Synthesis and biological activity of conformationally restricted analogues of milnacipran: (1S,2R)-1-phenyl-2-[(R)-1-amino-2-propynyl]-N,N- diethylcyclopropanecarboxamide is a novel class of NMDA receptor channel blocker

Article abstract of DOI:10.1021/jm980238m

Conformationally restricted analogues of (±)-(Z)-2-aminomethyl-1- phenyl-N,N-diethylcyclopropanecarboxamide [milnacipran, (±)-1] were designed on the basis of its characteristic cyclopropane structure and were synthesized enantioselectively to develop eff

Full text of DOI:10.1021/jm980238m

J . Med. Chem. 1998, 41, 3507-3514
3507
Syn th esis a n d Biologica l Activity of Con for m a tion a lly Restr icted An a logu es of
Miln a cip r a n : (1S,2R)-1-P h en yl-2-[(R)-1-a m in o-2-p r op yn yl]-N,N-
d ieth ylcyclop r op a n eca r boxa m id e Is a Novel Cla ss of NMDA Recep tor Ch a n n el
Block er
Satoshi Shuto,*,† Shizuka Ono,† Hiroaki Imoto,‡ Kiyonori Yoshii,‡ and Akira Matsuda*,†
Graduate School of Pharmaceutical Sciences, Hokkaido University, Kita-12, Nishi-6, Kita-ku, Sapporo 060-0812, J apan,
Department of Biochemical Engineering and Science, Kyushu Institute of Technology, Iizuka 820-8502, J apan
Received April 20, 1998
Conformationally restricted analogues of (()-(Z)-2-aminomethyl-1-phenyl-N,N-diethylcyclo-
propanecarboxamide [milnacipran, (()-1] were designed on the basis of its characteristic
cyclopropane structure and were synthesized enantioselectively to develop efficient NMDA
receptor antagonists. Among these analogues, (1S,2R)-1-phenyl-2-[(R)-1-amino-2-propynyl]-
N,N-diethylcyclopropanecarboxamide (2d ) had one of the most potent affinities for the receptor,
with a Ki value of 0.29 µM. The blockade of NMDA receptor channels expressed by Xenopus
oocytes by 2d was investigated in detail, and 2d was identified as a new class of open channel
blocker against this receptor.
In tr od u ction
Various antagonists to NMDA (N-methyl-D-aspartic
acid) receptors have been developed1-5 since such recep-
tors may be involved in both chronic and acute neuro-
degenerative disorders.1 Some of these antagonists
have been shown to be effective in experimental models
of epilepsy and stroke.1-3 Unfortunately, currently
available noncompetitive inhibitors have serious behav-
ioral effects4a,b and cause neuronal vacuolization,4c while
competitive inhibitors are often inactive in vivo because
of their poor permeability through the blood-brain
barrier.5 Therefore, another type of efficient NMDA
receptor antagonist is eagerly desired.
(()-(Z)-2-Aminomethyl-1-phenyl-N,N-diethylcyclopro-
panecarboxamide [milnacipran, (()-1],6 a clinically ef-
ficient antidepressant due to competitive inhibition of
the re-uptake of serotonin (5-HT) in the CNS,7 has also
F igu r e 1.
been recognized as a new class of noncompetitive NMDA
receptor antagonist.8 However, the binding affinity of
(()-1 for the NMDA receptor is not very high. We
designed and synthesized four types of conformationally
restricted analogues of (()-1 with different stereochem-
istries; i.e., 2 (type-1) and 3 (type-2), and their enanti-
omers en t-2 (type-3) and en t-3 (type-4), as shown in
Figure 1. In these analogues, an alkyl group introduced
at the R-position of the amino function of (()-1 restricts
the location of the amino group in space, which is
essential for the binding to the NMDA receptor,8 due
to steric repulsion from the diethylcarbamoyl group.9a
Therefore, the conformation of these compounds can be
limited, depending on the configuration of the alkyl
group introduced, and this hypothesis has been sup-
ported by X-ray crystallographic analyses,9a NMR ex-
periments,10 and molecular orbital calculations.9d Bio-
logical evaluations of these compounds have shown that
(1) conformational restriction can improve the activity;
(2) analogues with a (1S,2R)-configuration (type-1 and
type-2) are more potent than the corresponding enan-
tiomers (type-3 and type-4), and type-1 analogues are
more potent than the corresponding type-2 analogues;
and (3) introduction of a substituent bulkier than an
ethyl group, such as a butyl or a phenyl group, at the
1′-position significantly reduces the activity. Thus, we
found that analogues with a type-1 configuration, i.e.;
2a , 2b, and 2c, were potent NMDA receptor antagonists
which significantly inhibited the binding of [3H] MK-
801, with IC50 values about 30-fold stronger than that
of (()-1.9d
In this paper, we describe the synthesis and biological
evaluation of other conformationally restricted ana-
logues of milnacipran as NMDA receptor antagonists.
Considering the above findings, we designed conforma-
tionally restricted analogues with a type-1 configuration
and a sterically small carbon substituent, such as an
ethynyl or a cyano group, at the 1′-position, as shown
* To whom all correspondence and reprint requests should be
addressed. Phone: +81-11-706-3228. Fax: +81-11-706-4980. E-mail:
matuda@pharm.hokudai.ac.jp.
† Hokkaido University.
‡ Kyushu Institute of Technology.
S0022-2623(98)00238-6 CCC: $15.00 © 1998 American Chemical Society
Published on Web 08/01/1998

3508 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 18
Shuto et al.
Sch em e 2a
F igu r e 2.
a
Reagents: (a) Co2(CO)8; (b) 1) TFA, 2) NaN3, 3) CAN; (c) Ph3P,
py, then NH4OH; (d) H2, Pd-C.
chemistry at the 1′-position of 6 was determined to be
R, since the catalytic hydrogenation of 6 with Pd-C in
MeOH gave 1′S-ethyl12 derivative 7, the stereochemistry
of which was previously determined.9a,c As we reported,
the addition reactions of Grignard reagents on cyclo-
propylcarbaldehyde 5 proceed from the least-hindered
si-face in the bisected s-trans conformation, which would
be preferred due to the peculiar stereoelectronic effects
of the cyclopropane ring, to give 1′-addition products
highly stereoselectively.9 Introduction of an azide group
at the 1′-position of 6 with a NaN3/Ph3P/CBr4 system,13
which was effective in the synthetic route for 2a and
2b,9a did not give the corresponding azide product 9 at
all. However, when (trimethylsilyl)ethynyl derivative
8a , which was prepared by the addition reaction of
lithium TMS-acetylide on 5, was used as a substrate,
the desired 1′R-azide derivative 10 was obtained ste-
reoselectively in 78% yield. This reaction is thought to
give the configuration-retained azide product 10 via
participation of the neighboring amide group as an
intermediate I (Scheme 1).9a Treatment of 10 with Ph3P
and NH4OH in pyridine14 afforded a 1′-ethynyl analogue
with type-1 configuration 2d in 77% yield. Catalytic
hydrogenation of 2d with Pd-C in MeOH gave the
known 1′S-ethyl derivative 2b (PPDC),9a so that the 1′-
configuration of 2d was confirmed to be R.
As described below, the 1′-ethynyl derivative 2d with
type-1 configuration was identified as one of the most
potent NMDA receptor antagonists in this series of
compounds. Therefore, we planned to synthesize its 1′-
diastereomer 3d to confirm the stereochemistry-activ-
ity relationship. Compound 3d with type-2 configura-
tion was synthesized via a stereoselective SN1-type
replacement reaction using an azide anion as a nucleo-
phile (Scheme 2). It has been shown that alkynes
readily form complexes with Co2(CO)6 which can be used
efficiently in SN1-type substitution reactions, since they
significantly stabilize the carbocation at the adjacent
position.15 It has also been recognized that cyclopro-
pylmethyl carbocations can be significantly stabilized
by the interaction between a vacant p-orbital of the
carbocation and electrons of the cyclopropane ring,
which are characterized as a strong π-donor.16 Nucleo-
philic substitution reactions at the cyclopropylmethyl
position are facilitated by this interaction, which is
maximal when the cyclopropylmethyl carbocations exist
in either the bisected s-trans or s-cis conformation.16
Considering these previous findings, we presumed that
the nucleophilic substitution reaction between an azide
anion and the carbocation generated from a Co2(CO)6
complex 11 would not proceed via participation of the
F igu r e 3.
Sch em e 1a
a
Reagents: (a) RMgBr or RLi; (b) H2, Pd-C; (c) NaN3, CBr4,
Ph3P; (d) Ph3P, py, then NH4OH.
in Figure 2. We also synthesized 1′-carbamoyl, -meth-
oxycarbonyl, and -carboxylic acid derivatives, 2f-h , to
investigate the effects of the functional characteristics
of the 1′-substituent on the activity. In this study,
(1S,2R)-1-phenyl-2-[(R)-1-amino-2-propynyl]-N,N-di-
ethylcyclopropanecarboxamide (2d ) was identified as
one of the most potent NMDA receptor antagonists
among the conformationally restricted analogues. The
effect of 2d on NMDA receptor channels expressed by
Xenopus oocytes was investigated in detail. Its selectiv-
ity for NMDA receptors among other glutamate receptor
subtypes and the mechanism of its blocking effect on
NMDA receptors were investigated using Xenopus oo-
cytes under voltage-clamp conditions. Thus, 2d is a new
class of an open channel blocker that is different from
other blockers such as MK-801 and phencyclidine (PCP)
(Figure 3).
Resu lts a n d Discu ssion
Ch em istr y. All of the target compounds were syn-
thesized from an optically active cyclopropylcarbalde-
hyde derivative 5 with a (1S,2R)-configuration,9a which
was readily prepared from (R)-epichlorohydrin via a
lactone 411 (Scheme 1). The Grignard reaction of 5 with
HCtCMgBr in THF at -20 °C gave 1′R-product 6
highly diastereoselectively in 89% yield. The stereo-

Analogues of Milnacipran
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 18 3509
Sch em e 3
ester 2g was hydrolyzed with 6 M HCl to give amino
acid 2f as a hydrochloride in 66% yield. After the amino
group of 2g was protected by a Boc group, 14 was
converted to 1′-vinyl derivative 17, the stereochemistry
of which has been identified previously.9d Thus, the
configurations of the type-1 analogues 2e-h were
confirmed.
Bin d in g Affin ity for NMDA Recep tor . The syn-
thesized compounds were evaluated for their binding
affinity for the NMDA receptor of cerebral cortical
synaptic membranes from rats with [3H] MK-801 as a
radioligand.18 The binding affinity was significantly
affected by the substituent at the 1′-position, and the
results are shown in Table 1.
Sch em e 4a
Notably, the 1′-ethynyl analogue 2d (PPYDC) with
type-1 configuration [(1S,2R)-1-phenyl-2-[(R)-1-amino-
2-propynyl]-N,N-diethylcyclopropanecarboxamide] sig-
nificantly inhibited the binding of [3H] MK-801 with an
IC50 value of 0.29 ( 0.2 µM, which is about 20-fold
stronger than that of (()-1 (IC50 ) 6.3 ( 0.29 µM).
Analogues with a CN, CO2Me, or CO2H group at the
1′-position, i.e., 2e-g, were almost inactive, and the 1′-
carbamoyl analogue 2h showed weak binding affinity
to the receptor (IC50 ) 27 ( 1.8 µM). These results
suggest that the electron-withdrawing effect of the 1′-
substituents, which decreases the basicity of the amino
group, may diminish the potency. It is especially
interesting that the 1′-CN analogue 2e with type-1
configuration is inactive, although a cyano group has
steric and electronic features very similar to those of
an ethynyl group.
Considering these results, we synthesized a 1′-ethynyl
derivative with type-2 configuration, 3d , and evaluated
its binding affinity for the receptor. As a result, 3d
showed an only weak activity. This is consistent with
a previous finding that analogues with type-1 configu-
ration are more potent than the corresponding ana-
logues with type-2 configuration.9d
a
Reagents: (a) 1) TMSCN, Et3N, 2) NH3/MeOH; (b) 1) HCl/
MeOH, 2) H2O; (c) 6 M HCl; (d) Boc2O; (e) DIBAL-H; (f) Swern
ox.; (g) Ph3PCH3Br, BuLi.
neighboring group (I in Scheme 1) as mentioned above
but rather via an SN1 reaction due to effective stabiliza-
tion of the 1′-carbocation by both the Co2(CO)6-alkyne
complex and the cyclopropane ring. If this is the case,
the reaction would stereoselectively give the desired 1′S-
azide 12 since, with regard to the intermediate carboca-
tion, a bisected s-trans conformation would be preferred
over a bisected s-cis conformation due to steric repulsion
between the cobalt-bound alkynylmethyl group and the
N,N-diethylcarbamoyl group, and an azide anion would
attack the intermediate at the least-hindered face to
give the desired product 13 with high selectivity (Scheme
3). In fact, when complex 11, prepared from 6 and
Co2(CO)8, was successively treated with TFA and NaN3
in CH2Cl2, the desired azide derivative was obtained as
a sole product. Oxidative cleavage of the cobalt complex
with cerium nitrite (CAN) in acetone gave the 1′S-azide
12 in 56% yield from 6. Treatment of 12 with Ph3P and
NH4OH in pyridine gave 3d with type-2 configuration.
The stereochemistry was confirmed by converting 3d
into the previously reported 3b9a by catalytic hydroge-
nation.
As described above, 2d (PPYDC) had the most sig-
nificant affinity for the receptor among the conforma-
tionally restricted analogues.
In h ibitor y Effects on th e Up ta k e of 5-HT. The
inhibitory effects of the compounds on the uptake of
5-HT by nerve terminals of cerebral cortical synaptic
membrane from rats were evaluated with [3H] parox-
etine as a radioligand, and the results are shown in
Table 1. Among the newly synthesized compounds,
type-1 ethynyl derivative 2d (PPYDC) was the most
potent 5-HT uptake inhibitor (Ki ) 0.19 ( 0.2 µM)
although its effect was about 210-fold less than that of
milnacipran [(()-1, Ki ) 0.0085 ( 0.0006 µM]. On the
basis of these results, it appears as though the inhibi-
tory potency of the conformationally restricted ana-
logues with type-1 configuration against the 5-HT
uptake is significantly affected by the bulkiness of the
1′-substituent (Me > CtCH > Et).
The synthesis of the 1′-CN, -CO2Me, and -CONH2
analogues with type-1 configuration is summarized in
Scheme 4. Treatment of aldehyde 5 with TMSCN and
17
Et3N in CH2Cl2 gave a 1′-diastereomeric mixture of
the corresponding TMS-cyanohydrins, which was im-
mediately heated with NH3/MeOH in a steel tube to
afford amines 2e and 3e from 5 in isolated yields of 64%
and 32%, respectively, after silica gel column chroma-
tography. The 1′R-cyano analogue 2e with type-1
configuration was successively treated with HCl/MeOH
and water to give type-1 methyl ester 2g and type-1
amide 2h in yields of 50% and 24%, respectively. The
Effect on NMDA Recep tor Exp r essed by Oo-
cytes. The effect of PPYDC on glutamate receptor
subtypes expressed by Xenopus oocytes injected with
total brain mRNA of the Wistar-strain rat under two-
electrode voltage-clamp conditions was investigated as
described previously.19 The injected oocytes produced
various glutamate receptor subtypes (metabotropic and
ionotropic receptors), namely, NMDA, kainate, and

3510 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 18
Shuto et al.
Ta ble 1. Effects of Compounds on the NMDA Receptor Binding and the 5-HT Uptake
compd
stereochemistry
1′-substituent
NMDA receptor bindinga (IC50, µM)
5-HT uptakeb (Ki, µM)
(()-1
6.3 ( 0.3
0.35 ( 0.08
0.20 ( 0.02
0.29 ( 0.2
96 ( 10
0.0085 ( 0.0006
0.014 ( 0.002
24 ( 0.9
2a
type-1
type-1
type-1
type-1
type-1
type-1
type-1
type-2
type-2
type-2
type-2
Me
Et
2b
2d
2e
2f
2g
2h
3a
3b
CtCH
CtN
CO2H
CO2Me
CONH2
Me
Et
CtCH
CtN
0.19 ( 0.2
>100
>100
25 ( 5
>100
2.4 ( 0.3
>100
15 ( 0.8
>100
>100
27 ( 1.8
6.5 ( 0.5
8.2 ( 2.1
16 ( 4.8
>100
3d
3e
not tested
not tested
ketamine
0.61( 0.46
a
b
Assay was done with cerebral cortical synaptic membrane of rats using [3H]MK-801. Assay was done with cerebral cortical synaptic
membrane of rats using [3H]paroxetine.
F igu r e 5. Suppression curve. Oocytes were stimulated with
30 µM NMDA supplemented with 10 µM glycine in the
presence of different concentrations of PPYDC. Plotted relative
responses are the mean ( SD of five oocytes. Holding potential,
-50 mV.
F igu r e 4. Effect of 10 µM PPYDC on different neurotrans-
mitter receptor responses. Oocytes were voltage-clamped at
-50 mV and stimulated with selective agonists in the presence
and absence of 10 µM PPYDC. Lower bars on each current
trace show the duration of stimulation, and upper bars show
the duration of PPYDC application. Scale: 40 nA for NMDA;
10 nA for kainate, GABA, and ACh; and 2 nA for AMPA.
by fitting the following equation to the experimental
data
R ) 1 - {1/(1 + (IC50/[PPYDC])n)}
(1)
Ta ble 2. Effect of 10 µM PPYDC (2d ) on Different
Neurotransmitter Receptors
relative response
where R is the relative response in the presence of a
given concentration of PPYDC, [PPYDC].
receptor
(%)a mean ( SD
32.1 ( 0.07b
agonist
NMDA receptors
30 µM NMDA +
10 µM glycine
30 µM kainate
10 µM AMPA
30 µM GABA
1 µM ACh
As Figure 5 shows, eq 1 fits the plotted data points
with an IC50 of 4.75 µM, an n of 0.983, and a correlation
coefficient between eq 1 and the experimental data of
0.994. The n value was close to unity, indicating 1:1
stoichiometry; one PPYDC molecule blocks one NMDA
receptor molecule.
kainate receptors
AMPA receptors
GABAA receptors
AChM1 receptors
5HT2C receptors
96.9 ( 0.05
100.3 ( 0.08
99.6 ( 0.03
102.0 ( 0.05
100.3 ( 0.16
0.1 µM 5HT
a
NMDA receptor channels require both NMDA and
glycine to open. PPYDC decreased the saturation levels
of the concentration-response curve for NMDA supple-
mented with 10 µM glycine (Figure 6a) and that for
glycine supplemented with 30 µM NMDA (Figure 6b)
without shifting the concentration-response curves.
These results indicate that PPYDC blocks NMDA
receptors without competing with NMDA or glycine.
To test this conclusion, we compared the concentra-
tion-response curves shown in Figure 6 with an equa-
tion (eq 2) that assumes the uncompetitive blockade of
NMDA receptors
The numerals are means and SD of relative responses obtained
b
from 5 oocytes. p < 0.05, paired two-tailed t-test.
AMPA (R-amino-3-hydroxy-5-methyl-4-isoxazole propi-
onate) receptors, and other neurotransmitter receptors,
such as GABAA, AChM1, and 5HT2C receptors. In the
presence of PPYDC (10 µM), the responses of NMDA
receptors were decreased (Figure 4, Table 2). In con-
trast, the responses of other glutamate receptor sub-
types and neurotransmitter receptors remained un-
changed in the presence of 10 µM of PPYDC. PPYDC
had no effect on NMDA receptor responses when it was
applied prior to stimulating NMDA solution (data not
shown). These results indicate that PPYDC selectively
blocks opened NMDA receptor channels.
The blockade of the responses to 30 µM NMDA
increased with an increase in the concentration of
PPYDC (Figure 5). An IC50 value (the concentration of
a blocker at which receptor responses are reduced by
half) and an empirical Hill coefficient n were determined
n
Kd
+ 1
{
}
[Acontrol
n
]
R )
(2)
Kd
[PPYDC]
KPPYDC
+
+ 1
{
}
[A]
where R is the relative response, [Acontrol] is the con-

Analogues of Milnacipran
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 18 3511
F igu r e 7. Blocking and unblocking by 10 µM PPYDC and 1
µM MK-801 in the continuous presence of 30 µM NMDA
supplemented with 10 µM glycine. These traces were obtained
from the same oocyte to eliminate differences in the open
probability of NMDA receptors. Broken line, zero current level.
Holding potential, -50 mV.
to measure with our present system. This difference is
not surprising because the structure of PPYDC is quite
different from those of other known blockers (Figure 3).
The moderate recovery time constant of PPYDC sug-
gests that its pharmacological effect on NMDA receptors
is different from those of known blockers of this receptor.
In conclusion, we developed conformationally re-
stricted analogues of milnacipran [(()-1] and found a
potent NMDA receptor antagonist, PPYDC (2d ), which
was identified as a new class of channel blockers against
this receptor. PPYDC may be a desirable NMDA
antagonist, since milnacipran, the prototype of PPYDC,
has been shown clinically to be free from serious side-
effects and to be transportable to the brain.7
F igu r e 6. Concentration-response curves for NMDA (A) and
glycine (B) in the absence (b) and presence (O) of 3 µM PPYDC.
Different concentrations of NMDA (A) and glycine (B) were
supplemented with 10 µM glycine and 30 µM NMDA, respec-
tively. Holding potential, -50 mV. Lines were drawn using
the nonlinear least-squares method following eq 2. Each point
represents the mean ( SD of five oocytes.
centration of the control (30 µM NMDA in Figure 6a
and 10 µM glycine in Figure 6b), [A] is a given
concentration of test stimulus (NMDA in Figure 6a and
glycine in Figure 6b), Kd and KPPYDC are the dissociation
constants for the agonist (NMDA in Figure 6a and
glycine in Figure 6b) and PPYDC, respectively, [PPYDC]
is a fixed concentration of PPYDC (3 µM), and n is the
empirical Hill coefficient for the agonists. The Hill
coefficient for PPYDC is taken as unity based on an
analysis of the suppression curve shown in Figure 5.
As Figure 6 shows, the concentration curves for both
NMDA and glycine are described by eq 2 using the
nonlinear least-squares method, and the correlation
coefficients for the fitting ranged from 0.976 to 0.999.
This good fit supports the above conclusion that the
mode of action is uncompetitive. These results, together
with the above result that PPYDC blocked only acti-
vated NMDA receptors, indicate that PPYDC acts as
an open channel blocker against NMDA receptors.
The estimated dissociation constants of PPYDC were
1.81 µM in the presence of 10 µM glycine and 3.72 µM
in the presence of 30 µM NMDA; i.e., these values were
comparable to each other. The estimated dissociation
constants for NMDA and glycine were 27.0 µM and 0.25
µM, respectively, and the empirical Hill coefficients were
1.04 for NMDA and 0.96 for glycine.
Exp er im en ta l Section
Melting points were determined on a Yanagimoto MP-3
micro-melting point apparatus and are uncorrected. The NMR
spectra were recorded with a J EOL EX-270, -400, or Bruker
AMX 500 spectrometer with tetramethylsilane as an internal
standard. Chemical shifts were reported in parts per million
(δ), and signals are expressed as s (singlet), d (doublet), t
(triplet), q (quartet), m (multiplet), or br (broad). Mass spectra
were measured on a J EOL J MS-D300 spectrometer. Thin-
layer chromatography was done on Merck coated plate 60F254
.
Silica gel chromatography was done with Merck silica gel 5715.
Reactions were performed under argon.
(1S,2R)-1-P h en yl-2-[(R)-1-h yd r oxy-2-p r op yn yl]-N,N-d i-
eth ylcyclop r op a n eca r boxa m id e (6). A THF solution of
HCtCMgBr (0.5 M, 24 mL, 12 mmol) was added slowly to a
solution of 5 (1.94 g, 7.9 mmol) in THF (20 mL) at -10 °C.
The mixture was stirred at the same temperature for 3 h and
was quenched with aqueous saturated NH4Cl (20 mL). After
the mixture was concentrated in vacuo, the mixture was
partitioned between EtOAc and H2O. The organic layer was
washed with brine, dried (Na2SO4), evaporated, and purified
by column chromatography (silica gel; hexane/EtOAc, 3:1) to
1
give 6 (2.06 g, 96% as a solid): mp 77-80 °C; H NMR (500
MHz, CDCl3) 0.90 (3 H, t, J ) 7.0 Hz), 1.13 (3 H, t, J ) 7.0
Hz), 1.20 (1 H, dd, J ) 5.5, 6.5 Hz), 1.69 (1 H, ddd, J ) 6.5,
9.0, 10.0 Hz), 1.80 (1 H, dd, J ) 5.5, 9.0 Hz), 2.44 (1 H, d, J )
2.5 Hz), 3.31-3.44 (3 H, m), 3.48 (1 H, dq, J ) 14.0, 7.0 Hz),
3.96 (1 H, ddd, J ) 2.5, 2.5, 10.0 Hz), 5.50 (1 H, d, J ) 2.5
Several potent open channel blockers against the
NMDA receptor, such as MK-801 and PCP, which have
similar structural and biological features, are
known.1,2,4,20 Although PPYDC may block the same site
that MK-801 blocks within the NMDA channel, they
typically gave different results in unblocking processes.
As Figure 7 shows, the time constant for the recovery
of NMDA receptors blocked by PPYDC was about 8 s,
which is much shorter than the recovery time constant
of MK-801-blocked NMDA receptors, which was too long
Hz), 7.21-7.32 (5 H, m); MS (EI) m/z 271 (M+). Anal. (C17H21
NO2) C, H, N.
-
Ca ta lytic Hyd r ogen a tion of 6. A mixture of 6 (30 mg,
0.11 mmol) and 10% Pd-charcoal (5 mg) in MeOH (2 mL) was
stirred under atmospheric pressure of hydrogen at room
temperature for 2 h, and then the catalyst was filtered off with
Celite. The filtrate was evaporated, and the residue was
purified by column chromatography (silica gel; hexane/EtOAc,
2:1) to give 7 (27 mg, 89% as an oil), of which spectral data
were in accord with those reported previously.9a

3512 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 18
Shuto et al.
(1S,2R)-1-P h en yl-2-[(R)-1-h yd r oxy-3-(tr im eth ylsilyl)-2-
p r op yn yl]-N,N-d ieth ylcyclop r op a n eca r boxa m id e (8) a n d
Its 1′-Dia ster eom er . A hexane solution of BuLi (1.64 M, 2.3
mL, 3.9 mmol) was added to a solution of TMSCtCH (0.57
mL, 4.0 mmol) in THF (6 mL) at -78 °C, and the resulting
solution was stirred at the same temperature for 1 h. To the
resulting solution was added slowly a solution of 5 (490 mg,
2.0 mmol) in THF (10 mL) at -78 °C. The mixture was stirred
at the same temperature for 1 h and was quenched with
aqueous saturated NH4Cl (5 mL). After the mixture was
concentrated in vacuo, the mixture was partitioned between
EtOAc and H2O. The organic layer was washed with brine,
dried (Na2SO4), evaporated, and purified by column chroma-
tography (silica gel; hexane/EtOAc, 5:1) to give 8 and its 1′S-
diastereomer. Physical data of 8 (585 mg, 85% as crystals)
Ca ta lytic Hyd r ogen a tion of 2d . Compound 2d (free
amine, 11 mg, 0.040 mmol) was hydrogenated as described
above for 6. After the residue was purified by column
chromatography (silica gel; CHCl3
/MeOH/28% NH OH, 45:5:
4
0.1), 2b (8 mg, 74% as an oil) was obtained, of which spectral
data were in accord with those reported previously.9d
(1S,2R)-1-P h en yl-2-[(S)-1-a zid o-2-p r op yn yl]-N,N-d ieth -
ylcyclop r op a n eca r boxa m id e (12). A mixture of 6 (271 mg,
1.0 mmol) and Co2(CO)8 (380 mg, 1.0 mmol) in CH2Cl2 (10 mL)
was stirred at room temperature for 20 min, and then TFA
(770 µL, 10 mmol) was added at 0 °C. After the mixture was
stirred at 0 °C for 10 min, NaN3 (650 mg, 10 mmol) was added,
and the resulting mixture was stirred at room temperature
for 1.5 h. Water was added, and the resulting mixture was
partitioned. The organic layer was washed with brine, dried
(Na2SO4), and evaporated. A mixture of the residue and CAN
(274 mg, 5.0 mmol) in acetone (15 mL) was stirred at room
temperature for 1 h. The resulting mixture was partitioned
between EtOAc and brine, and the organic layer was dried
(Na2SO4), and evaporated. The residue was purified by column
chromatography (neutral silica gel; hexane/EtOAc, 5:1) to give
12 (165 mg, 56% as an oil): 1H NMR (270 MHz, CDCl3) 0.46
(3 H, t, J ) 7.0 Hz), 1.11 (3 H, t, J ) 7.0 Hz), 1.14 (1 H, dd, J
) 5.5, 9.5 Hz), 1.74 (1 H, dd, J ) 5.5, 6.0 Hz), 2.18 (1 H, ddd,
J ) 6.0, 9.5, 10.0 Hz), 2.65 (1 H, d, J ) 2.5 Hz), 3.04 (1 H, dq,
J ) 14.0, 7.0 Hz), 3.13 (1 H, dq, J ) 14.0, 7.0 Hz), 3.55 (1 H,
dq, J ) 14.0, 7.0 Hz), 3.61 (1 H, dq, J ) 14.0, 7.0 Hz), 3.79 (1
H, dd, J ) 2.5, 10.0 Hz), 7.20-7.35 (5 H, m); IR (neat), 2120
cm-1(νN3).
1
are as follows: mp 107-109 °C; H NMR (500 MHz, CDCl3)
0.17 (9 H, s), 0.91 (3 H, t, J ) 7.0 Hz), 1.12 (3 H, t, J ) 7.0
Hz), 1.21 (1 H, dd, J ) 5.5, 6.5 Hz), 1.68 (1 H, ddd, J ) 6.5,
9.0, 10.0 Hz), 1.79 (1 H, dd, J ) 5.5, 9.0 Hz), 3.33-3.41 (3 H,
m), 3.49 (1 H, dq, J ) 14.0, 7.0 Hz), 3.95 (1 H, dd, J ) 2.5,
10.0 Hz), 5.52 (1 H, d, J ) 2.5 Hz), 7.20-7.31 (5 H, m); MS
(EI) m/z 343 (M+). Anal. (C20H29NO2Si) C, H, N. Physical
data of the 1′S-diastereomer of 8 [(1S,2R)-1-Phenyl-2-[(S)-1-
hydroxy-3-(trimethylsilyl)-2-propynyl]-N,N-diethylcyclopro-
panecarboxamide, 89 mg, 13% as crystals] are as follows: mp
75-77 °C; 1H NMR (500 MHz, CDCl3) 0.15 (9 H, s), 0.98 (3 H,
t, J ) 7.0 Hz), 1.16 (3 H, t, J ) 7.0 Hz), 1.65-1.72 (3 H, m),
3.15 (1 H, dq, J ) 14.0, 7.0 Hz), 3.31 (1 H, dq, J ) 14.0, 7.0
Hz), 3.61 (2 H, dq, J ) 14.0, 7.0 Hz), 4.90 (1 H, dd, J ) 5.5,
10.5 Hz), 5.96 (1 H, d, J ) 10.5 Hz), 7.19-7.31 (5 H, m); MS
(EI) m/z 343 (M+). Anal. (C20H29NO2Si) C, H, N.
(1S,2R)-1-P h en yl-2-[(S)-1-a m in o-2-p r op yn yl]-N,N-d i-
eth ylcyclopr opan ecar boxam ide Hydr och lor ide (3d). Com-
pound 3d was prepared from 12 (150 mg, 0.51 mmol) as
described above for 2d . After the residue was purified by
column chromatography (silica gel; CHCl3/MeOH/28% NH4-
OH, 45:5:0.1) and treated with ion-exchange resin (Diaion WA-
30, Cl- form), 3d (83 mg, 60% as crystals) was obtained as a
hydrochloride: mp 95-98 °C; [R]23D ) -24.0 (c 0.620, MeOH);
1H NMR (400 MHz, CDCl3) 0.97 (3 H, t, J ) 7.0 Hz), 1.12 (3
H, t, J ) 7.0 Hz), 1.73-1.86 (2 H, m), 2.05-2.10 (1 H, m), 2.55
(1 H, s), 3.25 (1 H, dq, J ) 14.0, 7.0 Hz), 3.36-3.46 (2 H, m),
3.58 (1 H, dq, J ) 14.0, 7.0 Hz), 4.91 (1 H, br s), 7.20-7.31 (5
(1S,2R)-1-P h en yl-2-[(R)-1-a zid o-3-tr im eth ylsilyl-2-p r o-
p yn yl]-N,N-d ieth ylcyclop r op a n eca r boxa m id e (10). To a
solution of 8 (343 mg, 1.0 mmol) in DMF (8 mL) at 0 °C were
added NaN3 (1.17 g, 18 mmol), Ph3P (787 mg, 3.0 mmol), and
CBr4 (995 mg, 3.0 mmol), and the resulting mixture was stirred
at room temperature for 3 h. After the addition of water, the
resulting mixture was evaporated, and the residue was
partitioned between brine and EtOAc. The organic layer was
dried (Na2SO4), evaporated, and purified by column chroma-
tography (silica gel; hexane/EtOAc, 1:10) to give 10 (287 mg,
78% as an oil): 1H NMR (500 MHz, CDCl3) 0.21 (9 H, s), 0.44
(3 H, t, J ) 7.0 Hz), 1.07 (1 H, dd, J ) 5.5, 9.0 Hz), 1.10 (3 H,
t, J ) 7.0 Hz), 1.78 (1 H, dd, J ) 5.5, 6.0 Hz), 2.14 (1 H, ddd,
J ) 6.0, 9.0, 9.5 Hz), 3.02 (1 H, dq, J ) 14.0, 7.0 Hz), 3.10 (1
H, dq, J ) 14.0, 7.0 Hz), 3.52-3.61 (2 H, m), 3.85 (1 H, d, J )
9.5 Hz), 7.20-7.31 (5 H, m); MS (EI) m/z 368 (M+). Anal.
(C20H28N4OSi‚0.25H2O) C, H, N.
H, m), 9.51 (3 H, br s); MS (EI) m/z 270 (M+). Anal. (C17H23
ClN2O‚1.6H2O) C, H, N.
-
Ca ta lytic Hyd r ogen a tion of 3d . Compound 3d (free
amine, 17 mg, 0.063 mmol) was hydrogenated as described
above for 6. After the residue was purified by column
chromatography (silica gel; CHCl3/MeOH/28% NH4OH, 45:5:
0.1), 3b (17 mg, 98% as an oil) was obtained, of which spectral
data were in accord with those reported previously.9a
(1S,2R)-1-P h en yl-2-[(R)-1-a m in o-2-p r op yn yl]-N,N-d i-
eth ylcyclop r op a n eca r boxa m id e Hyd r och lor id e (2d ). Af-
ter a mixture of 10 (154 mg, 0.42 mmol) and Ph3P (219 mg,
0.84 mmol) in pyridine (6 mL) was stirred at room temperature
for 30 min, 28% NH4OH (4 mL) was added, and the resulting
mixture was stirred at room temperature for 36 h. The
mixture was evaporated, and the residue was partitioned
between Et2O and water. The organic layer was separated
and was washed with brine, dried (Na2SO4), evaporated, and
purified by column chromatography (silica gel; CHCl3/MeOH/
28% NH4OH, 300:10:0.1) to give 2d as an oil (87 mg, 77%).
The oil was partitioned between CHCl3 and 1 M NaOH, and
then the CHCl3 layer was washed twice with brine, dried (Na2-
SO4), and evaporated. The residue was dissolved in MeOH (1
mL), and the solution was put on a column of Diaion WA-30
resin (2 × 8 cm, Cl- form), which was developed with MeOH.
The solvent was evaporated, and the residue was treated with
Et2O to give white crystals of 2d (95 mg, 71% from 10 as a
hydrochloride): mp 96-100 °C; [R]21D ) +73.1 (c 0.840, CHCl3);
1H NMR (500 MHz, CDCl3) 0.90 (3 H, t, J ) 7.0 Hz), 1.09 (3
H, t, J ) 7.0 Hz), 1.24 (1 H, dd, J ) 6.0, 6.0 Hz), 1.95 (1 H, dd,
J ) 6.0, 9.0 Hz), 2.01 (1 H, ddd, J ) 6.0, 9.0, 10.0 Hz), 2.78 (1
H, d, J ) 1.5 Hz), 3.29 (1 H, dq, J ) 14.0, 7.0 Hz), 3.34-3.49
(2 H, m), 3.52 (1 H, dd, J ) 10.0, 1.5 Hz), 7.20-7.31 (5 H, m),
9.46 (3 H, br s); MS (EI) m/z 270 (M+). Anal. (C17H23ClN2O‚
0.4H2O) C, H, N.
(1S,2R)-1-P h en yl-2-[(R)-a m in ocya n om et h yl]-N,N-d i-
eth ylcyclop r op a n eca r boxa m id e (2e) a n d (1S,2R)-1-P h en -
yl-2-[(S)-a m in ocya n om eth yl]-N,N-d ieth ylcyclop r op a n e-
ca r boxa m id e (3e). A solution of 5 (858 mg, 3.50 mmol),
TMSCN (600 µL, 4.4 mmol), and Et3N (62 µL, 0.44 mmol) in
CH2Cl2 (10 mL) was stirred at room temperature for 16 h. To
the mixture was added NH3/MeOH (saturated at 0 °C, 40 mL),
and the whole was heated at 50 °C for 4 h in a steel tube.
After the solvent was evaporated, the residue was partitioned
between EtOAc and brine. The organic layer was dried (Na2-
SO4), evaporated, and purified by column chromatography
(silica gel; hexane/EtOAc, 1:2, then EtOAc) to give 2e (free
amine, 605 mg, 64% as yellow crystals) and 3e (free amine,
305 mg, 32% as a yellow oil). The hydrochlorides of the
diastereomers were prepared as described above for 2d . 2e
(hydrochloride): mp 140-142 °C; [R]27 ) -47.7 (c 0.760,
D
1
CHCl3); H NMR (500 MHz, CDCl3) 0.87 (3 H, t, J ) 7.0 Hz),
1.09 (3 H, t, J ) 7.0 Hz), 1.44 (1 H, dd, J ) 6.0, 6.5 Hz), 2.05
(1 H, dd, J ) 6.0, 8.5 Hz), 2.16 (1 H, ddd, J ) 6.5, 8.5, 10.5
Hz), 3.26 (1 H, dq, J ) 14.0, 7.0 Hz), 3.31-3.47 (3 H, m), 3.95
(1 H, d, J ) 10.5 Hz), 7.20-7.33 (5 H, m), 9.87 (3 H, br s); MS
(EI) m/z 271 (M+). Anal. (C16H22ClN3O) C, H, N. 3e (hydro-
chloride): mp 145-147 °C; [R]27 ) +20.4 (c 0.850, CHCl3);
D
1H NMR (500 MHz, CDCl3) 0.92 (3 H, t, J ) 7.0 Hz), 1.16 (3
H, t, J ) 7.0 Hz), 1.68 (1 H, dd, J ) 6.5, 6.5 Hz), 2.02 (1 H, dd,

Analogues of Milnacipran
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 18 3513
J ) 6.5, 9.0 Hz), 2.09 (1 H, ddd, J ) 5.5, 6.5, 9.0 Hz), 3.29 (1
H, dq, J ) 14.0, 7.0 Hz), 3.42-3.53 (3 H, m), 5.33 (1 H, d, J )
5.5 Hz), 7.23-7.34 (5 H, m), 9.96 (3 H, br s); MS (EI) m/z 271
(M+). Anal. (C16H22ClN3O‚0.3H2O) C, H, N.
NMR (500 MHz, CDCl3) 0.82 (3 H, t, J ) 6.5 Hz), 1.15 (3 H, t,
J ) 7.0 Hz), 1.41-1.58 (m, 12 H), 1.88 (1 H, dd, J ) 7.2, 15.8
Hz), 3.22-3.53 (4 H, m), 3.79 (s, 3 H), 4.27 (br s, 1 H), 5.30 (br
s, 1 H), 7.19-7.31 (5 H, m); MS (EI) m/z 404 (M+).
(1S,2R)-1-P h en yl-2-[(S)-1-a m in o-2-p r op en yl]-N,N-d i-
eth ylcyclop r op a n eca r boxa m id e (17). To a solution of
oxalyl chloride (75 µL, 0.088 mmol) in CH2Cl2 (2 mL) was
added slowly a mixture of DMSO (16 µL, 0.18 mmol) and CH2-
Cl2 (2 mL) at -78 °C, and the mixture was stirred at the same
temperature for 30 min. To the resulting mixture was added
slowly a solution of 15 (18 mg, 0.048 mmol) in CH2Cl2 (2 mL);
the whole was stirred at the same temperature for 2 h, and
then Et3N (49 µL, 0.35 mmol) was added. After being stirred
at -78 °C further for 1 h, the reaction mixture was quenched
with aqueous saturated NH4Cl, and then CH2Cl2 (10 mL) was
added. The organic layer was separated and washed with
brine, dried (Na2SO4), evaporated, and purified by column
chromatography (silica gel; hexane/EtOAc, 1:1) to give 16 as
an oil, which was used immediately in the next reaction.
To a suspension of Ph3PCH3Br (10 mg, 0.029 mmol) in THF
(1 mL) was added a BuLi solution (1.63 M in hexane, 16 µL,
0.029 mmol) at -21 °C, and the mixture was stirred at the
same temperature for 30 min. To the resulting mixture was
added slowly a solution of 16 (8 mg, 0.021 mmol) in THF (1
mL) at -21 °C. The mixture was stirred at the same
temperature for 1.5 h and then was quenched with aqueous
saturated NH4Cl (2 mL). After the mixture was concentrated
in vacuo, EtOAc and water were added, and then the mixture
was partitioned. The organic layer was washed with brine,
dried (Na2SO4), evaporated, and purified by column chroma-
tography (silica gel; hexane/EtOAc, 1:1) to give 17 (4 mg, 33%
as on oil), of which spectral data were in accord with those
reported previously.9d
(1S,2R)-1-P h en yl-2-[(R)-am in o(m eth oxycar bon yl)m eth -
yl]-N,N-d iet h ylcyclop r op a n eca r b oxa m id e (2g) a n d
(1S,2R)-1-P h en yl-2-[(R)-a m in e(ca r ba m oyl)m eth yl]-N,N-
d iet h ylcyclop r op a n eca r b oxa m id e (2h ). HCl gas was
bubbled and saturated into a solution of 2e (free amine, 691
mg, 2.55 mmol) in MeOH (15 mL) at 0 °C, and the resulting
solution was stirred at the same temperature for 1 h. After
ice-water (75 mL) was added, the resulting mixture was
stirred at 0 °C and was neutralized with NaHCO3. The
resulting mixture was extracted with EtOAc, and the organic
layer was washed with brine, dried (Na2SO4), and evaporated.
The residue was purified by column chromatography (silica
gel; CHCl3/MeOH, 20:1 and then 5:1) to give 2g (free amine,
385 mg, 50% as an oil) and 2h (free amine, 173 mg, 24% as a
solid). 2g: 1H NMR (500 MHz, CDCl3) 0.81 (3 H, t, J ) 7.0
Hz), 1.13 (3 H, t, J ) 7.0 Hz), 1.41 (1 H, dd, J ) 5.0, 6.0 Hz),
1.50 (1 H, ddd, J ) 6.0, 9.0, 9.5 Hz), 1.56 (1 H, dd, J ) 5.0, 9.0
Hz), 2.55 (2 H, br s), 3.21 (1 H, d, J ) 9.5 Hz), 3.27 (1 H, dq,
J ) 14.0, 7.0 Hz), 3.29 (1 H, dq, J ) 14.0, 7.0 Hz), 3.43 (1 H,
dq, J ) 14.0, 7.0 Hz), 3.53 (1 H, dq, J ) 14.0, 7.0 Hz), 3.75 (3
H, s), 7.19-7.31 (5 H, m); MS (EI) m/z 304 (M+). Anal.
(C17H24N2O3) C, H, N. 2h : mp 117-119 °C; [R]22 ) +86.1 (c
D
1
0.906, CHCl3); H NMR (500 MHz, CDCl3) 0.89 (3 H, t, J )
7.0 Hz), 1.13 (3 H, t, J ) 7.0 Hz), 1.36 (1 H, ddd, J ) 6.5, 9.0,
9.5 Hz), 1.54 (1 H, dd, J ) 6.0, 6.5 Hz), 1.72 (1 H, dd, J ) 6.0,
9.0 Hz), 2.32 (2 H, br s), 3.11 (1 H, d, J ) 9.5 Hz), 3.28-3.43
(3 H, m), 3.50 (1 H, dq, J ) 14.0, 7.0 Hz), 5.55 (1 H, br s), 7.16
(1 H, br s), 7.20-7.31 (5 H, m); MS (EI) m/z 289 (M+). Anal.
(C16H23N3O2‚0.2H2O) C, H, N.
Bin d in g Assa y. The binding affinity for the NMDA
receptor was investigated according to previously reported
methods.18
(1S,2R)-1-P h en yl-2-[(R)-a m in oca r boxym eth yl]-N,N-d i-
eth ylcyclop r op a n eca r boxa m id e Hyd r och lor id e (2f).
A
solution of 2g (304 mg, 1.0 mmol) in 6 M HCl (8 mL) was
stirred at room temperature for 3 days. The mixture was
evaporated, and the residual powders were washed with Et2O.
The powders were treated with CHCl3/hexane/benzene to give
In h ibitor y Effects on th e Up ta k e of 5-HT. The assay
was investigated according to the previously reprted method.9d
Assa y w ith Volta ge-Cla m p ed Oocytes. The blocking
effects of the compound were investigated on glutamate
receptor subtypes expressed by Xenopus oocytes injected with
total brain mRNA of the Wistar-strain rat under two-electrode
voltage-clamp conditions.19 Electrodes (∼1 MΩ) were filled
with 3 M KCl, and the ground electrode was a salt bridge.
Oocytes were perfused with a bathing solution (in mM: 96
NaCl, 2 KCl, 1 CaCl2, 10 HEPES/NaOH, pH 7.4). Selective
agonists for glutamate receptor subtypes and other neu-
rotransmitter receptors were added to the bathing solution in
the absence or presence of the compound. The stimulation
intervals were ∼2 min. NMDA was added to the bathing
solution together with 10 µM glycine unless otherwise noted.
The NMDA-induced currents consisted of a transient current
and a following steady-state current. The magnitude of the
steady-state current at a holding potential of -50 mV was
measured as the magnitude of the responses. The magnitude
of the responses to given test solutions was calculated relative
to the mean of the magnitude of the responses to 30 µM NMDA
in the absence of blockers (control responses) determined
before and after the application of the test solutions. When
the magnitude of the control responses changed by more than
20% in the experimental period of 60 min, the preparation was
discarded.
white crystals of 2f as a hydrochloride (215 mg, 66%): mp
1
159-161 °C; [R]21 ) +67.3 (c 0.545, MeOH); H NMR (500
D
MHz, CDCl3) 0.86 (3 H, t, J ) 7.0 Hz), 1.07 (3 H, t, J ) 7.0
Hz), 1.36 (1 H, ddd, J ) 6.5, 9.0, 11.0 Hz), 1.65 (1 H, dd, J )
6.0, 6.5 Hz), 2.14 (1 H, dd, J ) 6.0, 9.0 Hz), 3.23-3.42 (4 H,
m), 3.57 (1 H, d, J ) 11.0 Hz), 7.26-7.37 (5 H, m), 8.53 (3 H,
br s), 13.8 (1 H, br s); MS (EI) m/z 290 (M+). Anal. (C16H23
ClN2O3‚0.2H2O) C, H, N.
-
(1S,2R)-1-P h en yl-2-[(S)-(ter t-b u t oxyca r b on yla m in o)-
(m et h oxyca r b on yl)m et h yl]-N,N-d iet h ylcyclop r op a n e-
ca r boxa m id e (14). A solution of 2g (152 mg, 0.50 mmol) and
Boc2O (0.17 mL, 0.75 mmol) in MeCN (5 mL) was stirred at
room temperature for 1.5 h. After water was added, the
mixture was concentrated in vacuo and then extracted with
CHCl3. The organic layer was washed with brine, dried (Na2-
SO4), and evaporated. The residue was purified by column
chromatography (silica gel; hexane/EtOAc, 1:1) to give 14 (194
mg, 96% as a foam): 1H NMR (500 MHz, CDCl3) 0.68 (3 H, br
s), 1.10 (3 H, t, J ) 7.1 Hz), 1.33-1.46 (m, 11 H), 1.88 (1 H,
dd, J ) 7.2, 15.8 Hz), 3.22-3.34 (2 H, m), 3.48-3.58 (2 H, m),
3.74 (1 H, m), 3.83-3.89 (2 H, m), 5.13 (1 H, br s), 5.57 (1 H,
br s), 3.79 (s, 3 H), 7.20-7.31 (5 H, m); MS (EI) m/z 376 (M+).
Anal. (C22H32N2O5‚0.3H2O).
(1S,2R)-1-P h en yl-2-[(S)-1-(ter t-bu toxyca r bon yla m in o)-
2-h yd r oxyeth yl]-N,N-d ieth ylcyclop r op a n eca r boxa - m id e
(15). A hexane solution of DIBAL-H (0.93 M, 0.56 mL, 0.52
mmol) was added to a solution of 14 (70 mg, 0.17 mmol) in
THF (4 mL) at -78 °C, and the mixture was stirred at the
same temperature for 1 h and then at room temperature for
15 h. After being quenched with aqueous saturated NH4Cl (5
mL), the mixture was concentrated in vacuo. The mixture was
partitioned between EtOAc and H2O, and the organic layer
was separated and washed with brine, dried (Na2SO4), evapo-
rated, and purified by column chromatography (silica gel;
CHCl3/MeOH, 10:1) to give 15 (39 mg, 60% as a foam): 1H
Ack n ow led gm en t. We are grateful to Drs. D. Mo-
chizuki and R. Tsujita for their help and encouraging
discussions. We are also grateful to Daiso Co., Ltd. for
their gift of chiral epichlorohydrins. This investigation
was supported in part by a Grant from the Ministry of
Education, Science, Sports, and Culture of J apan.
Refer en ces
(1) Collingridge, G. L.; Watkins, J . C. The NMDA Receptor; IRL
Press at Oxford University Press: Oxford, U.K., 1994.

3514 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 18
Shuto et al.
(2) (a) Wong, E. H. F.; Kemp, J . A.; Priestley, T.; Knight, A. R.;
Woodruff, G. N.; Iversen, L. L. The anticonvulsant MK-801 is a
potent N-methyl-D-aspartate antagonist. Proc. Natl. Acad. Sci.
U.S.A. 1986, 83, 7104-7108. (b) Gill, R.; Foster, A. C.; Woodruff,
G. N. Systemic administration of MK-801 protects against
ischaemia-induced hippocampal neurodegeneration in the gerbil.
J . Neurosci. 1987, 7, 3343-3349. (c) Anis, N. A.; Berry, S. C.;
Burton, M. R.; Lodge, D. The dissociative anaesthetics, ketamine
and phencyclidine, selectively reduce excitation of central mam-
malian neurones by N-methylaspartate. Br. J . Pharmacol. 1983,
79 565-575. (d) Huettner, J . E.; Beam. B. P. Block of N-methyl-
D-aspartate-activated current by the anticonvulsant MK-801:
Selective binding to open channels. Proc. Natl. Acad. Sci. U.S.A.
1988, 85, 1307-1311.
(3) Hutchison, A. J .; Williams, M.; Angst, C.; de J esus, R.; Blan-
chard, L.; J ackson, R. H.; Wilusz, E. J .; Murphy, D. E.; Bernard,
P. S.; Schneider, J .; Campbell, T.; Guida, W.; Sills, M. A.
4-(Phosphonoalkyl)- and 4-(phosphonoalkenyl)-2-piperidinecar-
boxylic acids: synthesis, activity at N-methyl-D-aspartic acid
receptors, and anticonvulsant activity. J . Med. Chem. 1989, 32,
2171-2178.
(4) (a) Willetts, J .; Balster, R. L.; Leander, J . D. The behavioral
pharmacology of NMDA receptor antagonist. Trends Pharmacol.
Sci. 1990, 11, 423-428. (b) Tricklebank, M. D.; Singh, L.; Oles,
R. J .; Preston, C.; Iversoen, S. D. The behavioral effects of MK-
801: a comparison with antagonists acting noncompetitively at
the NMDA receptor. Eur. J . Pharmacol. 1989, 167, 127-135.
(c) Olney, J . W.; Labruyere, J .; Price, M. T. Pathological changes
induced in cerebrocortical neurons by phencyclidine and related
drugs. Science 1989, 244, 1360-1362.
(5) (a) Meldrum, B. S.; Croucher, M. J .; Czuczwar, S. J .; Collins, J .
F.; Curry, K.; J oseph, M.; Stone, T. W. A comparison of the
anticonvulsant potency of (()-2-amino-5-phosphonopentanoic
acid and (()-2-amino-7-phosphonoheptanoic acid. Neuroscience
1983, 9, 925-930. (b) Hogan, M. J .; Gjedde, A.; Hakem, A. M.
In vivo distribution of CGS-19755 within brain in a model of
focal cerebral ischemia. J . Neurochem. 1992, 58, 186-191.
(6) Compound (()-1 had previously been called “midalcipran”.
However, since the clinical studies began, the name “milnacip-
ran” has generally been used for (()-1.
(7) (a) Briley, M. Midalcipran hydrochloride. Drugs Future 1986,
11, 21-23. (b) Moret, C.; Charveron, M.; Finberg, J . P. M.;
Cozinier, J .; M. Briley, M. Biochemical profile of midalcipran
(F-2207). (Z)-1-Phenyl-1-diethylaminocarbonyl-2-aminomethyl-
cyclopropane hydrochloride, a potential fourth generation anti-
depressant drug. Neuropharmacology 1985, 24, 1211-1219. (c)
Bonnaud, B.; Cousse, H.; Mouzin, G.; Briley, M.; Stenger, A.;
Fauran, F.; Couzinier, J .-P. 1-Aryl-2-(aminomethyl)cyclopropan-
ecarboxylic acid derivatives. A new series of potential antide-
pressants. J . Med. Chem. 1987, 30, 318-325. (d) Ansseau, M.;
Papart, P.; Troisfontaines, B.; Bartholome, F.; Bataille, M.;
Charles, G.; Schittecatte, M.; Darimont, P.; Devoitille, J . M.
Controlled comparison of milnacipran and fluoxetine in major
depression. Psychopharmacology 1994, 114, 131-137. (e) Arti-
gas, F. Selective serotonin/noradrenaline reuptake inhibitors
(SNRIs). Pharmacology and therapeutic potential in the treat-
ment of depressive disorders. CNS Drugs, 1995, 4, 79-89.
(8) Shuto, S.; Takada, H.; Mochizuki, D.; Tsujita, R.; Hase, Y.; Ono,
S.; Shibuya, N.; Matsuda, A. (()-(Z)-2-Aminomethyl-1-phenyl-
cyclopropanecarboxamide derivatives as a new prototype of
NMDA receptor antagonists. J . Med. Chem. 1995, 38, 2964-
2968.
nucleophilic addition to cyclopropyl carbonyls: the facial selec-
tivity in the cyclopropyl ketones is opposite to that in the
corresponding aldehyde. Tetrahedron Lett. 1996, 37, 221-224.
(c) Shuto, S.; Ono, S.; Hase, Y.; Kamiyama, N.; Matsuda, A.
Synthesis of (+)- and (-)-milnaciprans and their conformation-
ally restricted analogs. Tetrahedron Lett. 1996, 37, 641-644. (d)
Shuto, S.; Ono, S.; Hase, Y.; Ueno, Y.; Noguchi, T.; Yoshii, K.;
Matsuda, A. Synthesis and biological activity of conformationally
restricted analogs of milnacipran: (1S,1R)-1-Phenyl-2-[(S)-1-
aminopropyl]-N,N-diethylcyclopropanecarboxamide, an efficient
noncompetitive N-methyl-D-aspartic acid receptor antagonist. J .
Med. Chem., 1996, 39, 4844-4852.
(10) Ogawa, K.; Yamashita, K., unpublished results.
(11) The enantiomeric purity of the lactone
4 was 96% ee, as
measured by HPLC with a Chiralcel-OJ column (Daicel Chemi-
cal Co., Ltd.): see ref 9a.
(12) The R,S indication for the configuration at the 1′-position of
compounds can be changed depending on the 1′-substituents (R),
irrespective of their absolute structure.
(13) Yamamoto, I.; Sekine, M.; Hata, T. One-step synthesis of
5′-azide-nucleosides. J . Chem. Soc., Perkin Trans. 1 1980, 306-
310.
(14) Mungall, W. S.; Greene, G. L.; Heavner, G. A.; Letsinger, R. L.
Use of the azido group in the synthesis of 5′-terminal aminode-
oxythymidine oligonucleotides. J . Org. Chem. 1975, 40, 1659-
1662.
(15) Nicholas, K. M. Chemistry and synthetic utility of cobalt-
complexed propargyl cations. Acc. Chem. Res. 1987, 20, 207-
215.
(16) (a) Haywood-Farmer, J . Long-range interactions of cyclopropyl
groups with carbonium ion centers. Chem. Rev. 1974, 74, 315-
350. (b) Wilcox, C. F.; Loew, M. M.; Hoffmann, R. Why a
cyclopropyl group is good at stabilizing a cation but poor at
transmitting substituent effects. J . Am. Chem. Soc. 1973, 95,
8192-8193.
(17) Kobayashi, S.; Tsuchiya, Y.; Mukaiyama, T. A facile synthesis
of cyanohydrin trimethylsilyl ethers by the addition reaction of
trimethylsilylcyanide with aldehydes under basic conditions.
Chem. Lett. 1991, 537-540.
(18) Asano, T.; Ikegaki, I.; Satoh, S.; Mochizuki, D.; Hidaka, H.;
Suzuki, Y.; Shibuya, M.; Sugita, K. Blockade of intracellular
actions of calcium may protect against ischemic damage to the
gerbil brain. Br. J . Pharmacol. 1991, 103, 1935-1938.
(19) (a) Yoshii, K.; Kurihara, K.; Inward rectifier produced by
Xenopus oocytes injected with mRNA extracted from carp
olfactory epithelium. Synapse 1989, 3, 234-238. (b) Yoshii, K.;
Yu, L.; Mixter-Mayne, K.; Davidson, N.; Lester, H. A. Equilib-
rium properties of mouse-torpedo acetylcholine receptor hybrids
expressed in Xenopus oocyte. J . Gen. Physiol. 1987, 90, 553-
557. (c) Kawano, H.; Sashihara, S.; Mita, T.; Ohno, K.; Kawa-
mura, M.; Yoshii, K. Phenytoin, an antiepileptic drug, competi-
tively blocked non-NMDA receptor produced by Xenopus oocytes.
Neurosci. Lett. 1994, 166, 183-186.
(20) (a) Ford, L. M.; Sanberg, P. R.; Norman, A. B.; Fogelson, M. H.
MK-801 prevents hippocampal neurodegeneration in neonatal
hypoxic-ischemic rats, Arch Neurol. 1989, 46, 1090-1096. (b)
Foster, A. C.; Gill, R.; Woodruff, G. N. Neuroprotective effects
of MK-801 in vivo: selectivity and evidence for delayed degen-
eration mediated by NMDA receptor activation, J . Neurosci.
1988, 8, 4745-4754. (c) Olney, J .; Price, M.; Salles, K. S.;
Labruyere, J .; Frierdich, G. MK-801 powerfully protects against
N-methyl aspartate neurotoxicity, Eur. J . Pharmacol. 1987, 141,
357-361. (d) Kroemer, R. T.; Koutsilieri, K.; Hecht, P.; Liedl,
K. R.; Riederer, P.; Kornhuber, J . Quantitative analysis of the
structural requirements for blockade of the N-methyl-D-aspar-
tate receptor at the phencyclidine binding site. J . Med. Chem.
1998, 41, 393-400.
(9) (a) Shuto, S.; Ono, S.; Hase, Y.; Kamiyama, N.; Takada, H.;
Yamashita, K.; Matsuda, A. Conformational restriction by
repulsion between adjacent substituents of a cyclopropane
ring: design and enantioselective synthesis of 1-phenyl-2-(1-
aminoalkyl)-N,N-diethylcyclopropanecarboxamides as potent
NMDA receptor antagonists. J . Org. Chem. 1996, 61, 915-923.
(b) Ono, S.; Shuto, S.; Matsuda, A. Highly stereoselective
J M980238M