First enantiospecific synthesis of a 3,4-dihydroxy-L-glutamic acid [(3S,4S)-DHGA], a new mGluR1 agonist

Article abstract of DOI:10.1016/S0960-894X(99)00641-1

The first synthesis of one of the 4 possible stereoisomers of 3,4- dihydroxy-L-glutamic acid ((3S,4S)-DHGA 3), a natural product of unknown configuration, is described. The synthesis is based on the Lewis acid catalyzed reaction of benzyl alcohol with a D

Full text of DOI:10.1016/S0960-894X(99)00641-1

Bioorganic & Medicinal Chemistry Letters 10 (2000) 129±133
First Enantiospeci®c Synthesis of a 3,4-Dihydroxy-L-glutamic
Acid [(3S,4S)-DHGA], a New mGluR1 Agonist
Philippe Dauban, a Carole de Saint-Fuscien, a Francine Acher, b Laurent Prezeau, c
Isabelle Brabet, c Jean-Philippe Pin c and Robert H. Dodd a,*
aInstitut de Chimie des Substances Naturelles, CNRS, 91198 Gif-sur-Yvette, France
b
Â
Laboratoire de Chimie et Biochimie Pharmacologique et Toxicologique (UMR 8601 CNRS), Universite Rene Descartes-Paris V, 45,
Â
Á
rue des Saints-Peres, 75270 Paris Cedex 06, France
 Â
CNRS-UPR 9023, Mecanismes Moleculaires des Communications Cellulaires, CCIPE, 141, rue de la Cardonille,
c
34094 Montpellier Cedex 5, France
Received 26 May 1999; accepted 15 November 1999
AbstractÐThe ®rst synthesis of one of the 4 possible stereoisomers of 3,4-dihydroxy-l-glutamic acid ((3S,4S)-DHGA 3), a natural
product of unknown con®guration, is described. The synthesis is based on the Lewis acid catalyzed reaction of benzyl alcohol with
a d-ribose-derived 2,3-aziridino-g-lactone 4-benzyl carboxylate (6). Preliminary pharmacological studies showed that (3S,4S)-3 is
an agonist of metabotropic glutamate receptors of type 1 (mGluR1) and a weak antagonist of mGluR4 but has no discernible
activity with respect to mGluR2. This activity pro®le can be rationalized by ®tting extended conformations of (3S,4S)-3 in proposed
models of each of these receptor subtypes. # 2000 Elsevier Science Ltd. All rights reserved.
Glutamic acid is the major excitatory neurotransmitter
of the central nervous system. It acts through four
major receptor classes termed NMDA, AMPA, kainate
and metabotropic receptors each of which is in turn
comprised of several subclasses.1±5 While the ®rst three
types of receptors, collectively referred to as ionotropic
glutamate receptors (iGluRs), function via cation
(Ca2+, K+, Na+) permeable channels, the metabo-
tropic receptors (mGluRs) are coupled to e€ector sys-
tems (e.g. phospholipase C and phosphoinositide (PI)
hydrolysis; adenylyl cyclase) through GTP-binding
proteins. Eight mGluR subtypes (mGluR1 to mGluR8)
have been isolated and these are classi®ed in three
groups (groups I, II and III) based on their amino acid
sequence homology. Group I, composed of mGluR1
and mGluR5, is positively coupled to phospholipase C
and intracellular production of inositol phosphate via a
Gq G-protein subtype. On the other hand, groups II
(mGluR2 and mGluR3) and III (mGluR4, 6, 7 and 8)
are negatively coupled to adenylyl cyclase via Gi/Go
G-protein subtypes. The mGluRs are currently thera-
peutic targets for the treatment of such neurodegenera-
tive syndromes as Parkinson's and Alzheimer's diseases,
as well as brain ischemia and epilepsy.4,5
In order to delineate the physiological role of each
mGluR subtype, there is presently a great need for spe-
ci®c ligands. For this purpose, a large number of ana-
logues of glutamic acid have been synthesized in recent
years and some have been shown to exhibit such speci-
®city.6±17 Thus, (2S,4S)-4-methylglutamic acid (1) is
an agonist of the mGluR1 and 2 subtypes18±20 while
(2S,4S)-4-(o,o-diphenylalkyl)glutamic acids 2a and 2b
are speci®c mGluR2 antagonists.14,15 Because naturally-
occurring compounds have very often been shown to be
biologically active and therapeutically useful and in an
e€ort to discover potentially selective mGluR ligands,
we recently undertook a literature survey of natural
glutamic acid derivatives. Interestingly, the isolation of
3,4-dihydroxyglutamic acid (DHGA, 3) from the seeds
of Lepidum sativum as well as from the leaves of Rheum
rhaponticum was reported over 40 years ago21 and this
amino acid was subsequently shown to be a constituent
of a large variety of ¯owering plants, ferns, mosses and
mushrooms.22,23 However, nothing is known concerning
the relative and, a fortiori, the absolute con®gurations
of the two chiral centers of this natural compound.
Moreover, until now, neither chiral nor achiral synth-
eses of DHGA, or of 3,4-disubstituted glutamic acid
derivatives in general, have been reported. In order to
evaluate the possible interactions of this class of natural
substances with mGluR subtypes, the synthesis of one
*Corresponding author. Tel.: +33-1-69-82-45-94; fax: +33-1-69-07-
72-47; e-mail: robert.dodd@icsn.cnrs-gif.fr
0960-894X/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(99)00641-1

130
P. Dauban et al. / Bioorg. Med. Chem. Lett. 10 (2000) 129±133
of the four possible isomers of 3 was undertaken utiliz-
ing our 2,3-aziridino-g-lactone approach to the enan-
tiospeci®c synthesis of substituted a- or b-amino
acids.24±29
led to isolation of a single species, identi®ed as the
amino acid 3.32
In preliminary pharmacological studies, the e€ect of
(3S,4S)-DHGA on mGluRs was determined by mea-
suring IP production relative to glutamate itself, on one
representative receptor subtype from each group (i.e.
mGluR1 for group I; mGluR2 for group II; mGluR4
for group III) using previously described methodol-
ogy.33±35 The coupling of mGluR2 and mGluR4 to PLC
was made possible by co-expressing these receptors with
the chimeric G-protein Gqi9, as previously described.33±35
As shown in Figure 1A, (3S,4S)-DHGA (1 mM) sig-
ni®cantly stimulated IP production (IP conversion
ratio=0.29, corresponding to a 4.5-fold increase of the
basal IP production) in cells expressing mGluR1 though
not as strongly as glutamate (IP conversion ratio=0.44
at 1 mM, corresponding to a 6.6-fold increase of the
basal IP production). The e€ect was dose-dependent
with an EC50 of 257‹64 mM (n=3) showing that
(3S,4S)-3 is less potent than Glu (EC50=1.05‹0.20
mM, n=5). In contrast, (3S,4S)-DHGA did not activate
mGluR2 or mGluR4, the measured IP production being
in both cases practically identical to basal values (i.e.
with no ligands or in the presence of the mGluR speci®c
antagonists). (3S,4S)-DHGA is thus an agonist of
mGluR1. In order to verify for receptor-speci®c
antagonist e€ects of (3S,4S)-3, IP production resulting
from simultaneous application of this compound and
glutamate to each receptor subtype was measured (Fig.
1B). As expected for mGluR1, where it acts as an ago-
nist, (3S,4S)-DHGA did not inhibit the IP production
induced by Glu (1 mM). On the other hand, (3S,4S)-3
inhibited the glutamate-induced stimulation of mGluR4
(57.0‹11.9% inhibition, n=3) but had little e€ect
on mGluR2 (29.0‹10.7% inhibition, n=3). (3S,4S)-
DHGA can thus be considered a weak mGluR4
antagonist.
The starting material for our study was the uronic acid
derivative 4 (a,b mixture, Scheme 1) which was pre-
pared from d-ribose as previously described.24 Previous
experience has shown that hydroxyglutamic acid deri-
vatives are prone to degradation when subjected to
acidic or basic conditions. Our objective was thus to
prepare a synthon which would allow complete depro-
tection of the amino acid derivative simply by hydro-
genolysis in the ®nal step. For this reason, compound 4
was ®rst transformed into the benzyl ester 5 in 86%
yield using Steglich conditions30 and then into the 2,3-
aziridino-g-lactone 6 using the general sequence of
reactions we have developed for the synthesis of this
class of compounds.27,31
The conversion of aziridino-g-lactone 6 into (3S,4S)-
DHGA (3S,4S-3) is depicted in Scheme 2. Thus, treat-
ment of a chloroform solution of 6, with a 10-fold
excess of benzyl alcohol and 2 eq of boron tri¯uoride
etherate at 0 ꢀC for 20 h, a€orded an equimolar mixture
of the protected glutamic acid derivative 7 (resulting
from nucleophilic attack of both the aziridine and lac-
tone functionalities by benzyl alcohol) and the lactone 8
(resulting from partial cyclization of 7). While both
these products could be separated by chromatography,
their facile interconversion encouraged us to proceed
directly with hydrogenolytic deprotection of the mix-
1
ture. H NMR spectroscopy of the crude reaction pro-
duct showed, as expected, the presence of both (3S,4S)-
DHGA 3 and its lactonized form. Passage of this mate-
rial through a basic anion exchange column (AG1-X4
resin), using acetic acid (0.2±0.5 M) as the eluent, then
These biological activities of (3S,4S)-3 can be tenta-
tively interpreted on the basis of mGluR1, mGluR2,
mGluR4 pharmacophore models which have been
recently described.20,37 In all three models, glutamic
Scheme 1.
Scheme 2.

P. Dauban et al. / Bioorg. Med. Chem. Lett. 10 (2000) 129±133
131
Figure 1. E€ects of (3S,4S)-DHGA on mGluR1, 2 and 4 transiently expressed in HEK 293 cells: HEK 293 cells expressing mGluR1, mGluR2 (plus
Gqi9) or mGluR4 (plus Gqi9) were stimulated by the indicated compounds and the intracellular activity of the phospholipase C enzyme was mea-
sured by quantifying the production of inositol phosphates (IP). A: (3S,4S)-DHGA is an agonist on mGluR1. HEK 293 cells expressing mGluR1, 2
or 4 are incubated in a medium containing glutamate (Glu, 1 mM), (3S,4S)-DHGA (1 mM) or the appropriate antagonist36 (4CPG, MCCG-I or
MAP4, respectively; 1 mM). Glu activated all three receptors but (3S,4S)-DHGA activated only mGluR1. As expected, none of the antagonists
induced any agonist e€ect. B: (3S,4S)-DHGA is a weak antagonist on mGluR4. HEK 293 cells expressing mGlu R1, 2 or 4 were stimulated by Glu
(1 mM on mGluR1; 20 mM on mGluR2; 30 mM on mGluR4) in the presence or absence of their respective antagonists (4CPG, MCCG-I and MAP4)
or (3S,4S)-DHGA. (3S,4S)-DHGA (1 mM) inhibited Glu-induced IP production to a signi®cant level only on mGluR4. As a control, the Glu-
induced stimulation of the three receptors was inhibited by their respective antagonists (1 mM). Results are expressed as the amount of IP produced
over the radioactivity present in the membranes (IP conversion ratio). ***=P<0.001.
acid, which can adopt numerous conformations, lies in
an extended conformation (aa, g+a, g a)38 and is
bound to the receptor by means of hydrogen bonds or
ionic interactions with the amino and carboxylic func-
tions (S1, S2, S3a,b sites). Selective agonists for each of
the three mGluR groups such as quisqualic acid
(mGluR1), LY354740 (mGluR2), ACPT-I (mGluR4)39
were shown to bind to these sites and to take advantage
of the di€ering protein environments resulting in spe-
ci®c interactions located in the S4, S5 and S6 regions as
depicted in Figure 2.
(3S,4S)-DHGA might result from the additional hy-
droxyl group located in the restricted S5 region.
In the mGluR2 and mGluR4 models, glutamic acid
adopts an aa conformation.20,37 In the analogous
(3S,4S)-DHGA conformation, the 4-hydroxyl group
points to the S6 hydrophobic region located in the
(3S,4S)-DHGA conformations were generated by
molecular dynamics, minimized and clustered using
InsightII/Discover softwares (Molecular Simulations,
San Diego, CA) as previously described.20,37 (3S,4S)-
DHGA was taken in its zwitterionic state so that the
protonation of the distal acidic functions allows all
conformations to be generated with a dielectric constant
of 80 or 5 within 3.5 kcal energy di€erence. Selected
conformers were superimposed to the pharmacophore
models. Interestingly, while S1, S2, S3a,b sites are ®tted
with the amino and carboxylic groups of (3S,4S)-
DHGA, the mGluR1 hydrophilic site in the S4 region is
®tted with the 4-hydroxyl group of the g+a conforma-
tion or the 3-hydroxyl of the g a conformation. The
folded ag+ conformation accommodates the same 5
sites with the 4-hydroxyl at S3a and the distal carbox-
ylate at S3b and S4. Thus, several (3S,4S)-DHGA
conformations with similar energies display binding
analogous to quisqualate, the most potent mGluR1
agonist. However, the apparent weaker anity of
Figure 2. Superposition of the aa, g+a, g a and ag+ (3S,4S)-DHGA
conformations onto quisqualate (magenta), LY354740 (blue), ACPT-I
(orange) in their respective mGluR1, mGluR2, mGluR4 pharmaco-
phore model conformations:20,37 (3S,4S)-DHGA carbon atoms are
colored in green, nitrogen in blue, carboxylic oxygens in red and
hydroxylic oxygens in yellow. Common glutamate binding sites S1, S2,
S3a,b and selective S4, S7 regions are indicated. S4 and S6 are selective
hydrophilic sites of respectively mGluR1 and mGluR4, in contrast S6
is hydrophobic at mGluR2. S5 is a region with a di€erent steric
allowance according to each mGluR. S7 is an unexplored region.

132
P. Dauban et al. / Bioorg. Med. Chem. Lett. 10 (2000) 129±133
foreground region of Figure 2 (lining the LY354740
ring as de®ned previously).20 This situation could be
responsible for the observed lack of activity of (3S,4S)-
DHGA on mGluR2 compared to the strong potency of
(2S,4S)-4-MeGlu.20 In contrast, this S6 region is
hydrophilic in mGluR4 as shown by ACPT-I binding,
so that the antagonist property of (3S,4S)-DHGA on
mGluR4 would be due to the position of the 3-hydroxyl
in the unexplored S7 region. However, as mentioned
above, the g+a and g a conformations of (3S,4S)-
DHGA position a hydroxyl group in the S4 region
which was proposed as a restricted area in mGluR2 and
mGluR4.20,37 It was suggested that a hydrophilic group
in this S4 region might be responsible for an antagonist
property at mGluR4. Thus, all three extended forms of
(3S,4S)-DHGA can account for the observed biological
activities at mGluR2 and mGluR4.
10. Moody, C. M.; Young, D. W. Tetrahedron Lett. 1994, 35,
7277.
11. Kozikowski, A. P.; Steensma, D.; Araldi, G. L.; Tuck-
mantel, W.; Wang, S.; Pshenichkin, S.; Surina, E.; Wro-
blewski, J. T. J. Med. Chem. 1998, 41, 1641.
12. Ouerfelli, O.; Ishida, M.; Shinozaki, H.; Nakanishi, K.;
Ohfune, Y. Synlett 1993, 409.
13. Ornstein, P. L.; Bleisch, T. J.; Arnold, M. B.; Wright, R.
A.; Johnson, B. G.; Schoepp, D. D. J. Med. Chem. 1998, 41,
346.
14. Wermuth, C. G.; Mann, A.; Schoenfelder, A.; Wright, R.
A.; Johnson, B. G.; Burnett, J. P.; Mayne, N. G.; Schoepp, D.
D. J. Med. Chem. 1996, 39, 814.
15. Escribano, A.; Ezquerra, J.; Pedregal, C.; Rubio, A.;
Yruretagoyena, B.; Baker, S. R.; Wright, R. A.; Johnson, B.
G.; Schoepp, D. D. Bioorg. Med. Chem. Lett. 1998, 8, 765.
16. Helaine, V.; Bolte, J. Tetrahedron: Asymmetry 1998, 9,
3855.
17. Pin, J.-P.; De Colle, C.; Bessis, A.-S.; Acher, F. Eur. J.
Pharmacol. 1999, 375, 277.
In conclusion, use of 2,3-aziridino-g-lactone methodol-
ogy has now been successfully applied to the ®rst
enantiospeci®c preparation of a 3,4-disubstituted glu-
tamic acid, namely (3S,4S)-dihydroxy-l-glutamic acid
[(3S,4S)-3]. The selective activity of this compound on
mGluR1 relative to mGluR2 and mGluR4 which can be
rationalized in terms of receptor models, suggests that
the other isomers of this compound may also constitute
a valuable source of speci®c ligands of the metabotropic
glutamate receptors of the central nervous system. This
is presently being investigated.
18. Todeschi, N.; Gharbi-Benarous, J.; Acher, F.; Larue, V.;
Pin, J.-P.; Bockaert, J.; Azerad, R.; Girault, J.-P. Bioorg. Med.
Chem. 1997, 5, 335.
19. Brauner-Osborne, H.; Nielsen, B.; Stensbùl, T. B.; Johan-
sen, T. N.; Skjaerbaek, N.; Krogsgaard-Larsen, P. Eur. J.
Pharmacol. 1997, 335, R1.
20. Jullian, N.; Brabet, I.; Pin, J.-P.; Acher, F. J. Med. Chem.
1999, 42, 1546.
21. Virtanen, A. I.; Ettala, T. Suomen Kemistilehti. 1956, B29,
107.
22. Virtanen, A. I.; Ettala, T. Acta Chem. Scand. 1957, 11,
182.
23. Muller, A. L.; Uusheimo, K. Acta Chem. Scand. 1965, 19,
1987.
24. Dubois, L.; Dodd, R. H. Tetrahedron 1993, 49, 901.
25. Dubois, L.; Mehta, A.; Tourette, E.; Dodd, R. H. J. Org.
Chem. 1994, 59, 434.
26. Dauban, P.; Dubois, L.; Tran Huu Dau, E.; Dodd, R. H.
J. Org. Chem. 1995, 60, 2035.
27. Dauban, P.; Chiaroni, A.; Riche, C.; Dodd, R. H. J. Org.
Chem. 1996, 61, 2488.
28. Dauban, P.; Hofmann, B.; Dodd, R. H. Tetrahedron 1997,
53, 10743.
Acknowledgements
We are grateful for fellowships from the French Minis-
try of Defence (P.D.) and from the Ministry of Higher
Education (C.S.-F.). The pharmacological work and
molecular modeling was supported by grants from the
``Action Incitative Physique et Chimie du Vivant''
(PCV97-115), the European Community Biomed 2
(BMH4-CT96-0228) and Biotech 2 (BI04-CT96-0049)
programs, the Fondation pour la Recherche Medicale
and Bayer Company (France and Germany).
29. Dauban, P.; Dodd, R. H. J. Org. Chem 1997, 62, 4277.
30. Neises, B.; Steglich, W. Angew. Chem., Int. Ed. Engl. 1978,
17, 522.
31. Full details of the experimental procedures and character-
ization of the products have been published elsewhere: Dau-
ban, P.; De Saint-Fuscien, C.; Dodd, R. H. Tetrahedron 1999,
References and Notes
55, 7589.
28
d
32. Compound (3S,4S)-3: ‰aŠ
0.8 (c 1.0, H2O); 1H NMR
1. Krogsgaard-Larsen, P.; Ebert, B.; Lund, T. M.; Brauner-
Osborne, H.; Slok, F. A.; Johansen, T. N.; Brehm, L.; Mad-
sen, U. Eur. J. Med. Chem. 1996, 31, 515.
2. Knopfel, T.; Kuhn, R.; Allgeier, H. J. Med. Chem. 1995, 38,
1417.
3. Nakanishi, S. Science 1992, 258, 597.
4. Conn, P.; Pin, J.-P. Ann. Rev. Pharmacol. Toxicol. 1997, 37,
205.
5. Madge, D. J.; Batchelor, A. M. Ann. Rep. Med. Chem.;
Academic Press, 1996; Chapter 4.
(250 MHz, D2O) d 3.80 (s, 1H), 4.24 (d, 1H, J4,3=3.3 Hz),
4.48 (d, 1H, J=3.2 Hz) ; 13C NMR (75 MHz, D2O) d 56.9,
71.2, 76.2, 173.9, 178.3; mass spectrum (FAB) m/z 180
(MH)+.
33. Gomeza, J.; Mary, S.; Brabet, I.; Parmentier, M.-L.; Res-
tituito, S.; Bockaert, J.; Pin, J.-P. Mol. Pharmacol. 1996, 50,
923.
34. Joly, C.; Gomeza, J.; Brabet, I.; Curry, K.; Bockaert, J.;
Pin, J.-P. J. Neurosci. 1995, 15, 3970.
35. As indicated in the introduction, mGluR2 and mGluR4
are negatively coupled to adenylyl cyclase. Because this trans-
duction cascade is more dicult to measure in transient
transfection assays than the stimulation of IP formation, the
coupling of these receptors to phospholipase C was made
possible by co-expressing them with a chimeric G-protein
alpha subunit, Gqi9 (ref 36).
6. Shimamoto, K.; Ishida, M.; Shinozaki, H.; Ohfune, Y. J.
Org. Chem. 1991, 56, 4167.
7. Yanagida, M.; Hashimoto, K.; Ishida, M.; Shinozaki, H.;
Shirahama, H. Tetrahedron Lett. 1989, 30, 3799.
8. Gu, Z. Q.; Hesson, D. P.; Pelletier, J. C.; Maccechini, M. L.;
Zhou, L. M.; Skolnick, P. J. Med. Chem. 1995, 38, 2518.
9. Ezquerra, J.; Pedregal, C.; Mico, I.; Najera, C. Tetrahedron:
Asymmetry 1994, 5, 921.
36. Watkins, J. C.; Collingridge, G. Trends Pharmacol. Sci.
1994, 15, 333.

P. Dauban et al. / Bioorg. Med. Chem. Lett. 10 (2000) 129±133
133
37. Bessis, A.-S.; Jullian, N.; Coudert, E.; Pin, J.-P.; Acher, F.
Neuropharmacology 1999, in press.
39. Structures of the selective agonists of the three mGluR
groups depicted in Figure 2.
38. The various conformations of l-glutamic acid are described
as follows: