Stereoselective synthesis and preliminary evaluation of new D-3-heteroarylcarbonylalanines as ligands of the NMDA receptor

Article abstract of DOI:10.1016/j.bmcl.2004.06.062

New N-heteroarylcarbonylalanines of the d-series were stereoselectively prepared from enoates derived from d-mannitol. Their activities in binding and functional assays for the NMDA sub-type of glutamate receptors are described. New N-heteroarylcarbonylal

Full text of DOI:10.1016/j.bmcl.2004.06.062

Bioorganic & Medicinal Chemistry Letters 14 (2004) 4399–4403
Stereoselective synthesis and preliminary evaluation of new
D-3-heteroarylcarbonylalanines as ligands of the NMDA receptor
c
c
c
Paulo G. Lima,a,b Rodrigo R. B. Caruso, Simone O. Alves, Renata F. Pessoa,
ˆ
Dayde L. Mendonc¸a-Silva,c Ricardo J. Nunes,d Franc¸ois Noe¨l,c
Newton G. Castroc, and Paulo R. R. Costa
a,
*
*
a
´ ´ ˆ ´
Laboratorio de Quımica Bioorganica, Nucleo de Pesquisas de Produtos Naturais, Centro de Ciencias da Saude, Bloco J,
ˆ
Universidade Federal do Rio de Janeiro, RJ 21941-590, Brazil
´
b
´
ˆ
´
Departamento de Quımica Organica, Instituto de Quımica, Universidade Federal do Rio de Janeiro, RJ 21941-590, Brazil
´
c
´ ´
Departamento de Farmacologia Basica e Clınica, ICB, Centro de Ciencias da Saude, Bloco J, Universidade Federal do Rio de Janeiro,
ˆ
RJ 21941-590, Brazil
d
´
Departamento de Quımica, CCFM, Universidade Federal de Santa Catarina, SC 88040-900 Brazil
Received 11 May 2004; revised 21 June 2004; accepted 21 June 2004
Abstract—New N-heteroarylcarbonylalanines of the D-series were stereoselectively prepared from enoates derived from D-mannitol.
These compounds were active in binding and functional assays of the NMDA sub-type of glutamate receptors. A pyridine derivative
inhibited MK801 binding, protected neurons from excitotoxic damage and blocked NMDA-induced currents in neurons. A thio-
phene derivative positively modulated the NMDA receptor, possibly through the allosteric glycine site.
ꢀ 2004 Elsevier Ltd. All rights reserved.
1. Introduction
(S)-AMPA (2), is the prototype agonist for the AMPA
sub-type of ionotropic GluR while (S)-4C3HPG (3) is
a metabotropic receptor agonist. On the other hand,
the NMDA sub-type of ionotropic GluR generally
exhibits affinity for amino acids bearing the nonnatural
(S)-Glutamic acid is the major excitatory aminoacid neu-
rotransmitter in the vertebrate central nervous system,
where it operates on two types of receptors (GluR): ion-
otropic (ligand-gated ion channels) and metabotropic
(G-protein coupled).1 Glutamatergic neurotransmission
plays an important role in memory, learning, and devel-
opmental plasticity, as well as in pathological processes.2
GluR hypoactivity has been implicated in schizophrenia,
while excessive activity leads to acute and chronic neuro-
degenerative disorders.3
The interest in nonnatural aromatic a-aminoacids aim-
ing at GluR started with the discovery of the strong hal-
lucinogenic effect of ibotenic acid (1), a naturally
occurring excitotoxin that is active as agonist at all
sub-types of GluR. A structurally related compound,
Keywords: D-Aminoacid; NMDA; Glycine; Glutamate.
*
Corresponding authors. Tel.: +55-21-22804694; fax: +55-21-
25626733; e-mail: ngcastro@farmaco.ufrj.br; lqb@nppn.ufrj.br
0960-894X/$ - see front matter ꢀ 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.06.062

4400
P. G. Lima et al. / Bioorg. Med. Chem. Lett. 14 (2004) 4399–4403
R-configuration, including the prototype agonist N-
methyl-D-aspartate (4), and the antagonist MDL
100453 (5).4 The peculiar stereoselectivity of NMDA re-
ceptors is also present at the coagonist glycine site,
where D-serine is the main endogenous agonist in most
regions of the brain.
Considering the potential use of NMDA receptor lig-
ands in therapeutics, we started a program aiming at
the synthesis and pharmacological evaluation of new
heteroaromatic amino acids related to known agonists
and antagonists. As initial targets, we designed com-
pounds 6a,b (series D) as analogues of 5, where a meth-
ylenephosphonate moiety has been substituted for a
heteroaromatic group. Although 6a,b have longer chains
than the usual aromatic analogues such as 1, 2, and 3,
the presence of a carbonyl group introduces conforma-
tional restrictions that might favorably affect binding
to the glutamate site. The role of the distal acidic group
in NMDA antagonists is well established, but the effect
of replacement with basic (pyridine, 6a) or neutral (thio-
phene, 6b) heteroaromatic groups has not been de-
scribed.
Scheme 1. Synthesis of syn-8 and syn-9. Reagents and conditions: (i)
BnNH2 (3equiv), 30ꢁC, 3d, 80%, syn/anti>90%; (ii) (a) H2 (1atm),
MeOH, Pd/C 10% p/p, 1h; (b) (Boc)2O, 95%.
Scheme 2. Synthesis of heteroarylalanines 6a,b. Reagents and condi-
tions: (iii) ArLi (2.5equiv), THF, À78ꢁC; (iv) HCl 1M(2.equiv),
MeOH/H2O, 0ꢁC, neat; (v) (a) NaIO4, THF/H2O; (b) NaClO2,
NaH2PO4, H2O2, 10–15ꢁC; (vi) HCl(g)/MeOH (dry), 0ꢁC.
A synthetic approach was designed to obtain these com-
pounds in enantiomerically pure form, using in the key
step a syn-stereoselective conjugate addition of benzyl-
amine to enoate 7,5 prepared from D-mannitol,6 to
introduce the configuration required for D-aminoacids.
A new transformation of b-aminoesters into b-amino-
arylketones, previously described by our group, was also
employed in the synthetic strategy.7 Since the enantio-
mer of 7 is available from vitamin C,8 the synthesis of
the L-series analogues of 6a,b is also possible using this
approach.
hydrolyzed under mild acidic conditions, leading to the
corresponding N-protected amino diols 12a,b. In con-
trast with what was reported by Dondoni and collabora-
tors for similar hydroxyketones,9 we did not observe the
presence of cyclic hemiketals as the result of the attack
of the primary hydroxy group at the benzylic ketone
group. The diols 12a,b were oxidized with sodium perio-
date6 and the resulting crude aldehydes were allowed to
react with sodium chlorite.10 The formation of a carb-
oxylic function was confirmed by preparation of the
methyl ester derivatives from 13a,b with diazomethane.
The target amino acids 6a,b were obtained as hydrochlo-
rides from 13a,b after cleavage of the protecting
Boc group under nonhumid acidic conditions. Similar
synthetic sequences have been employed by other
groups, without racemization of the a-stereogenic cen-
ter.11
The reaction between 7 and benzylamine led to our key
intermediate, syn-8, in good to excellent stereoselection
(syn/anti>9.5/0.5 at À30ꢁC) (Scheme 1). Hydrogenoly-
sis in MeOH of the N-benzyl group in 8, followed by
in situ addition of a (Boc)2O/MeOH solution, furnished
the N-Boc-aminoester 9 in excellent yield.
The reactional course of the addition of aryllithium rea-
gents to syn-8 and syn-9 was modulated by the nature of
R, as previously described by our laboratory.7 While the
unexpected arylketones 10a,b were obtained from syn-8,
the undesired elimination of benzylamine could be
avoided when the N-Boc derivative syn-9 was the sub-
strate, leading to 11a,b as the only product isolated, in
reasonable to good yields (Scheme 2). The dimethyl-
acetal group in compounds 11a,b could be selectively
Compounds 6a,b were initially screened in binding as-
says of [3H]MK-801 (dizocilpine), a noncompetitive
antagonist of the NMDA receptor that binds to and
blocks the channel pore mostly when the receptor is in
its active conformation. In rat brain homogenates, 6a
inhibited the binding of [3H]MK-801 in a concentra-
tion-dependent manner with IC50 around 300lM

P. G. Lima et al. / Bioorg. Med. Chem. Lett. 14 (2004) 4399–4403
4401
Scheme 3. Conformational equilibrium in 6a and 6b by molecular
modeling (PM3).
Figure 1. Modulation of [3H]MK-801 binding to rat cerebral cortical
membranes. Data are means and SEMfrom triplicate measurements in
1–4 experiments.
independent manner, protects neurons from glutamate
injury (mediated by activation of NMDA receptors),
and partially blocks NMDA-evoked whole-cell currents.
The main activity of 6b was a positive modulation of the
NMDA receptor, as evidenced by the increased binding
of MK-801 and increased channel activation observed in
electrophysiological experiments. The target of 6b for its
agonist-like effect is likely to be the glycine site, because
receptor activation appeared only in the absence of gly-
cine. In the presence of saturating glycine (10lM), 6b
actually inhibited currents activated by a low concentra-
tion of NMDA (10lM). This weak inhibition may sug-
gest that 6b competed with glycine, being an agonist
with slightly lower intrinsic efficacy. Alternatively, 6b
could be a full agonist at the glycine site, with an addi-
tional inhibitory action through other site(s).
(n=2). The effect of 6a was not modulated by addition
of glutamate (0.1–100lM) or glycine (0.01–10lM),
suggesting a noncompetitive mechanism. In contrast,
6b increased the binding of [3H]MK-801, suggesting
a positive allosteric modulation similar to that seen
with L-glutamate (Fig. 1). Neither 6a nor 6b (50–
500lM) significantly affected the binding of [3H]kainic
acid, showing selectivity for NMDA versus kainate
receptors.
Data from neuroprotection assays performed with cul-
tured rat cortical neurons agreed with the binding re-
sults. Compound 6a inhibited delayed neuronal death
induced by a brief application of glutamate (200lM)
and glycine (10lM), by 18.4% 3.9% at 50lM, and by
60.3% 9.4% at 500lM (P<0.05, n=3, ANOVA). In
the same assay, 6b was not significantly protective at
50–500lM (n=3).
Semi-empirical calculations using PM3 hamiltonian
(MOPAC) showed that compounds 6a,b prefer a cyclic
structure, maintained by hydrogen bond between the
hydrogen attached at the nitrogen and the carbonyl
group at the benzylic position, placing the aromatic moi-
eties near the carboxy group and almost coplanar with
the carbonyl group at the six-membered ring. In the
thiophene derivative, a transoid relationship between
the oxygen at the carbonyl group and the sulfur atom
is preferred whereas in the pyridine derivative, cisoid
and transoid conformers (O, N relationship) are almost
isoenergetic (Scheme 3). These conformational trends
will be useful in the design of new analogues and the
synthesis and pharmacological evaluation of derivatives
of 6 is under way in our laboratories.
In whole-cell patch-clamp recordings performed on cul-
tured hippocampal neurons, 6a (200lM) inhibited
NMDA-evoked currents by 14.3 3.1% (n=10). Com-
pound 6b (200lM) also produced a small but significant
inhibition of NMDA-evoked currents, reducing them by
12.5 2.2% at À60mV (n=5). These inhibitory effects
reached an apparent steady state in less than one second
after switching from the solution containing agonist
alone to that containing agonist plus 6a or 6b. Like
NMDA, neither of the two novel compounds produced
whole-cell currents in hippocampal neurons when ap-
plied alone (n=10), thus ruling out an agonist effect
on AMPA, kainate, or GABAA receptors, which are
abundant in this preparation. However, a brief (5s)
application of 6b (200lM) in the presence of NMDA
(10lM, without glycine) induced robust currents, simi-
lar in amplitude and time course to those elicited by gly-
cine (10lM) with NMDA.
2. Experimental section
Enoate 7 and adduct syn-8 were prepared as previously
described, using D-mannitol as starting material.5
2.1. Procedure for addition of 2-lithium-pyridine to 9
n-BuLi 1.4M(1.9mL; 2.7mmol) was diluted with 9mL
hexane and cooled at À78ꢁC. A solution of 2-bromopy-
ridine (0.23mL; 2.4mmol) in THF (2mL) was then
added slowly. After 10min, a solution of syn-9
(0.324g; 2.7mmol) in THF (1mL) was added drop-wise.
The reaction was completed in 15min, then quenched
These data suggest that the two novel D-amino acids
6a,b interact with NMDA receptors in distinct manners.
Results from three different assays suggest that 6a is a
weak noncompetitive antagonist of NMDA receptors,
since it inhibits the binding of MK-801 in an agonist-

4402
P. G. Lima et al. / Bioorg. Med. Chem. Lett. 14 (2004) 4399–4403
with AcOH 20% v/v in THF. The medium was neutral-
ized with NaHCO3 10%, the aqueous phase was sepa-
rated, and extracted with AcOEt. The final organic
layer was dried under anhydrous Na2SO4 and the sol-
vent was removed by evaporation under vacuum. After
flash chromatography, 11a was obtained as a yellow oil
(0.211g; 58% yield).
3.97 (2H, m); 4.53 (1H, dd, J=4.1 and 6.3Hz,); 8.10
(1H, m); 8.55 (2H, m); 8.55 (1H, d, J=5.0Hz), 6b: d
3.72 (1H, d, J=5.0Hz); 3.84 (1H, d, J=5.0Hz,); 4.56
(1H, t, J=5.0Hz); 7.23 (1H, m); 7.95 (2H, m).
2.6. Binding assays
About 100lg of protein from adult Wistar rat brain cor-
tical membranes were incubated in 0.5mL of 5mM
Tris–HCl buffer (pH7.2) and 2nM[ H]MK-801 for 1h
2.2. Procedure for addition of 2-lithium-thiophene to 9
3
To a solution of thiophene (0.4mL; 5mmol) in THF
(10mL) cooled at À78ꢁC was added n-BuLi 1.4M
(3.8mL; 5.3mmol) drop-wise. After 1h, the medium
was set at À15ꢁC for 1h more, then cooled at À78ꢁC
again. A solution of syn-9 (0.606g; 2.0mmol) in THF
(5mL) was added drop-wise. Performing the same pro-
cedure described above, 11b was obtained as a yellow
oil (0.571g; 80% yield).
at 25ꢁC.12 After separation of bound and free ligands
by rapid filtration, the glass fiber filters were rapidly
washed twice with 4mL of ice-cold Tris–HCl buffer.
Specific binding was defined as total binding minus
binding measured in the presence of 3mMketamine.
Similar procedures were used in the binding assay of
5nM[3H]kainic acid. Nonspecific binding was measured
in the presence of 500lMkainic acid.
2.3. General procedure for hydrolysis of the ketal group in
11a,b
2.7. Neuroprotection assay
Neocortical neurons were isolated from Wistar rat fe-
tuses at 18–20days of gestation and maintained in cul-
ture in 24-well plates for 14–16days. Cells were
washed with exposure saline (in mM: NaCl 165, KCl
5, CaCl2 2, D-glucose 30, HEPES 5, tetrodotoxin
15·10À5, NaOH 2; pH7.35) and the excitotoxic insult
was started by adding glutamate 200lMplus glycine
10lMfor 5min. Compounds were added to the expo-
sure saline 15min before glutamate. After 5min of gluta-
mate exposure, cells were washed with culture medium,
compounds were readded (without Glu or Gly) and the
plate returned to the incubator. Four hours later, cell
damage was evaluated by a kinetic spectrophotometric
assay of lactate dehydrogenase in the culture superna-
tant.13 Results from three experiments, each with tripli-
cate cultures, were evaluated by ANOVA.
To a solution of 11a,b in 7mL MeOH/H2 O 3:1 (v/v) was
added 2molequiv of HCl 1M. The medium was cooled
at 0–5ꢁC without agitation. After 16h, the reaction mix-
ture was neutralized with NaHCO3 satd and then it was
extracted with EtOAc (6·10mL). The organic layer was
dried under anhydrous Na2SO4 and the solvent was re-
moved by evaporation under vacuum. The diols 12a,b
were obtained as amorphous solids (92% and 77% yield,
respectively).
2.4. General procedure for oxidative cleavage of diol
group in 12a,b
A suspension of NaIO4 (190mg; 1.1mmol) in THF was
added to a solution of crude 12a,b in THF at rt. After
1h, the medium was diluted with 10mL H2O, then ex-
tracted with CH2Cl2 (6·10mL). The organic layer was
dried under anhydrous Na2SO4 and the solvent was re-
moved by evaporation under vacuum. The crude alde-
hyde was dissolved in MeCN (1mL), followed by
addition of a solution of NaH2PO4ÆH2O (80mg;
0.5mmol) in 0.5mL H2O and later 0.1mL H2O2 30%
(1mmol). The resultant medium was cooled at 0ꢁC, then
a solution of NaClO2 (125mg; 1.1mmol) in H2O (1mL)
was added drop-wise for 15min. After 3h at 10–15ꢁC,
the reaction mixture was acidified with HCl 10%, then
extracted with EtOAc (6·10mL). After usual work-
up, the crude oil was purified by flash chromatography
(AcOEt) furnishing 13a (0.202g; 76%) and 13b
(0.175g; 71%) as amorphous solids.
2.8. Electrophysiology
Hippocampal neurons were isolated from Wistar rat
fetuses and maintained in culture for 23–41days.
Whole-cell membrane currents were recorded by the
patch-clamp technique with membrane potential held
at À60mV and using a fast-switching motorized parallel
tube array system for solution application.14 The agon-
ists NMDA (10lM) and glycine (10lM) were applied
in assay saline (see above), with or without the test
compounds, in short pulses (5–15s). Data were ana-
lyzed with pCLAMP 6 software (Axon Instruments,
USA).
2.5. Procedure for N-Boc deprotection on 13a,b
Acknowledgements
An anhydrous methanolic solution of N-Boc-amino
acids 13a (0.177g) and 13b (0.070g) was cooled in ice
bath and bubbled with HCl gas for 15min. After 1h,
the solvent was removed by evaporation under vacuum,
furnishing the solid hydrochlorides 6a (0.152g; 94%)
Our research was supported by grants from PRONEX
Nos. 41.96.0888.00 (CNPq) and E26/171.162 (FAP-
ERJ), and by CAPES. F.N., N.G.C., and P.R.R.C are
CNPq fellows. S.O.A. received a CNPq (PIBIC) fellow-
ship.
1
and 6b (0.050g; 90%). H NMR (200MHz, D2O) 6a: d

P. G. Lima et al. / Bioorg. Med. Chem. Lett. 14 (2004) 4399–4403
4403
J. D.; Cleary, D. G. Tetrahedron 1989, 45, 391; (c) Mann,
J.; Partlett, N. K.; Tomas, A. J. Chem. Res. Synop. 1987,
369.
References and notes
1. (a) Moloney, M. G. Nat. Prod. Rep. 1998, 15, 205; (b)
Moloney, M. G. Nat. Prod. Rep. 2002, 19, 597.
7. Lima, P. G.; Sequeira, L. C.; Costa, P. R. R. Tetrahedron
Lett. 2001, 42, 3525.
8. Jurczak, J.; Pikul, S.; Bauer, T. Tetrahedron 1986, 42, 447.
9. Dondoni, A.; Boscarato, A.; Marra, A. Synlett, 1993, 256.
10. Dalcanale, E.; Montanari, F. J. Org. Chem. 1986, 51, 567.
11. Badorrey, R.; Cativiela, C.; Dias-de-Villegas, M. D.;
2. Bliss, T. V. P.; Collingridge, G. L. Nature 1993, 361, 31.
3. (a) Olney, J. W. Annu. Rev. Pharmacol. Toxicol. 1990, 30,
47; (b) Kemp, J. A.; McKernan, R. M. Nat. Neurosci.
2002, 5, 1039; (c) Marino, M. J.; Conn, P. J. Curr. Drug
Target. CNS. Neurol. Disord. 2002, 1, 1.
´
Galvez, J. A. Synthesis, 1997, 747.
4. Whitten, J. P.; Baron, B. M.; McDonald, I. A. Bioorg.
Med. Chem. Lett. 1993, 3, 23.
5. Matsunaga, H.; Sakamaki, T.; Nagaoka, H.; Yamada, Y.
Tetrahedron Lett. 1983, 24, 3009.
6. (a) Leonard, J.; Mohialdin, S.; Swain, P. A. Synth.
Commun. 1989, 19, 3529; (b) Marshall, J. A.; Trometer,
12. Gorodinsky, A.; Paas, Y.; Teichberg, V. I. Neurochem. Int.
1993, 23, 285.
13. Koh, J. Y.; Choi, D. W. J. Neurosci. Methods 1987, 20, 83.
14. Castro, N. G.; de Mello, M. C. F.; de Mello, F. G.;
Aracava, Y. Br. J. Pharmacol. 1999, 126, 1847.