (1.5 g, 5 mmol) with
ice cooling. Stirring at room temperature for 6 h, evaporation,
CC [eluent: toluene-EtOAc (4:1)], and recrystallization
(EtOAc-light petroleum) gave 17 (1.2 g, 65%): mp 167-169
°C. 1H NMR (CDCl3): δ 6.78 (s, 1H), 4.05 (s, 3H), 3.96 (s, 2H),
3.82 (s, 6H), 2.10 (s, 3H). Anal. (C12H15BrN2O6S) C, H, Br,
N, S.
(RS)-2-Am in o-3-(3-h yd r oxyisot h ia zol-5-yl)p r op ion ic
Acid Hyd r obr om id e (4b‚HBr ). A solution of 16 (500 mg,
1.6 mmol) in 48% aqueous HBr (30 mL) was refluxed for 10
min and evaporated. The residue was dissolved in H2O (10
mL) and treated with charcoal. Filtration, evaporation, and
recrystallization (EtOH-Et2O) gave 4b‚HBr (105 mg, 25%):
mp 159-166 °C dec. 1H NMR (D2O): δ 6.23 (s, 1H), 3.93 (t,
1H, J ) 5.6 Hz), 3.30 (d, 2H, J ) 5.6 Hz). 13C NMR [D2O-
H2O (1:9), pH ) 13.1]: δ 183.9 (CO2H), 180.8 (C-3), 164.5 (C-
5), 117.2 (C-4), 59.0 (CH), 36.6 (CH2). 13C NMR [D2O-H2O
(1:9), pH ) 0.73]: δ 173.2 (CO2H), 173.5 (C-3), 160.8 (C-5),
116.4 (C-4), 55.5 (CH), 31.5 (CH2). Anal. (C6H8N2O3S‚HBr)
C, H, Br, N, S.
(RS)-2-Am in o-3-(4-b r om o-3-h yd r oxyisot h ia zol-5-yl)-
p r op ion ic Acid Hyd r och lor id e (5b‚HCl). A solution of 17
(1.1 g, 3 mmol) in 4 M HCl (30 mL) was refluxed for 4.5 h and
evaporated. The residue was dissolved in H2O (25 mL) and
extracted with EtOAc (2 × 25 mL). The aqueous phase was
evaporated. Recrystallization (EtOH-Et2O) gave 5b‚HCl (181
mg, 20%): mp 230-234 °C dec. 1H NMR (D2O): δ 4.27 (t, 1H,
J ) 6.0 Hz), 3.55 (dd, 1H, J ) 5.6, 15.0 Hz), 3.48 (dd, 1H, J )
5.6, 15.0 Hz). 13C NMR [D2O-H2O (1:9), pH ) 12.9]: δ 183.6
(CO2H), 176.0 (C-3), 157.9 (C-5), 105.1 (C-4), 58.4 (CH), 37.3
(CH2). 13C NMR [D2O-H2O (1:9), pH ) 1.09]: δ 173.4 (CO2H),
169.5 (C-3), 154.3 (C-5), 104.6 (C-4), 55.0 (CH), 32.1 (CH2).
Anal. (C6H7BrN2O3S‚HCl) C, H, N, S; Br: calcd, 26.32; found,
25.56. Cl: calcd, 11.68; found, 10.80.
NMR Titr a tion s. The following solutions in 0.6 mL of
H2O-D2O (9:1), containing 1 M KCl, were titrated (1) thio-
AMPA (2b) (6 mg, 0.03 mmol) and AMPA (2a )32 (10 mg, 0.06
mmol), (2) thio-ATPA (3b) (10 mg, 0.04 mmol) and ATPA (3a )19
(19 mg, 0.08 mmol), (3) 4b‚HBr (12 mg, 0.05 mmol) and HIBO
(4a )32 (17 mg, 0.09 mmol), and (4) 5b‚HCl (18 mg, 0.07 mmol)
and Br-HIBO (5a )32 (10 mg, 0.04 mmol). The solutions were
alkalized with 1 M NaOH and titrated with 0.5 M HCl
adjusted to an ionic strength of 1 M with KCl. Acidity was
measured directly in the NMR tubes, before and after record-
ing the NMR spectrum, using a glass microelectrode connected
to a Radiometer PHM 83 pH meter. The 13C{1H} NMR spectra
were recorded with a Bruker AMX 400 WB spectrometer,
operating at 100.62 MHz for 13C. Spectral parameters were
adjusted to obtain digital resolution in the frequency domain
of 0.8 Hz/data point. All spectra were recorded at 25.0 °C and
standardized to internal sodium 3-(trimethylsilyl)propane-
sulfonate set to δ 0.0. The nonlinear curve fitting was carried
out with the Ultrafit v.2.11 program (Biosoft, Cambridge, U.K.)
using Levenberg-Marquardt algorithm.
Recep tor Bin d in g Assa ys. Affinity for NMDA, AMPA,
and kainic acid receptors was determined using the ligands
[3H]CPP,45 [3H]AMPA,44 and [3H]kainic acid,46 respectively.
The membrane preparations used in all the receptor binding
experiments were prepared according to the method of Ransom
and Stec.49
In Vitr o Electr op h ysiology. A rat cortical wedge prepa-
ration for determination of excitatory amino acid-evoked
depolarizations described by Harrison and Simmonds50 was
used in a slightly modified version. Wedges (500 µm thick) of
rat brain, containing cerebral cortex and corpus callosum, were
placed through a grease barrier for electrical isolation with
each part in contact with an Ag/AgCl pellet electrode. The
cortex and corpus callosum parts were constantly superfused
with a Mg2+-free (and Ca2+-free for the corpus callosum)
oxygenated Krebs buffer at room temperature. The test
compounds were added to the cortex superfusion medium and
the potential difference between the electrodes recorded on a
chart recorder. Applications of agonists were done for 90 s at
each concentration tested. The sensitivity of agonist effects
to the AMPA receptor antagonist NBQX (5 µM) was tested at
agonist concentrations producing at least 50% of maximal
responses. Under these conditions, all of the recorded agonist
responses were reversibly reduced by at least 70%. In experi-
ments designed to detect antagonist effects of AMPA (2a )
analogs at 1 mM concentrations, the compounds were applied
alone for 90 s followed by coapplication of agonist (2a , 5 µM)
and potential antagonist for another 90 s.
In Vivo Con vu lsa n t Activity. Convulsant activity was
determined in male mice (NMRI/BOM, SPF, Bomholtgaard,
Denmark) weighing 24-26 g as described.27 Thio-AMPA (2b)
was dissolved in saline and given subcutaneously in doses of
1.25, 5, 20, and 40 mg/kg. The animals were observed for 30
min for the presence or absence of clonic/tonic convulsions.
Each dose group consisted of 5 mice. ED50 values were
calculated by log-probit analysis.
Meta bolism Stu d ies. Thio-AMPA (2b) was dissolved in
saline and given subcutaneously at a dose of 20 mg/kg. Urine
was collected from 20 male mice (NMRI/BOM, SPF, Bomholt-
gaard, Denmark), weighing 24-26 g, after 30 min. After
addition of 10 vol % of D2O, urine samples were analyzed
directly by 1H NMR. 1H NMR spectra were obtained with low-
power presaturation for elimination of the water signal. The
presence of thio-AMPA (2b) was proved by the observation of
signals at δ 2.40 (s, CH3) and 2.94-3.07 (CH2), the intensities
of which were increased after addition of 1 mg of authentic
material. Due to strong chemical noise around δ 4.3, the signal
of CH could not be detected.
Ack n ow led gm en t. This work was supported by
grants from the Lundbeck Foundation and the Danish
State Biotechnology Program (1991-1995). We thank
the Alfred Benzon Foundation for a grant for the
purchase of the Bruker AMX 400 WB spectrometer. The
secretarial assistance of Anne Nordly is gratefully
acknowledged.
Su p p or tin g In for m a tion Ava ila ble: Tables listing in-
trinsic chemical shift data for the ionic forms of 2a -5b (8
pages). Ordering information is given on any current mast-
head page.
Refer en ces
(1) Krogsgaard-Larsen, P., Hansen, J . J ., Eds. Excitatory Amino
Acid Receptors. Design of Agonists and Antagonists; Ellis Hor-
wood: Chichester, 1992.
(2) Collingridge, G. L., Watkins, J . C., Eds. The NMDA Receptor;
Oxford University Press: Oxford, 1994.
(3) Wheal, H. V., Thomson, A. M., Eds. Excitatory Amino Acids and
Synaptic Transmission; Academic Press: London, 1995.
(4) Conn, P. J ., Patel, J ., Eds. The Metabotropic Glutamate Recep-
tors; Humana Press: New J ersey, 1994.
(5) Danysz, W.; Parsons, C. G.; Bresink, I.; Quack, G. Glutamate
in CNS disorders. Drug News Perspect. 1995, 8, 261-277.
(6) Kno¨pfel, T.; Kuhn, R.; Allgeier, H. Metabotropic glutamate
receptors: Novel targets for drug development. J . Med. Chem.
1995, 38, 1417-1426.
(7) Deutsch, S. I.; Morihisa, J . M. Glutamatergic abnormalities in
Alzheimer’s disease and
a rationale for clinical trials with
L-glutamate. Clin. Neuropharmacol. 1988, 11, 18-35.
(8) Greenamyre, J . T.; Young, A. B. Excitatory amino acids and
Alzheimer’s disease. Neurobiol. Aging 1989, 10, 593-602.
(9) Bowen, D. M. Treatment of Alzheimer’s disease. Molecular
pathology versus neurotransmitter-based therapy. Br. J . Psy-
chiat. 1990, 157, 327-330.
(10) Madsen, U.; Ebert, B.; Krogsgaard-Larsen, P. Modulation of
AMPA receptor function in relation to glutamatergic abnormali-
ties in Alzheimer’s disease. Biomed. Pharmacother. 1994, 48,
305-311.
(11) Carlsson, M.; Carlsson, A. Interactions between glutamatergic
and monoaminergic systems within the basal ganglia - implica-
tions for schizophrenia and Parkinson’s disease. Trends Neuro-
sci. 1990, 13, 272-276.

+
+
AMPA Receptor Agonists
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 4 527
(12) Ulas, J .; Cotman, C. W. Excitatory amino acid receptors in
schizophrenia. Schizophrenia Bull. 1993, 19, 105-117.
(13) Deutsch, S. I.; Mastropaolo, J .; Schwartz, B. L.; Rosse, R. B.;
Morihisa, J . M. A “glutamatergic hypothesis” of schizophrenia.
Rationale for pharmacotherapy with glycine. Clin. Neuro-
pharmacol. 1989, 12, 1-13.
(14) Ebert, B.; Søby, K. K.; Madsen, U.; Krogsgaard-Larsen, P.
Glutamic acid receptors in schizophrenia and Alzheimer’s
disease: Functional partial agonism and receptor modulation
as potential therapeutic approaches. In Schizophrenia - An
Integrated View; Fog, R., Gerlach, J ., Hemmingsen, R., Eds.;
Munksgaard: Copenhagen, 1995; pp 379-395.
(15) Krogsgaard-Larsen, P.; Ebert, B.; Lund, T. M.; Bra¨uner-Osborne,
H.; Sløk, F. A.; J ohansen, T. N.; Brehm, L.; Madsen, U. Design
of excitatory amino acid receptor agonists, partial agonists and
antagonists: ibotenic acid as a key lead structure. Eur. J . Med.
Chem. 1996, 31, 515-537.
(16) Ebert, B.; Madsen, U.; Lund, T. M.; Lenz, S. M.; Krogsgaard-
Larsen, P. Molecular pharmacology of the AMPA agonist, (S)-
2-amino-3-(3- hydroxy-5-phenyl-4-isoxazolyl)-propionic acid [(S)-
APPA] and the AMPA antagonist, (R)-APPA. Neurochem. Int.
1994, 24, 507-515.
(17) Ebert, B.; Lenz, S. M.; Brehm, L.; Bregnedal, P.; Hansen, J . J .;
Frederiksen, K.; Bøgesø, K. P.; Krogsgaard-Larsen, P. Resolu-
tion, absolute stereochemistry, and pharmacology of the S-(+)-
and R-(-)-isomers of the apparent partial AMPA receptor agonist
(R,S)-2-amino-3-(3-hydroxy-5-phenylisoxazol-4-yl)propionic acid
[(R,S)-APPA]. J . Med. Chem. 1994, 37, 878-884.
(18) Ebert, B.; Madsen, U.; Søby, K. K.; Krogsgaard-Larsen, P.
Functional partial agonism at ionotropic excitatory amino acid
receptors. Neurochem. Int. 1996, 29, 309-316.
(19) Lauridsen, J .; Honore´, T.; Krogsgaard-Larsen, P. Ibotenic acid
analogues. Synthesis, molecular flexibility, and in vitro activity
of agonists and antagonists at central glutamic acid receptors.
J . Med. Chem. 1985, 28, 668-672.
(31) Krogsgaard-Larsen, P.; Hjeds, H.; Curtis, D. R.; Lodge, D.;
J ohnston, G. A. R. Dihydromuscimol, thiomuscimol and related
heterocyclic compounds as GABA analogues. J . Neurochem.
1979, 32, 1717-1724.
(32) Hansen, J . J .; Krogsgaard-Larsen, P. Isoxazole amino-acids as
glutamic acid agonists. Synthesis of some analogs and homologs
of ibotenic acid. J . Chem. Soc., Perkin Trans. 1 1980, 1826-
1833.
(33) Krogsgaard-Larsen, P.; Honore, T.; Hansen, J . J .; Curtis, D. R.;
Lodge, D. New class of glutamate agonist structurally related
to ibotenic acid. Nature (London) 1980, 284, 64-66.
(34) Krogsgaard-Larsen, P.; Hansen, J . J .; Lauridsen, J .; Peet, M.
J .; Leah, J . D.; Curtis, D. R. Glutamic acid agonists, stereo-
chemical and conformational studies of DL-R-amino-3-hydroxy-
5-methyl-4-isoxazolepropionic acid (AMPA) and related com-
pounds. Neurosci. Lett. 1982, 31, 313-317.
(35) Hansen, J . J .; Nielsen, B.; Krogsgaard-Larsen, P.; Brehm, L.;
Nielsen, E. O.; Curtis, D. R. Excitatory amino acid agonists.
Enzymic resolution, X-ray structure and enantioselective activi-
ties of (R)- and (S)- bromohomoibotenic acid. J . Med. Chem. 1989,
32, 2252-2260.
(36) Bischoff, F.; J ohansen, T. N.; Ebert, B.; Krogsgaard-Larsen, P.;
Madsen, U. Excitatory amino acid receptor ligands: Asymmetric
synthesis, absolute stereochemistry and pharmacology of (R)-
and (S)-homoibotenic acid. Bioorg. Med. Chem. 1995, 3, 553-
558.
(37) Goerdeler, J .; Mittler, W. Synthese von 3-Hydroxy-, 3-Alkoxy-
und 3-Amino-isothiazolen. (Synthesis of 3-Hydroxy-, 3-Alkoxy-
and 3-Amino-isothiazoles.) Chem. Ber. 1963, 96, 944-954.
(38) Begtrup, M.; Sløk, F. A. Equilibrium control in bromomethyl-
ation: an expedient route to 2-amino-3-(3-hydroxy-5-methyl-
isoxazol-4-yl)propionic acid (AMPA). Synthesis 1993, 861-863.
(39) Lykkeberg, J .; Krogsgaard-Larsen, P. Structural analogues of
GABA. Synthesis of 5-aminomethyl-3-isothiazolol (Thiomusci-
mol). Acta Chem. Scand. 1976, B30, 781-785.
(40) Soai, K.; Oyamada, H. A chemoselective one-step reduction of
â-ketoesters to 1,3-diols. Synthesis 1984, 605-607.
(41) Freedman, M. H.; Lyerla, J . R.; Chaiken, I. N.; Cohen, J . S.
Carbon-13 nuclear magnetic resonance studies on selected amino
acids, peptides, and proteins. Eur. J . Biochem. 1973, 32, 215-
226.
(42) Frydenvang, K.; Matzen, L.; J aroszewski, J . W.; Norrby, P.-O.;
Liljefors, T.; Krogsgaard-Larsen, P. Structural characteristics
of the moieties of 3-isoxazolol and 3-isothiazolol. J . Chem. Soc.,
Perkin Trans. II 1996, submitted.
(20) Krogsgaard-Larsen, P.; Brehm, L.; J ohansen, J . S.; Vinzents,
P.; Lauridsen, J .; Curtis, D. R. Synthesis and structure-activity
studies on excitatory amino acids structurally related to ibotenic
acid. J . Med. Chem. 1985, 28, 673-679.
(21) Ebert, B.; Madsen, U.; Lund, T. M.; Holm, T.; Krogsgaard-
Larsen, P. Molecular pharmacology of cortical and spinal AMPA
receptors. Mol. Neuropharmacol. 1992, 2, 47-49.
(22) Madsen, U.; Frølund, B.; Lund, T. M.; Ebert, B.; Krogsgaard-
Larsen, P. Design, synthesis and pharmacology of model com-
pounds for indirect elucidation of the topography of AMPA
receptor sites. Eur. J . Med. Chem. 1993, 28, 791-800.
(23) Christensen, I. T.; Ebert, B.; Madsen, U.; Nielsen, B.; Brehm,
L.; Krogsgaard-Larsen, P. Excitatory amino acid receptor ligands.
Synthesis and biological activity of 3-isoxazolol amino acids
structurally related to homoibotenic acid. J . Med. Chem. 1992,
35, 3512-3519.
(43) Quirt, A. R.; Lyerla, J . R.; Peat, I. R.; Cohen, J . S.; Reynolds, W.
F.; Freedman, M. H. Carbon-13 nuclear magnetic resonance
titration shifts in amino acids. J . Am. Chem. Soc. 1974, 23, 570-
574.
(24) Skjærbæk, N.; Ebert, B.; Falch, E.; Brehm, L.; Krogsgaard-
Larsen, P. Excitatory amino acids. Synthesis of (RS)-2-amino-
3-(5-cyclopropyl-3-hydroxyisoxazol-4-yl)propionic acid,
potent and specific AMPA receptor agonist. J . Chem. Soc., Perkin
Trans. I 1995, 221-225.
a new
(44) Honore´, T.; Nielsen, M. Complex structure of quiqualate-
sensitive glutamate receptors in rat cortex. Neurosci. Lett. 1985,
54, 27-32.
(45) Murphy, D. E.; Schneider, J .; Boehm, C.; Lehmann, J .; Williams,
K. Binding of [3H]3-(2-carboxypiperazin-4-yl)propyl-1-phosphonic
acid to rat membranes. A selective high affinity ligand for
N-methyl-D-aspartate receptors. J . Pharmacol. Exp. Ther. 1987,
240, 778-783.
(25) Sløk, F. A.; Ebert, B.; Lang, Y.; Krogsgaard-Larsen, P.; Lenz, S.
M.; Madsen, U. Excitatory amino acid receptor agonists. Syn-
thesis and pharmacology of analogues of 2-amino-3-(3-hydroxy-
5-methylisoxazol-4-yl)propionic acid (AMPA). Eur. J . Med. Chem.
1996, in press.
(26) Turski, L.; J acobsen, P.; Honore´, T.; Stephens, D. N. Relief of
experimental spasticity and anxiolytic/anticonvulsant actions of
the R-amino-3-hydroxy-5-methyl-4-isoxazolepropionate antago-
nist 2,3-dihydroxy-6-nitro-7-sulfamoylbenzo(F)quinoxaline. J .
Pharmacol. Exp. Ther. 1992, 260, 742-747.
(27) Arnt, J .; Sa´nchez, C.; Lenz, S. M.; Madsen, U.; Krogsgaard-
Larsen, P. Differentiation of in vivo effects of AMPA and NMDA
receptor ligands using drug discrimination methods and con-
vulsant anticonvulsant activity. Eur. J . Pharmacol. 1995, 285,
289-297.
(28) Krogsgaard-Larsen, P. γ-Aminobutyric acid agonists, antago-
nists, and uptake inhibitors. Design and therapeutic aspects. J .
Med. Chem. 1981, 24, 1377-1383.
(46) Braitman, D. J .; Coyle, J . T. Inhibition of [3H]kainic acid receptor
binding by divalent cations correlates with ion affinity for the
calcium channel. Neuropharmacology 1987, 26, 1247-1251.
(47) Sheardown, M. J .; Nielsen, E. Ø.; Hansen, A. J .; J acobsen, P.;
Honore´, T. 2,3-Dihydroxy-6-nitro-7-sulfamoyl-benzo(F)quinox-
aline. A neuroprotectant for cerebral ischemia. Science 1990,
247, 571-574.
(48) Swamer, F. W.; Hauser, C. R. Claisen acylations and carb-
ethoxylations of ketones and esters by means of sodium hydride.
J . Am. Chem. Soc. 1950, 72, 1352-1356.
(49) Ransom, R. W.; Stec, N. L. Cooperative modulation of [3H]MK-
801 to the N-methyl-D-aspartate receptor ion channel complex
by L-glutamate, glycine and polyamines. J . Neurochem. 1988,
51, 830-836.
(29) Krogsgaard-Larsen, P.; Mikkelsen, H.; J acobsen, P.; Falch, E.;
Curtis, D. R.; Peet, M. J .; Leah, J . D. 4,5,6,7-Tetrahydroiso-
thiazolo[5,4-c]pyridin-3-ol and related analogues of THIP. Syn-
thesis and biological activity. J . Med. Chem. 1983, 26, 895-900.
(30) Krogsgaard-Larsen, P.; Frølund, B.; J ørgensen, F. S.; Schousboe,
(50) Harrison, N. L.; Simmonds, M. A. Quantitative studies on some
antagonists of N-methyl-D-aspartate in slices of rat cerebral
cortex. Br. J . Pharmacol. 1985, 84, 381-391.
receptor agonists, partial agonists and antagonists.
A. GABAA
Design and therapeutic prospects. J . Med. Chem. 1994, 37,
2489-2505.
J M9607212