Some synthetic approaches to glutamate AMPA receptor agonists based on isoxazolones

Article abstract of DOI:10.1071/CH04041

Several approaches to the synthesis of derivatives of the antifungal antibiotic TAN-950A, which is also an agonist of glutamate at hippocampal neurons, are reported. Additions of isoxazolon-4-yl anions to methyleneoxazolidinones were not useful because ad

Full text of DOI:10.1071/CH04041

Rapid Communication
CSIRO PUBLISHING
www.publish.csiro.au/journals/ajc
Aust. J. Chem. 2004, 57, 685–688
Some Synthetic Approaches to Glutamate AMPA Receptor Agonists
Based on Isoxazolones∗
Matthew Cox,A Saba Jahangiri,A Michael V. Perkins,A and Rolf H. PragerA,B
A School of Chemistry, Physics and Earth Sciences, Flinders University of South Australia, Adelaide
SA 5001, Australia.
B Author to whom correspondence should be addressed (e-mail: rolf.prager@flinders.edu.au).
Several approaches to the synthesis of derivatives of the antifungal antibioticTAN-950A, which is also an agonist of
glutamate at hippocampal neurons, are reported. Additions of isoxazolon-4-yl anions to methyleneoxazolidinones
were not useful because addition occurred predominantly through N-2. Similarly addition of the isoxazolon-4-yl
radicals to model Michael acceptors occurred predominantly through N-2. Racemic analogues of TAN-950A were
prepared by reaction of isoxazolone Mannich bases with acetylaminomalonate or addition of β-ketoester anions
to dehydroalanines. The best approach to enantiomerically pure analogues was by acylation of pyroglutamates,
followed by reaction with hydroxylamine.
Manuscript received: 17 February 2004.
Final version: 20 April 2004.
(S)-Glutamic acid 1 (Scheme 1) is the predominant excitatory
neurotransmitter in the mammalian central nervous system
(CNS). Excitatory amino acid receptors are believed to play
a variety of physiological roles in mammals, ranging from
the processing of sensory information, to coordinated move-
ment patterns, cognitive processes, learning, and memory.[1]
Changes in glutamate transmission have been associated with
several neurological disorders including acute neurodegener-
ation, pain, epilepsy, and possibly chronic neurodegenerative
diseases such as Alzeimer’s, Parkinson’s, and Huntington’s
diseases, depression, drug dependency, and schizophrenia.[1]
Glutamate combines with two types of CNS receptors.
These are ionotropic receptors (iGluRs), which open an asso-
ciated ion channel, and metabotropic receptors (mGluRs),
which have seven transmembrane spanning regions and
modulate secondary messenger systems.
Nonetheless, several agonists of the AMPA receptor, such as
ACPA 3[5] and heterocyclic derivatives ofAMPA,[6] or antag-
onists, such as ACMP 4[7] and AMOA 5,[8] are known. Some
structure–activity studies have been carried out on the mam-
malian glutamate receptors,[1,9] which have helped to elu-
cidate the molecular structural requirements for competitive
agonist–antagonist interaction.Three structural requirements
of the amino acid receptor site were suggested to include the
presence of one basic and two acidic groups, the basic amino
group to be situated α to a carboxyl group and the second
acidic group to be situated at the end of a chain of three to six
atoms separating the two acidic groups. The ω-acidic group
was capable of wide structural variation.
TAN-950A (6, R = H), structurally related to AMPA, is
an antifungal antibiotic that was isolated in 1991 from the
culture filtrate of Streptomyces platensis A-136.[10] TAN-
950A has a high affinity for glutamate receptors of the CNS
and exhibited agonistic activity for hippocampal neurons.
Tamura and coworkers[11] have investigated the structure–
activity relationship of TAN-950A analogues for glutamate
receptors. Derivatives of 6 (R = methyl, ethyl, cyclopropyl,
and cyclopentyl) showed high affinity towards the AMPA
We are interested in one of the three subtypes of
ionotropic receptors of (S)-glutamic acid, the receptors acti-
vated by AMPA (2-amino-3-(3-hydroxy-5-methylisoxazol-
4-yl)propionate) 2.[2] Although extensive research has
been carried out on the other sub-receptor sites,[3,4] the
AMPA receptor subtype has received much less attention.
H
N
O
CH3
CH3
R
CH3
CH3
HO
O
O
O
O
N
N
N
N
O
HO2CH2CH2C
HO
HO2C
HO2CH2CO
H
O
H2N
H2N
ACMP 4
Scheme 1.
H2N
CO2H
CO2H
H2N
H2N
CO2H
H2N
CO2H
CO2H
CO2H
1
AMPA 2
ACPA 3
AMOA 5
TAN-950A 6
∗ The authors of this paper congratulate Lew Mander, a friend and colleague, on his 65th birthday.
© CSIRO 2004
10.1071/CH04041
0004-9425/04/070685

686
M. Cox et al.
receptor; in particular the analogues containing cycloalkyl
groups were up to one thousand times more active than (S)-
glutamic acid, and bound more strongly than AMPA. The
analogues containing sterically hindered groups at C-3 (iso-
propyl) or an aromatic ring (phenyl) had lower binding values
at the AMPA receptor.
to that at C-2 (Scheme 3). However, the degree of diastereo-
selectivity depends on the nature of the radical and the N-acyl
group.[14]
In light of Scheme 3, it was hoped that treatment of 8
with a radical generated, initially, from the 3-substituted 4-
bromoisoxazolone 9[17] would lead to the desired (S)-amino
acids. Unfortunately, all trial radical addition reactions with
9 were not useful in the present context. Treatment of 9 with
tributyltin hydride and diethyl ethoxymethylenemalonate,
using photochemical initiation, led to the isolation of diethyl
aminomethylenemalonate 10 which we suggest is formed
as shown in Scheme 3. This indicates that the desired is-
oxazolone radical added to the substrate through N-2 rather
than C-4, giving the adduct 11 rather than the desired 12.
Subsequent photochemical decomposition of 11 would lead
to a carbene[18] and ultimately 10. When the radical was
generated under thermal conditions, product 11 was formed
(Scheme 4).A similar addition under thermal conditions with
ethyl propiolate gave 13,[19] which hydrolyzed on isolation
to 14 (Scheme 5).
Since the acidity of an isoxazol-5(2H)-one is compara-
ble to that of a carboxyl group, and capable of modulation
by substituents at C-3, we wished to further extend the ana-
logues of 6, in particular C-3 aryl-substituted isoxazolone and
pyrazolone amino acids. Variation of the C-3 aryl substituent
was chosen to increase the hydrophobicity of the compound,
thereby facilitating transmembrane permiability. In addition,
aryloxy, arylamino, or arylthio analogues having substituents
capable of hydrogen-bonding were also included, to inves-
tigate whether secondary intermolecular interactions could
enhance the affinity of the analogues towards the receptors.
Methyleneoxazolidinone 7 has been shown to undergo
conjugate additions both with nucleophiles[12,13] and
radicals[14] with high diastereoselectivity (Scheme 2), pro-
viding a facile route to enantiomerically pure amino acids.
The synthesis of the required 4-methyleneoxazolidin-5-
ones has been developed by Seebach,[15] and modified by
Beckwith and Pyne.[16] Beckwith[14] showed that radical
additiontothemethyleneoxazolidinone8oftheenantiomeric
series gave a product with sterechemistry predominantly anti
Attempts to modify the electronic nature of the radical
species were abandoned when it was found that reac-
tions with the N-acetylbromoisoxazolone or N-methylbro-
moisoxazolone (15 or 16, Scheme 6) were unsuccessful; no
radical formation was observed with the former and only
reduction with the latter.
CO2Et
O
O
O
O
R
O
R
N
N
ꢀ
4
Ph
Ph
R
CO32ꢀ, H2O
O
O
H
Bu3SnH
CO2Et
2
N
N
H
OH
H
N
NH2
Ph
RCO
RCO
O
H
H
But
But
O
O
Br
7
O
9
13
14
Scheme 2.
Scheme 5.
O
O
O
COCH3
CH3
R
R
N
N
MH
Ph
Ph
R
O
O
O
O
H
O
N
N
N
PhCO
PhCO
PhCO
Ph
Ph
Ph
H
H
H
O
Br
O
Br
8
15
16
Scheme 3.
Scheme 6.
H
CO2Et
Ph
CO2Et
CO2Et
N
H
N
N
N
Ph
Ph
Ph
O
O
O
O
Bu3SnH
EtO
CO2Et
Ph
H
N
O
O
O
O
EtO2C
Br
O
CO2Et
O
9
11
12
hn
CO2Et
CO2Et
CO2Et
CO2Et
Ph
H
N
H2N
10
Scheme 4.

Synthetic Approaches to Glutamate AMPA Receptor Agonists
687
Molecular orbital calculations (AM1, Macspartan), com-
paring the electron distribution in the highest occupied
molecular orbitals of the radicals from diethyl malonate and
the 3-phenylisoxazolone were informative. In the former, as
expected,[20] almost all the electron spin is localized on car-
bon, but for the isoxazolone radical more electron spin is
localized on nitrogen than on carbon, in accord with our
experimental observations.
While conjugated additions to the chiral oxazolidinones
7 and 8 have been reported,[12,13] alkylation of the read-
ily formed anions of 4-unsubstituted isoxazolones has been
problematic, leading to mixtures of 2- and 4-substituted
products,[19,21,22] and, indeed, attempted reaction with 8 was
not encouraging. This approach was also abandoned.
Boc, Tr; R2 = But, Me) were investigated, but reactions
were complicated by predominant elimination to the dehy-
droalanine product or formation of the aziridine (Scheme 8).
Racemic products 26 were again available through conju-
gate addition of enolates of β-keto esters to the protected
dehydroalanine 25 followed by acid hydrolysis (Scheme 9,
R = Me, Ph, 4-ClC6H4, 4-MeOC6H4).
We found that the alkylation of the benzamidoglutamate
ester 27 with aldehydes gave the chiral hydroxy compounds
28, which could be oxidized by Jones’reagent to the ketones,
whereupon reaction with hydroxylamine gave the required
benzamides 29 (Scheme 10, R =Ar). However, we were
unable to hydrolyze the benzamide without considerable
decomposition.
A synthesis of one set of the desired compounds 6 in
racemic form was achieved by use of the Mannich reac-
tion. While treatment of 14 with formaldehyde and piperidine
has been reported to give the internal salt 17,[23] we found
that treatment of 14 with formaldehyde and dimethylamine
hydrochloride gave the ether 18, which reacted with the
sodium salts of diethyl malonate and diethyl acetylaminoma-
lonate to give the addition compounds 19 and 20 respectively.
Hydrolysis of the ester 20 gave the desired amino acid 21.The
corresponding aryl amino acids 22 and 23 were prepared by
the same route (Scheme 7). However, the observation that the
(R)- and (S)-enantiomers of similar analogues have different
glutamate binding sites, limits the use of racemates.[11]
In an attempt to generate enantiomerically pure com-
pounds 21–23, several substitution reactions on variously
protected serine derivatives (24, X = Hal, OMs; R1 =Ac,
As observed by Tamura,[11] the alkylation or acylation of
pyroglutamate esters proved to be the best general approach.
While alkylation of the Boc ester 30 with CS2 and MeI readily
gave the dithioacetal 31, we were unable to successfully treat
this compound with hydroxylamine or hydrazine to give the
isoxazolone or pyrazolone (Scheme 11). Acylation of 30
with chlorophenyl chloroformate at low temperature gave the
thionoester 32, which unexpectedly[24] gave the thionoiso-
xazolone 33 on treatment with hydroxylamine. Methylation
of this compound occurred readily to give 34. Acylation of
30 with aromatic acid chlorides gave a mixture of diastereo-
meric ketones 35, which gave the required isoxazolones 36
or pyrazolones 37, which could be hydrolyzed to the desired
(S)-amino acids (Scheme 12). Likewise, the use of aryl thio-
cyanates led to the arylamino analogues 38, which were
hydrolyzed to the corresponding amino acids by sequential
reaction with hydroxide followed by trifluoroacetic acid.
We are currently investigating the biological activity of the
isoxazolones described herein, both as glutamate analogues
and as allosteric modulators of GABAB receptors.[25]
In conclusion, of the approaches investigated, those
involving acylation of pyroglutamate esters appear to provide
the most general approach to enantiomerically pure 2-amino-
3-(5-oxo-2,5-dihydroisoxazol-4-yl) propionic acids.
N
Ph
O
Ph Ph
N
N
ꢀ
O
O
O
H
Nꢁ
H
H
O
O
CH2OCH2
17
18
X
H
N
H
N
Accessory Materials
Ph
O
O
Experimental details for the preparation of compounds
21–23, 36, 38 (Ar = 2-naphth), and hydrolysis of 36
O
O
HO2C
EtO2C
NH2
CO2Et
X
X
H
R1HN
19 X ꢂ H
20 X ꢂ NHAc
21 X ꢂ H
22 X ꢂ Cl
ꢁ
R1HN
CO2R2
R1NH
CO2R2
CO2R2
23 X ꢂ OMe
24
Scheme 7.
Scheme 8.
H
N
1. NH2OH
2. H3Oꢁ
O
R
R
O
EtO
ꢀ
RCOCHCO2Et
ꢁ
AcHN
CO2But
O
O
25
AcHN
HO2C
CO2But
NH2
H
H
26 R ꢂ Me, Ph, 4-ClC6H4, 4-MeOC6H4
Scheme 9.

688
M. Cox et al.
H
N
HO
R
R
O
1. LHMDS
2. RCHO
1. Jones’
2. NH2OH
MeO2C
BzHN
MeO2C
O
CO2Me
BzHN
CO2Me
BzHN
CO2Me
27
28
29
Scheme 10.
H
N
Me
N
SMe
OC6H4Cl
S
S
SMe
O
O
MeS
H
H
H
O
O
O
N
Boc
CO2Me
O
N
Boc
CO2Me
O
N
Boc
CO2Me
BocHN
BocHN
CO2Me
CO2Me
30
31
32
33
34
Scheme 11.
H
N
H
N
H
N
Ar
O
Ar
Ar
NHAr
O
HN
O
O
H
O
O
O
N
Boc
CO2Me
BocHN
BocHN
BocHN
CO2Me
CO2Me
CO2Me
35
36
37
38
Ar ꢂ Ph, 4-ClC6H4, 4-MeOC6H4, 1-naphth, 2-naphth
Ar ꢂ Ph, 2-ClC6H4,
3-MeOC6H4, 2-naphth
Scheme 12.
(Ar = 2-naphth) to the acid are available from the author or,
until July 2009, the Australian Journal of Chemistry.
[11] N. Tamura, Y. Matsushita, T. Iwama, S. Harada, S. Kishimoto,
K. Itoh, Chem. Pharm. Bull. Jpn. 1991, 39, 1199.
[12] S. G. Pyne, A. Javidan, B. W. Skelton, A. H. White, Tetrahedron
1995, 51, 5157. doi:10.1016/0040-4020(95)98711-P
[13] S. G. Pyne, B. Dikic, P. A. Gordon, B. W. Skelton, A. H. White,
Aust. J. Chem. 1993, 46, 73.
Acknowledgments
S.J. thanks Flinders University for anAustralian Postgraduate
Award.
[14] J. R. Axon, A. L. J. Beckwith, J. Chem. Soc., Chem. Commun.
1995, 549. doi:10.1039/C39950000549
[15] D. Seebach, A. Fadel, Helv. Chim. Acta 1985, 68, 1243.
[16] A. L. Beckwith, S. G. Pyne, B. Dikic, C. L. Chai, P. A. Gordon,
B. W. Skelton, M. J. Tozer, A. H. White, Aust. J. Chem. 1993,
46, 1425.
[17] A. J. Boulton, A. R. Katritzky, Tetrahedron 1961, 12, 41.
doi:10.1016/0040-4020(61)80097-5
References
[1] J. C. Watkins, P. Krogsgaard-Larsen, T. Honore, Trends Phar-
macol. Sci. 1990, 11, 25. doi:10.1016/0165-6147(90)90038-A
[2] P. Krogsgaard-Larsen, T. Honore, J. J. Hansen, D. R. Curtis,
D. Lodge, Nature 1980, 284, 64.
[18] R. H. Prager, D. S. Millan, in Advances in Nitrogen Heterocycles
4 (Ed. C. J. Moody) 2000 (Elsevier: Amsterdam).
[19] M. Cox, F. Heidarizadeh, R. H. Prager, Aust. J. Chem. 2000,
53, 665. doi:10.1071/CH00098
[3] A. Foster, G. Fagg, Brain Res. Rev. 1984, 7, 103. doi:10.1016/
0165-0173(84)90020-1
[4] H. McLennon, Progr. Neurobiol 1983, 20, 251. doi:10.1016/
0301-0082(83)90004-7
[20] S. Oae, Y. Kadoma, Can. J. Chem. 1986, 64, 1184.
[21] C. H. DePuy, E. W. Della, J. Filley, J. J. Grabowski,
V. M. Bierbaum, J. Am. Chem. Soc. 1983, 105, 2481.
[22] Y. Makisumi, T. Sasatani, Tetrahedron Lett. 1969, 10, 543.
doi:10.1016/S0040-4039(01)87743-1
[5] U. Madsen, E. H. F. Wong, J. Med. Chem. 1992, 35, 107.
[6] J. Valgeirsson, J. K. Christensen, A. S. Kristensen,
D. S. Pickering, B. Nielsen, C. H. Fischer, H. Brauner-Osborne,
E. O. Nielsen, et al., Bioorg. Med. Chem. 2003, 11, 4341.
doi:10.1016/S0968-0896(03)00485-1
[23] W. Barbrieri, L. Bernadi, S. Coda, S. Colo, G. Diqual,
G. Palamidessi, Tetrahedron 1967, 23, 4409. doi:10.1016/S0040-
4020(01)88838-4
[24] (a) D. E. Worral, J. Am. Chem. Soc. 1923, 45, 3092.
(b) D. E. Worral, J. Am. Chem. Soc. 1918, 40, 415.
[25] D. I. B. Kerr, J. Ong, N. M. Puspawati, R. H. Prager, Eur. J.
Pharmacol. 2002, 451, 69. doi:10.1016/S0014-2999(02)02195-7
[7] U. Madsen, F. A. Slok, T. B. Stensbol, H. Brauner-
Osborne, H.-C. H. Lutzhoft, M. V. Poulsen, L. Eriksen,
P. Krogsgaard-Larsen, Eur. J. Med. Chem. 2000, 35, 69.
doi:10.1016/S0223-5234(00)00104-5
[8] P. Krogsgaard-Larsen, J. W. Ferkany, E. O. Nielsen, U. Madsen,
B. Ebert, J. S. Johansen, J. Med. Chem. 1991, 34, 123.
[9] L. Brehm, F. S. Jorgensen, J. J. Hansen, P. Krogsgaard-Larsen,
Drug News Perspect. 1988, 1, 138.
[10] T. Iwama, Y. Nagai, N. Tamura, S. Harada, A. Nagaoka,
Eur. J. Pharmacol. 1991, 197, 187. doi:10.1016/0014-2999(91)
90520-Z