Drug design, in vitro pharmacology, and structure-activity relationships of 3-acylamino-2-aminopropionic acid derivatives, a novel class of partial agonists at the glycine site on the N-methyl-D-aspartate (NMDA) receptor complex

Article abstract of DOI:10.1021/jm900363q

Retaining agonistic activity at the glycine coagonist site of the NMDA receptor in molecules derived from glycine or D-serine has proven to be difficult because in the vicinity of the α-amino acid group little substitution is tolerated. We have solved thi

Full text of DOI:10.1021/jm900363q

J. Med. Chem. 2009, 52, 5093–5107 5093
DOI: 10.1021/jm900363q
Drug Design, in Vitro Pharmacology, and Structure-Activity Relationships of
3-Acylamino-2-aminopropionic Acid Derivatives, a Novel Class of Partial Agonists at the
Glycine Site on the N-Methyl-D-aspartate (NMDA) Receptor Complex
Stephan Urwyler,§ Philipp Floersheim, Bernard L. Roy, and Manuel Koller*
Novartis Institutes for BioMedical Research, Neuroscience, 4056 Basel, Switzerland. §Current affiliation: Department of Chemistry and
Biochemistry, University of Berne, Switzerland.
Received March 23, 2009
Retaining agonistic activity at the glycine coagonist site of the NMDA receptor in molecules derived
from glycine or D-serine has proven to be difficult because in the vicinity of the R-amino acid group little
substitution is tolerated. We have solved this problem by replacing the hydroxy group of D-serine with
an amido group, thus keeping the hydrogen donor function and allowing for further substitution and
exploration of the adjacent space. Heterocyclic substitutions resulted in a series of 3-acylamino-
2-aminopropionic acid derivatives, with high affinities in a binding assay for the glycine site. In a
functional assay assessing the activation of the glycine site, these compounds displayed a wide range of
intrinsic efficacies, from antagonism to a high degree of partial agonism. Structure-activity relation-
ships reveal that lipophilic substituents, presumably filling an additional hydrophobic pocket, are
accepted by the glycine site, provided that they are separated from the R-amino acid group by a short
linker.
Introduction
use of partial agonists at this site might be more promising
than that of “silent” antagonists, since such compounds
would tune down receptor activity to a certain level rather
than completely blocking it. Moreover, the concept of partial
agonism might be useful not only in clinical indications
requiring inhibition of NMDA receptor function (e.g., epi-
lepsy, neurodegeneration) but also in others, such as possibly
schizophrenia, where a certain stimulation might be desirable
(for reviews, see refs 19 and 20).
Retaining agonistic activity in molecules derived from
glycine 1 (Figure 1) has proven to be difficult because of
severe space restrictionsfor substituents atorinthe immediate
vicinity of the amino acid function.21-23 D-Alanine 2,21 ami-
nocyclopropanecarboxylic acid (ACC, 3), and D-serine 4 are
among the few examples of full agonists at the glycine site.
However, substitution at the β-C atom of D-alanine 2 destroys
its affinity for the glycine site (2d-f) (with the noteworthy
exception of some E-γ-substituted vinylglycine derivatives,
2a-c;24 see below). Similarly, stepwise expansion of the ring
size from ACC 3 to aminocyclohexanecarboxylic acid drasti-
cally reduces not only the affinity (3a-c) but also the intrinsic
efficacy of compounds at the glycine site.23 Also, no substitu-
tion is tolerated at or in the vicinity of the hydroxy group of
D-serine 4 (4a-c).
The N-methyl-D-aspartate (NMDAa) receptor, a subtype
of the ionotropic glutamate receptor family, is involved in key
physiological functions, such as learning, memory and devel-
opmental processes on the one hand, but also in a variety of
disease states (epilepsy, acute and chronic neurotoxicity) on
the other hand. It is therefore believed to have a considerable
potential as a site for therapeutic intervention.1 The receptor
comprises multiple functional domains interacting with each
other, thus allowing a subtle regulation of receptor activity
and efficacy. In particular, glycine has been shown to be a
mandatory coagonist necessary for the opening of NMDA
receptor ion channels.2,3 The glycine recognition site, located
on the NR1 subunit of the NMDA receptor complex, can be
distinguished from the inhibitory glycine receptor by its
insensitivity to strychnine. It has attracted considerable inter-
est as a potential drug target, as an alternative to the NMDA
recognition site itself.1,4,5 A great number of glycine site
antagonists from different structural classes have been
described in recent years.6-17 However, serious side effects
(e.g., memory impairment, neurotoxic or psychotomimetic
effects) have been observed with the use of competitive or
channel blocking NMDA receptor antagonists in animal
models and/or in human clinical trials.18 Since glycine is an
obligatory coagonist, one would expect the same to happen
from an antagonist blocking the glycine site. Therefore, the
The E-γ-substituted vinylglycine derivatives 2a-c24 retain
agonistic activity at the glycine site very likely because of the
slender shape of the 3,4-double bond in the close neighbor-
hood of the R-C atom (compare with the inactive compound
2e with a 4,5-double bond!). The activity of these compounds
suggests the existence of an extra hydrophobic binding pocket
at the glycine site, in addition to the hydrophilic pocket with
which the amino acid function interacts. On the other hand,
it seems clear that the inactivity of the O-methyl derivative of
*To whom correspondence should be addressed. Phone: þ41 61 324
66 08. Fax: þ44 61 696 24 55. E-mail: manuel.koller@novartis.com.
a Abbreviations: ACBC, 1-aminocyclobutane carboxylic acid; ACC,
1-aminocyclopropanecarboxylic acid; 5,7-DCK, 5,7-dichlorokynurenic
acid; DCS, D-cycloserine; NMDA, N-methyl-D-aspartate; Z, benzyloxy-
carbonyl.
r
2009 American Chemical Society
Published on Web 07/30/2009
pubs.acs.org/jmc

5094 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 16
Urwyler et al.
Figure 1. Affinities (pKi values) of different R-substituted derivatives of glycine 1, D-alanine 2, 1-aminocyclopropanecarboxylic acid (ACC, 3),
and D-serine 4 for the glycine recognition site on the NMDA receptor. The values were obtained from a [3H]glycine binding assay performed as
described in the Experimental Section, or they were drawn from ref 24 (in the case of compounds 2a-c) or from ref 22 (in the cases of
compounds 4b and 4c).
Chart 1
Several starting materials RCOX (Scheme 1) were com-
mercial, such as the acid chlorides (X=Cl) for target com-
pounds 8, 10, 11 or the carboxylic acids (X = OH) used for
target compounds 17, 19-22, 38 (Tables 1-3). Most others
were known from literature and synthesized following the
published procedures, as for compounds 9, 12-16, 18, 24, 25,
30, 32, 34-37, 46, 51 (Tables 1-4). For the target compounds
33, 47, 48, and 50, the intermediate indole carboxylic
acids 33c, 47c, 48b, and 50d were not known in the literature
but could be prepared from the hydrazones 33a, 47a
(Scheme 2), and 50b (obtained in four steps from 1-chloro-
4-iodo-2-nitrobenzene as depicted in Scheme 3), respectively,
by the Fischer indole synthesis. The cyclization of 50b deliv-
ered the desired product 50c, together with the abnormal
Fischer indolization product28 50c0 (Scheme 3).
The 2-ethylphenyl-substituted naphthalenecarboxylic acid
28a (for target compound 28) and most of the 4-aryl sub-
stituted benzothiophene-2- (for 27, 31) and indole-2-car-
boxylic acids (for 26, 40-42, 44, 45) were prepared from
their bromo counterparts by Pd-catalyzed coupling with the
corresponding arylboronic acids (Scheme 4).
R-26 was prepared in good yields as described in Scheme 5.
Benzyl ester R-5c was synthesized from R-5a26 (Chart 2) and
condensed with the carboxylic acid 26a to the Z-protected
aminoester 26b, enabling deprotection of the product in one
step by hydrogenolysis under mild conditions.
D-serine (4c in Figure 1) is not due to steric hindrance but to
the removal of the hydrogen bond donor function of the
hydroxy group, which is involved in the agonism by D-serine 4
at the glycine site.22,25 Therefore, we started a synthetic
program based on the replacement of the hydroxy group in
D-serine 4 by an amido group, leading to 3-acylamino-
2-aminopropionic acid derivatives (Chart 1) and thus preser-
ving its potential for hydrogen bond formation and allowing
for further substitution to explore the space adjacent to this
functional group.
Chemistry
Most of the 3-acylamino-2-aminopropionic acid derivatives
(III) were prepared as depicted in Scheme 1. Condensation of
the appropriate carboxylic acid or acid chloride (I) with 3-
amino-2-benzyloxycarbonylaminopropionic acid methyl ester
5b (either racemic or, for the optically active target molecules
(III), as the enantiomersR-5b26 and S-5b27) gave the esters (II).
Deprotection was accomplished by saponification with aqu-
eous base followed by removal of the Z-group either by
hydrogenolysis or by reaction with iodotrimethylsilane.
The racemic diaminopropionic acid rac-5a and its methyl
ester rac-5b were prepared starting from rac-asparagine and
using the procedure described for R-5b26 (structures shown in
Chart 2).
For target compound 39, the intermediate 22a was used as
coupling partner for phenylboronic acid in the Pd-catalyzed
reaction (Scheme 6), generating in one step carboxylic acid
39a that was subsequently transformed into 39. On the other
hand, for 43 the stannylated indole ester 43a was coupled with
1-benzyl-2-bromobenzene to 43b (Scheme 7). Saponification
to 43c and further reaction according to Scheme 1 delivered
43.
Compounds 6 and 7 (Table 1) were obtained by reacting
rac-5a with methyl chloroformate and ethyl isocyanate,
respectively, followed by deprotection.
Several functional group transformations in positions 3 and 4
of the indole and benzothiophene scaffolds were accomplished

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 16 5095
Scheme 1a
a (i) For X = Cl: CH2Cl2-pyridine 1:1, -10 °C. (ii) For X = OH: 1,10-carbonyldiimidazole, pyridine, room temp. (iii) NaOH aq, THF or dioxane,
room temp. (iv) H2/Pd-C, THF-H2O 1:1, room temp. (v) ISiMe3, CH2Cl2, 0 °C.
Chart 2
a comparison of the racemic compound rac-22 (pKi = 6.8)
and its R-enantiomer R-22 (pKi = 7.0) with the unsubstituted
compound R-14 (pKi = 5.3, Table 1). Other ring systems were
much less affected by such a bromo substituent (cf. benzothio-
phenes 17/24; 2-naphthyl derivatives 20/25). In the case of the
3-indolyl compound 15, bromo substitution in position 4 of
the indole ring resulted in a strong loss of affinity (23). The
compound with the highest affinity at the glycine site prepared
in this work was amino acid 26 with a 2-ethylphenyl group
attached in position 4 of the indole nucleus, with a pKi value of
7.6 for the R-enantiomer. Similar to the bromine atom in this
position, the 2-ethylphenyl group did not induce or only
slightly induced an increase in affinity when it was attached
to other ring systems, e.g., as illustrated by the pairwise
comparison of the benzothiophene derivatives 17 (pKi=5.9)
and 27 (pKi=6.2) or the 2-naphthyl derivatives 20 (pKi = 5.9)
and 28 (pKi=5.7).
Therefore, besides a few more benzothiophenes (29, 30, 31),
the derivatization of molecules with the carbonyldiaminopro-
pionic acid side chain attached in position 2 of an indole ring
was further pursued. The results from several single enantio-
mers suggested that the R-form has the strongest activity at
the glycine site (cf. Table 1, R-14/S-14; Table 2, R-22/rac-22,
R-26/rac-26), indicating that the binding mode of the amino
acid part of the 3-acylamino-2-aminopropionic acids dis-
cussed here resembles that of the parent compound D-serine.
For simplicity, however, in most cases only the racemic
compounds were prepared (compounds 9 and 10 only as
R-enantiomers). Of course, this needs to be considered when
comparisons of their affinities with those of R-enantiomers
are made.
delivering the desired substituted acids for the condensation
reaction with rac-5b. These reactions are summarized in
Scheme 8. 4-Bromo-1H-indole-3-carboxylic acid 23b, used for
target compound 23, was prepared from 4-bromo-1H-indole by
trichloroacetylation to 23a and hydrolysis of the trichloro-
methyl group under mild conditions to avoid decarboxyla-
tion.29 4-Cyanobenzothiophene-2-carboxylic acid 29c, the
precursor for target compound 29, was obtained by cyanation
of the bromo ester 29a with cuprous cyanide at high tempera-
ture, followed by selective hydrolysis of the ester group of 29b.
Target compound 49, carrying a nitro group in position 3 of the
indole nucleus, required the indole acid 49a that was obtained
from nitration of 4-bromo-1H-indole-2-carboxylic acid.
To obtain the N-methylated indole derivatives 52 and 53,
4-bromo-1H-indole-2-carboxylic acid methyl ester was
N-methylated to 52a. Pd-catalyzed coupling of 52a with
2-ethylphenylboronic acid gave 53a. Saponification of the
methyl esters and condensation with rac-5b followed by
deprotection provided the two target compounds (Scheme 9).
Results
Some prototype 3-acylamino-2-aminopropionic acid deri-
vatives with various acyl substituents were tested in radioli-
gand binding assays for the strychnine-insensitive glycine site
at the NMDA receptor complex ([3H]glycine) and, in most
cases, also the glutamate recognition site ([3H]CGP39653).
The resulting pKi values are given in Table 1. Whereas the
carbamate 6 and the urea 7 were virtually inactive, amino
acids with aroyl or heteroaroyl groups at the β-nitrogen atom
displayed micromolar affinities for the glycine site. The
activity of compounds with single aromatic ring substitution
in this position (8-11) was moderate, whereas bicyclic aro-
matic rings attached to the carbonyl group conferred affinities
in the low micromolar range to compounds 13, R-14, 15,
17-21. A methylene spacer between the aromatic ring and the
carbonyl group (16) or the attachment of the ring system at a
saturated C-atom (compare the inactive indanyl derivative 12
with the active indenyl compound 13) were not tolerated.
Compounds with attachments of various substituents at
bicyclic aromatic ring systems are listed in Tables 2 and 3. A
most promising ring system turned out to be indole: A
bromine atom in position 4 of the indole nucleus significantly
enhanced the binding affinity for the glycine site, as shown by
To determine if the aforementioned position 4 of the indole
ring is the most promising one for further derivatization, the
derivatives with chloro at the four different positions of the
benzene ring were compared: whereas chloro substitution in
positions 5 and 6 of the indole ring (R3, R4, Table 3) did not
enhance the affinities of the compounds for the glycine site
(34, 35), substitution at position 7 (R5, 36, pKi = 5.5) resulted
in a little improvement and in position 4 (R2, 37, pKi = 6.0) in
even more improvement over the unsubstituted R-14 (pKi =
5.3). Substitution with other groups in position 4, such as
methoxy (38), was also beneficial.
Position 3 of the indole ring was varied by attaching alkyl
side chains (R1, Table 3). Whereas methyl (32) increased the
affinity only a little, a significant enhancement of glycine site
affinity was obtained with propyl (33) by a factor of 10,
compared with R-14. Methylation of the indolyl N-atom (R6,
Table 3) had no effect on binding affinity (46, pKi = 4.9) when
compared to the unsubstituted indolyl compound R-14.
Of particular interest was the observed increase of affinity
with the 4-phenyl substituent, leading to the question of
whether substitution of the phenyl ring in 39 was a means to
further strengthen binding affinity. This was indeed the case,

5096 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 16
Urwyler et al.
Table 1. Affinities of 3-Acylamino-2-aminopropionic Acid Derivatives
for the Glycine and Glutamate Recognition Sites of the NMDA
Receptor: Unsubstituted Chains and Ring Systems as R Groups
Table 2. Effect on Affinities of Bromine- and 2-Ethylphenyl Substitu-
tion at Indolyl, Benzothienyl, and Naphthyl in the Acyl Group of the
3-Acylamino-2-aminopropionic Acid Derivatives
a The affinities of the compounds for the glycine and glutamate
recognition sites are expressed as pKi values and were determined in
[3H]glycine and [3H]CGP39653 binding assays as described in the
Experimental Section. The data shown are mean values ( SEM from
at least three independent experiments for [3H]glycine binding (at least
two for [3H]CGP39653 binding).
The binding affinities of rac-22 and rac-26 for the glycine
site could not be improved by various disubstitution patterns
at the indole ring (Table 4). For example, a propyl substituent
in position 3 (R1), which by itself had a positive effect (such as
in 33) compared to R-14, showed no synergistic effect when
combined with a bromo- or 2-ethylphenyl group in position 4
but on the contrary lowered glycine site affinity (compare 47
with rac-22, or 48 with rac-26). The introduction of an
electron-withdrawing group in position 3, such as nitro, which
presumably enhances the NH-donor function of the indolyl
N-atom, reduced the affinity for the glycine site (compare 49
with rac-22). It is also noteworthy that a chloro atom in
position 7 (R5, Table 3), which by itself only slightly improved
the affinity for the glycine site (36 vs R-14), was rather
detrimental when combined with an 2-ethylphenyl substituent
in position 4 (50 vs rac-26). Similarly, whereas a methyl group
a The affinities of the compounds shown were determined in
[3H]glycine and [3H]CGP39653 binding assay as described in the
Experimental Section. The data shown are expressed as pKi values and
represent mean values ( SEM from at least three independent experi-
ments. n.t., not tested.
as alkyl substitution in position 2 of the phenyl ring
produced derivatives with high affinity for the glycine site:
benzyl (43), methyl (40), ethyl (rac-26), propyl (41), and
even the sterically demanding isopropyl group (42) gave a
3- to 100-fold increase of affinity compared to 39, the
optimum being found with ethyl in the already mentioned
compounds rac-26 and R-26. No increase in glycine site
affinity (compared to 39) was obtained with more polar
substituents in the ortho-position of the phenyl ring, such as
in 44 and 45.

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 16 5097
Table 3. Affinities of 3-Acylamino-2-aminopropionic Acid Derivatives for the Glycine and Glutamate Recognition Sites of the NMDA Receptor:
Monosubstituted Bicyclic Ring Systems in the Acyl Group
compd
X
R1
R2
R3
R4
R5
R6
glycinea pKi
glutamatea pKi
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
S
H
H
H
CN
Cl
H
H
H
H
H
Cl
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
Cl
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
Cl
H
H
H
H
H
H
H
H
H
H
5.9 ( 0.04
6.5 ( 0.03
4.7 ( 0.12
5.5 ( 0.15
6.3 ( 0.10
4.7 ( 0.09
4.9 ( 0.17
5.5 ( 0.05
6.0 ( 0.19
5.9 ( 0.07
5.6 ( 0.08
6.5 ( 0.11
7.3 ( 0.11
7.3 ( 0.07
6.0 ( 0.11
5.8 ( 0.04
5.6 ( 0.04
4.9 ( 0.09
<4
<4
S
S
2,4-dichlorophenyl
H
<4.5
n.t.
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
methyl
propyl
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
3.8 ( 0.16
n.t.
n.t.
H
H
H
H
H
Cl
n.t.
H
4.1 ( 0.08
4.1 ( 0.05
4.7 ( 0.03
5.2 ( 0.13
5.4 ( 0.05
5.6 ( 0.10
<5.5
H
OCH3
phenyl
o-tolyl
H
H
H
2-propylphenyl
2-isopropylphenyl
2-benzylphenyl
2-dimethylamino-phenyl
2-methoxyphenyl
H
H
H
H
4.1 ( 0.01
4.3 ( 0.12
n.t.
H
H
methyl
a The affinities of the compounds for the glycine and glutamate recognition sites are expressed as pKi values and were determined in [3H]glycine (N g
3) and [3H]CGP39653 (N g 2) binding assays as described in the Experimental Section. The data shown are mean values ( SEM. n.t.: not tested.
Table 4. Affinities of 3-Acylamino-2-aminopropionic Acid Derivatives
The results in Tables 1-4 show that many of the com-
for the Glycine and Glutamate Recognition Sites of the NMDA
pounds described in this study bind not only to the strychnine-
Receptor: Disubstituted Bicyclic Ring Systems in the Acyl Group
insensitive glycine coagonist site but also, albeit with much
weaker affinities, to the glutamate recognition site, labeled by
[3H]CGP39653. Their activity at this site was not due to
indirect (allosteric) inhibition of [3H]CGP39653 binding via
agonism at the glycine site,30,31 since the glycine antagonist
5,7-dichlorokynurenic acid was added to the assay mixture to
block such an interaction. On the other hand, a limited
selection of 3-acylamino-2-aminopropionic acid derivatives
(9, 10, 17, R-14, 15, 20, R-22, 24, R-26, 27, 29, 31, 33) was
compd R1
R2
R5 R6 glycinea pKi glutamatea pKi
tested in assays for the ion channel site ([3H]MK-801 binding
measured under equilibrium conditions) or the polyamine site
associated with the NMDA receptor ([3H]spermidine
binding). None of them showed any activity (data not shown).
When measured under nonequilibrium conditions, the stimu-
lation of the binding of [3H]MK-801 can be used as a functional
assay for compounds acting at the NMDA receptor complex.
47
48
49
50
51
52
53
propyl Br
propyl 2-ethylphenyl H
nitro Br
H
H
H
H
H
H
6.5 ( 0.02 4.2 ( 0.25
6.8 ( 0.07 5.3 ( 0.04
6.1 ( 0.05 4.9 ( 0.12
6.9 ( 0.06 <4.5
H
H
H
H
H
2-ethylphenyl Cl
Cl
Br
Cl
H
6.8 ( 0.15 4.0 ( 0.11
methyl 5.0 ( 0.04 n.t.
32,33
2-ethylphenyl H methyl 3.9 ( 0.05 n.t.
The rate of association of the noncompetitive NMDA
a The affinities for the glycine and glutamate recognition sites of the
receptor antagonist [3H]MK-801 to its binding site inside the
receptor-operated ion channel depends on the degree of channel
opening, which is, in turn, a function of the activation of the
glutamate and glycine sites on the receptor. In our experiments,
stimulation of the glycine coagonist site was monitored under
conditions where the transmitter recognition site was maximally
activated by a saturating concentration of glutamate. In this
setup glycine, ACC, and both enantiomers of serine stimulated
the binding of [3H]MK-801 to the same maximal level, with the
expected potencies and enantioselectivity and with D-serine being
about a 100-fold more potent than L-serine (data not shown in
detail). The well established partial agonist D-cycloserine,34 on its
own, stimulated the binding of [3H]MK-801 maximally by about
50% of the effect obtained with a saturating concentration of
compounds listed in the table are expressed as pKi values and were
determined in [3H]glycine (N g 3) and [3H]CGP39653 (N g 2) binding
assays as described in the Experimental Section. The data shown are
mean values ( SEM. n.t.: not tested.
at the indolyl N-atom had no effect in the otherwise
unsubstituted molecule 46, it dramatically decreased the
affinities of the compounds bearing a bromo- or 2-ethyl-
phenyl group in position 4 (52, 53). In other cases, how-
ever, appropriate disubstitution could well increase glycine
affinity when compared to the corresponding monosub-
stitutions. For example, the 4,7-dichloro compound 51 had
a higher affinity than the two monochloro derivatives 36
and 37.

5098 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 16
Urwyler et al.
Scheme 2a
a (i) H2SO4, ethanol, reflux; (ii) 2-ethylphenylboronic acid, Pd(PPh3)4, 2 M Na2CO3, ethanol, toluene, reflux; (iii) aqueous base, methanol or ethanol,
reflux; (iv) according to Scheme 1.
Scheme 3a
a (i) 2-Ethylphenylboronic acid, Pd(PPh3)4, 2 M Na2CO3, ethanol, toluene, reflux; (ii) Fe, NH4Cl, H2O, methanol, reflux; (iii) NaNO2, HCl aq, 0 °C,
5 h, then SnCl2, HCl conc, 30 min; (iv) 2-oxopropionic acid methyl ester, H2SO4 conc, H2O, ethanol, 10 °C; (v) polyphosphoric acid, 140 °C; (vi) 1 M
NaOH, methanol, reflux; (vii) according to Scheme 1.
Scheme 4a
a (i) Pd(PPh3)4 cat., 2 M Na2CO3, ethanol, toluene, reflux; (ii) according to Scheme 1.
glycine (Figure 2). In the presence of a maximally active glycine
concentration, however, D-cycloserine inhibited [3H]MK-801
binding down to a level similar to the maximal activation that
it produced when added alone. The same typically partial
agonistic behavior was found with compound R-26 (Figure 2).
However, the determination of the maximal degree of
agonistic effect (“efficacy”) of a compound in this assay is
complicated by the fact that under basal conditions (i.e.,
without addition of an agonist), the glycine site is already
activated to some extent by endogenous glycine presumably

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 16 5099
Scheme 5a
a (i) R-5c, 1,10-carbonyldiimidazole, pyridine, room temp (81%); (ii) H2, Pd-C, AcOH, room temp (89%).
Scheme 6a
a (i) Phenylboronic acid, Pd(PPh3)4, 2 M Na2CO3, ethanol, toluene, reflux; (ii) ISiMe3, CH2Cl2, 0 °C.
Scheme 7a
a (i) (SnBu3)2, Pd(PPh3)4, 2 M Na2CO3, ethanol, toluene, reflux; (ii) 1-benzyl-2-bromobenzene, PdCl2(PPh3)2, DMF, 120 °C; (iii) 1 M NaOH,
methanol, reflux; (iv) according to Scheme 1.
Scheme 8a
a (i) Cl3CCOCl, pyridine, dioxane, reflux; (ii) 1 M KOH, ethanol, room temp; (iii) CuCN, N-methylpyrrolidone, reflux; (iv) 1 equiv of Na2CO3, H2O,
methanol, dioxane, reflux; (v) HNO3, acetic anhydride, 0 °C; (vi) according to Scheme 1.
present in the membrane preparation.35 This phenomenon is
illustrated by the observation that [3H]MK-801 binding under
basal conditions is inhibited by the glycine site antagonist 5,7-
dichlorokynurenic acid (Figure 3).36 For this reason, we
calculated the efficacies of compounds with reference to a
“zero level” which was determined in the presence of a
maximally active concentration (30 μM) of 5,7-dichloroky-
nurenic acid (Figure 3). According to this procedure, a
compound like cis-methyl-HA966 (L-68741437), which also
inhibits [3H]MK-801 binding under basal conditions but not
to the same extent as the pure antagonist 5,7-dichlorokynure-
nic acid, is recognized as a partial agonist with an efficacy of
14%, which is less than the degree of basal receptor activation
by endogenous glycine (Figure 3).
Table 5 gives an overview of the potencies (pEC50 or pA2
values) and efficacies of a selection of 3-acylamino-2-amino-
propionic acid derivatives in this assay. Their potencies
showed an excellent correlation with the affinities found in

5100 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 16
Urwyler et al.
Scheme 9a
a (i) NaH, MeI, DMF, room temp; (ii) 2-ethylphenylboronic acid, Pd(PPh3)4, 2 M Na2CO3, ethanol, toluene, reflux; (iii) 1 M NaOH, methanol, reflux;
(iv) according to Scheme 1.
Figure 3. Method of determining the efficacy of partial agonists
acting at the glycine site. The nonequilibrium binding of [3H]MK-
801 under basal conditions (no agonist added, middle dashed line) is
already activated to about 25% relative to the maximally possible
stimulation (with a saturating concentration of glycine, upper
dashed line) because of the endogenous glycine present. This is
demonstrated by an inhibition of [3H]MK-801 binding under basal
conditions by the glycine site antagonist 5,7-dichlorokynurenic acid
(5,7-DCK, b). The efficacies (maximal stimulation obtained, rela-
tive to maximal effect of glycine) of partial agonists were therefore
determined relative to a zero level (lower dashed line) measured in
the presence of a saturating concentration of 5,7-DCK. This
principle is illustrated with ACBC, having an efficacy of 53%. A
low-efficacy partial agonist such as L-687414 (2) having an efficacy
that is lower than the basal activity because of endogenous glycine
will result in a partial inhibition of this basal activity.
Figure 2. Partial agonism at the glycine site by D-cycloserine (top)
and compound R-26 (bottom). The potentiation of [3H]MK-801
binding was measured under nonequilibrium conditions in the
presence of a maximally active concentration of glutamate, as
described in the methods section. The compounds were tested either
alone (squares) or in combination with 1 μM glycine (triangles).
Basal levels were determined in the absence of any glycine ligand.
Maximal stimulation was measured in the presence of 3 μM glycine,
and 10 μM glutamate was present in all samples.
compounds had only relatively weak binding affinities for
the glycine site (Table 1). On the other hand, the compound
with the highest binding affinity, R-26, had somewhat less
intrinsic efficacy (about 50%). Interestingly, some com-
pounds with very low efficacies (e.g., 29 around 20%) or even
fully antagonistic properties (31), were also found (Table 5).
In the search of a pharmacophore model accounting for the
affinities of the 3-acylamino-2-aminopropionic acids at the
strychnine-insensitiveglycine site, weassessedconformational
preferences of R-26, 27, and 53. The N-methylamides 54-56
were used as respective model systems for molecular modeling
(Table 6). On the basis of ab initio energy differences, the s-cis
conformation of the indoles 54 and 55 is favored by 4.7 and
5.0 kcal/mol, respectively, whereas no significant energy
difference between the s-cis and s-trans conformations results
with the benzothiophene 56. Therefore, we assume that the
the [3H]glycine binding assay (slope = 1.0, r2=0.82, p<0.01).
On the other hand, their intrinsic efficacies (maximal degrees
of stimulation) did not go in parallel with their binding
affinities. For example, the compound with the highest effi-
cacy (about 80% of the maximal effect of glycine, Table 5)
found in the series was the furane analogue 10, and the
corresponding pyrrolo and thiophene analogues 9 and 11 also
stimulated [3H]MK-801 binding to a relatively high level
(about 60% of the maximal effect of glycine). All three

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 16 5101
s-cis conformation corresponds to the active one which is,
according toourcalculations, also accessible tothe benzothio-
phene 56.
bond formation, i.e., by an amido group, might allow sub-
stitution and thereby exploration of the space adjacent to this
part of the molecule.
It quickly became clear that the resulting 3-acylamino-2-
aminopropionic acids indeed bear a considerable potential as
ligands at the strychnine-insensitive glycine coagonist site
associated with the NMDA receptor complex. Compounds
with bicyclic aromatic ring systems (13, R-14, 15, 17-21),
were more potent than monocyclic analogues (8-11), pro-
vided the conjugation between the ring system and the
carbonyl group of the side chain was maintained (see the
inactive compounds 12 and 16, Table 1). Attachments of
substituents to the bicyclic ring systems reveal that the 2-
indolyl scaffold was the most promising of those tried for
optimization of binding affinity. For example, the attachment
of a bromo- or 2-ethylphenyl substituent in position 4 of the
indole ring resulted in massive increases in glycine site binding
affinity, while similar substitutions at the 3-indolyl, 2-ben-
zothiophene, or the 2-naphthyl scaffolds lead to less or no
improvement of affinities (Table 2).
Discussion
Partial agonism at the glycine coagonist site of the NMDA
receptor complex might be an attractive therapeutic principle
because it would make it possible to stabilize the receptor
efficacy at a certain desired level without completely closing
down normal synaptic signal transmission. To date, only a
small number of partial glycine site agonists are known, such
as D-cycloserine,34,39 HA-966,40,41 or some cyclic homologues
of glycine.23 Retaining agonistic activity in glycine-derived
molecules has proven to be difficult because of severe space
restrictions for substituents at or in the immediate vicinity of
theaminoacidgroup. However, forthe reasons outlined inthe
introduction, we felt that mimicking the hydroxy group of
D-serine in a way that preserves its potential for hydrogen
Table 5. Potencies and Efficacies of 3-Acylamino-2-aminopropionic
Acid Derivatives in a Functional Assay (Stimulation of [3H]MK-801
Binding under Nonequilibrium Conditions) for the Glycine Recognition
Site of the NMDA Receptor
Structure-activity relationships in the indole-2-carbonyl-
diaminopropionic acid series showed that introducing sub-
stituents in the indole position 3 (R1) and, even more
importantly, position 4 (R2) was the most successful measure
to further increase the affinities of these types of compounds.
The increase of affinities (compared to the unsubstituted
R-14) obtained by different lipophilic substituents in position
4 of these compounds is a sign that the groups attached reach
into a specifically shaped lipophilic area of the glycine site.
Thus, groups such as methoxy (38), chloro (37), bromo (rac-
22), or phenyl (39) give rise to significant increases of affinity.
The highest increase was obtained with 2-ethylphenyl (rac-26)
in position 4. A comparison between phenyl and 2-ethylphe-
nyl indicates a special role for the ethyl group: as calculated by
quantum chemical geometry optimization of achiral model
systems (not shown), the torsion angle between the indolyl
system and the 4-phenyl ring (as in 39) as well as between the
indolyl system and the 4-(2-ethylphenyl) ring (as in rac-26)
corresponds for both at their minimum energy to a gauche
conformation;inother words, the ethyl group ofrac-26 has no
critical influence on the preferred orientation of the phenyl
ring. From this we conclude that the 2-ethyl substituent
improves the binding affinity likely by interacting with an
additionalpart of the glycinesite andnot bymerestabilization
of the active conformation.
compd
pEC50 or pA2
efficacy (%)a
N
10
41
5.1 ( 0.10
7.6 ( 0.16
5.0 ( 0.15
5.6 ( 0.10
6.9 ( 0.14
6.5 ( 0.05
7.7 ( 0.09
6.5 ( 0.10
6.6 ( 0.09
5.3 ( 0.10
6.0 ( 0.03b
6.2 ( 0.18b
5.7 ( 0.12b
5.5 ( 0.23b
5.1 ( 0.20b
81 ( 5
64 ( 1
63 ( 7
63 ( 3
62 ( 4
59 ( 2
55 ( 2
55 ( 10
45 ( 4
42 ( 1
38 ( 2.5
32 ( 2
28 ( 1
22 ( 2
none (antagonist)
8
5
4
5
4
3
4
2
4
3
3
3
3
3
3
9
11
40
44
R-26
43
R-22
R-14
47
51
33
29
31
a The compounds listed were tested in a functional NMDA receptor
assay (potentiation of [3H]MK-801 binding under nonequilibrium con-
ditions) as described in the Experimental Section. The efficacies are
expressed relative to the maximal effect of glycine (=100%) and are
based on a zero-level defined in the presence of 30 μM 5,7-dichloroky-
nurenic acid (Figure 3). b For the compounds with a low intrinsic efficacy
(similar to the basal level determined by the presence of endogenous
glycine or less), potencies are expressed as pA2 values and were deter-
mined from their apparent antagonism against glycine (Schild plots38).
The data shown are mean values ( SEM from N independent determi-
nations.
While methylation of the indolyl N-atom has no effect on
the affinity of the unsubstituted compound R-14 (compare
with 46), it dramatically reduces the one of the 2-ethylphenyl
Table 6. Energy Differences from ab Initio Geometry Optimization of Model Systems 54-56a
“s-cis”
energy (hartree)
“s-trans”
energy (hartree)
X
t (deg)
t (deg)
E(s-cis) - E(s-trans) (kcal/mol)
NH (54)
N-Me (55)
S (56)
0.0
20.4
17.9
-571.445 418
-610.721 292
-914.237 384
159.1
143.4
179.7
-571.437 954
-610.713 387
-914.238 429
-4.7
-5.0
0.7
a Structures 54-56 are taken as model systems of respective compounds 26, 27, and 53. Constrained geometry optimizations were carried out with the
Turbomole 5.5 program on the BP86 level of theory, taking advantage of the RI-DFT method using the SV(P) basis set. The torsion angle t was
constrained from 180° to 0° in steps of 20°. t: torsion angle X-C-CdO.

5102 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 16
Urwyler et al.
Figure 5. Hypothetical active s-cis conformation of R-26 displayed
with solvent-accessible surface colored by lipophilicity. Highly
polar 3-acylamino-2-aminopropionic acid substructure is colored
from blue to green, and highly lipophilic biphenyl substructure is
colored in brown. Model and lipophilic surface representation was
generated with MOLCAD within Sybyl8.0 (Tripos International,
1699 South Hanley Road, St. Louis, MO 63144) using default
parameters.
Figure 4. Display of the least-squares superimposition of S-cis
models of 55 and 56 on 54 with respect to their common substruc-
tures of non-hydrogen atoms. The z-clipped van der Waals surfaces
of the NH (H in cyan), S (in yellow), and NCH3 (C in white, H in
cyan) groups illustrate the significant change in volume require-
2
˚
ments of 54, 56, 55. The unclipped surface areas vary from 14 A (54)
2
2
and 26 A (56) to 37 A (55).
˚
˚
derivative rac-26 (compare with 53). It seems likely that the
decreaseof affinityproducedby indolyl-N-methylation comes
from a steric conflict evoked by the 4-phenyl substituent,
which would be moved into a position unfavorable for
receptor binding. The fact that the affinity of the 2-ethylphe-
nylbenzothiophene analogue 27 (pKi = 6.2) is between those
of the NH- and NCH3-indolyl derivatives (rac-26, pKi = 7.5;
53, pKi = 3.9) supports this contention, since the space
occupied by a sulfur atom is intermediate between that of an
NH and an N methyl groups, respectively (Figure 4). The
same rank order of affinities is found in the 4-bromo-sub-
stituted series 52 (4-bromoindolyl-N-methyl, pKi = 5.0), 24
(4-bromobenzothiophene, pKi = 6.4), and rac-22 (4-bro-
moindolyl-NH, pKi = 6.8), again illustrating the increase of
affinity with decreasing steric bulk at the heteroatom of the
bicyclic ring system. The differences are less pronounced,
however, which is in line with the smaller size of the bromo
atom compared with the 2-ethylphenyl group. With an even
smaller chloro atom in position 4, no steric hindrance is
produced anymore by replacing the indolyl by a benzothio-
phene ring system; in fact, compound 30 (pKi=6.5) is some-
what more potent than 37 (pKi = 6.0), suggesting that the
slightly increased bulk of the S-atom moves the 4-chloro atom
into a more favorable position for binding. A similar steric
influence is exerted by a chloro-atom in position 7 of the
indole nucleus. While it has no effect on glycine site affinity on
its own (compare 36 with R-14), it results in increased affinity
of the 4,7-dichloro substituted compound 51 in comparison
with 37, again suggesting that the slightly bulky 7-chloro
group moves the 4-chloro substituent into a favorable posi-
tion for binding. In conjunction with the much more space-
filling 2-ethylphenyl substituent, however, the 7-chloro atom
has a detrimental effect on glycine site affinity (50 vs rac-26),
although much less than indolyl-N-methylation.
approximation of the contour of the ligand-induced receptor
conformation.
In order to assess our pharmacophore hypothesis, we
attempted to model the binding of R-26 to the ligand binding
domain of NR1 using the conformation observed in the X-ray
structure of the NR1/cycloleucine complex.42 We chose the
cycloleucine complex for modeling because the binding cav-
ities revealed in the complexes with glycine, D-serine, DCS,
ACC, and ACBC were too small for accommodation of R-26.
Several different binding modes were modeled whereby each
mode involved the salt bridges with Arg523 and Asp732 also
reported for the complex of cycloleucine and related partial
agonists. Different regions ofthe NR1 glycine binding domain
were found to be accessible to the o-ethylphenyl moiety of
R-26, but in contrast to the observed hydrogen bonding of
D-serine, any hydrogen bond of its amide proton was absent.
However, giventhe significant ligand-inducedconformational
changes of the NR1 constructs, the validity of our models is
too uncertain to support any pharmacophore hypothesis.
3-Acylamino-2-aminopropionic acid analogues also bind
to the glutamate recognition site at the NMDA receptor,
labeled by [3H]CGP39653 (Tables 1-4). However, the
increases in glycine site affinities obtained by the derivatiza-
tions discussed above do not lead to similar increases in
affinities for the glutamate site; as a result, more than a 100-
fold selectivity for the glycine over the glutamate site were
found for a number of compounds, for example, R-22 and
many others; see Tables 1-4.
The structure-activity relationships that govern the affi-
nities of these types of compounds do not predict their
intrinsic efficacies at the glycine coagonist site of the NMDA
receptor. In the series of compounds presented here, the one
with the highest efficacy to stimulate [3H]MK-801 binding
under nonequilibrium conditions wasthe furane derivative10,
which had only a moderate binding affinity for the glycine site
(cf. Tables 1 and 5). At the other end of the scale, compound
31, which had a similar affinity, was found to be a “silent”
glycine antagonist without any intrinsic efficacy at all. At
present, it seems that no pattern of structure-activity rela-
tionships can be derived from our data to explain the rules
governing the intrinsic efficacies of 3-acylamino-2-aminopro-
pionic acids at the glycine site.
One hypothetical active conformation of R-26, as derived in
part by conformational analysis, is shown in Figure 5. Since
the phenyl ring on its own did not increase glycine site affinity
(compare R-14 and 39), it seems conceivable that the essential
part of the phenylethyl substituent is the ethyl group, reaching
into a separate attachment site, and that the phenyl ring merely
acts as a “scaffold” to support this ethyl group. The solvent-
accessible surface shown in Figure 5 may be considered as an

Article
In conclusion, the compounds of the 3-acylamino-2-ami-
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 16 5103
Data Analysis. The binding in the presence of 10 μM gluta-
mate, but in the absence of any added glycine agonist, was
named “basal binding”. However, it was found that a glycine
antagonist like 5,7-dichlorokynurenic acid inhibited this basal
binding to a considerable extent, suggesting that the glycine site
was already activated by about 25% by endogenous glycine
under basal conditions (Figure 3). A “zero level” was therefore
determined in the presence of 30 μM 5,7-dichlorokynurenic
acid, and the intrinsic efficacies of partial agonists were calcu-
lated with reference to this zero level. Figure 3 illustrates this
principle.
Molecular Modeling. Modeling was performed using our in-
house software WitNotP [Widmer, A. WITNOTP: A Computer
Program for Molecular Modelling; Novartis Pharma AG: Basel,
Switzerland, 1997] using the Tripos force field (TAFF) of Sybyl
6.4 [Tripos, Inc., St. Louis, MO] and Turbomole5.5 on a
HPxw8200 workstation under Linux.47 Geometries of mole-
cules were obtained by building molecular models from frag-
ments, followed by TAFF energy minimization. The vdW
surfaces were calculated with WitNotP.
nopropionic acid type presented here are a novel class of
ligands acting at the strychnine-insensitive glycine coagonist
site associated with the NMDA receptor. They provide the
proof that it is possible to retain high affinity and different
degrees of intrinsic efficacy in molecules derived from
D-serine. Extension of an amido analogue of D-serine by
monocyclic or bicyclic (preferably indolyl) aromatic ring
systems maintains moderate to high affinity for the glycine
site. Occupation of an additional lipophilic pocket by suitable
substituents in position 4 of the indolyl ring system confers
nanomolar affinity to this type of compound but leaves little
steric tolerance for further substitution. The resulting com-
pounds havevariousdegreesofintrinsic activity, rangingfrom
pure antagonism to high efficacy (80%) partial agonism.
Experimental Section
Binding Assays for Various Recognition Sites on the NMDA
Receptor Complex. The rat brain cortex membrane preparation
used was the same for all NMDA receptor related binding assays
and was based on procedures described earlier.43 The binding of
[3H]glycine to the strychnine-insensitive coagonist site was
measured as described previously.43 [3H]CGP3965344 was used
as a ligand to label the NMDA recognition site. The binding
assay was performed as described previously,45 with the follow-
ing modification. Because complex interactions between the
glutamate and glycine recognition sites are known to occur in
radioligand binding studies,30,31 30 μM 5,7-dichlorokynurenic
acid was added to the assay to occlude an allosteric inhibition of
[3H]CGP39653 binding produced by agonism of the test com-
pounds at the glycine site. To determine the affinities of com-
pounds for the MK-801 site within the NMDA receptor-
operated ion channel, [3H]MK-801 binding was measured under
equilibrium conditions, also as described earlier.45 The affinities
of compounds for the polyamine site associated with the
NMDA receptor complex were assessed in binding experiments
using [3H]spermidine as a radioligand.46 The assay mixture
(final volume of 1 mL) in 50 mM Tris-HCl buffer (pH 7.0)
consisted of an aliquot of membranes prepared as described
above (corresponding to approximately 10 mg of original tissue
weight), 20 nM [3H]spermidine, and the test compounds at the
appropriate concentrations. Nonspecific binding was measured
in the presence of 10 mM unlabeled spermidine. The samples
were incubated for 30 min at 0 °C and were then filtered through
Whatman GF/B filters which had been soaked in a 0.3%
polyethyleneimine solution for at least 1 h prior to use. The
filters were then washed with ice-cold incubation buffer and
further processed as described above.
In all receptor binding assays, pIC50 values were derived from
the curves describing the inhibition of specific [3H]ligand bind-
ing by the test compounds and converted into pKi values using
the Cheng-Prusoff relationship. Prism 3.0 software (GraphPad
Software, San Diego, CA) was used for these calculations.
Functional NMDA Receptor Test. MK-801 Potentiation As-
say. The final pellet of the rat brain membrane preparation43 was
resuspended in ice-cold Tris-HCl buffer (10 mM, pH 7.4) to give a
final concentration of 10-15 mg of original tissue weight (about
50-200 mgofprotein) persample. Inthe binding experiment, this
membrane suspension was preincubated for 15 min at 25 °C in the
presence of a saturating concentration (10 μM) of glutamate and
varying concentrations of compounds acting at the glycine site as
indicated in Results. [3H]-MK-801 (15-30 Ci/mMol, DuPont
NEN) was then added to give a concentration of 3 nM in a final
assay volume of 2 mL. After further incubation at 25 °C for
15 min (nonequilibrium conditions), bound and free radioligand
were separated by filtration through Whatman GF/B filters,
using a Brandel M24R cell harvester. The radioactivity on the
filters was then determined by liquid scintillation counting.
Chemistry. Chromatographic separations in preparative scale
were performed either by flash chromatography48 or by medium
pressure liquid chromatography (MPLC) using a UV detector
and a pump. In all cases 40-60 μm silica gel (Merck) was used. If
not otherwise specified, 1H NMR spectra for intermediates were
taken on a Gemini 200 MHz spectrometer. The target com-
pounds were uniformly obtained as white powders and char-
acterized by 1H NMR spectra (Bruker 360 MHz spectrometer).
1
Chemical shifts for H are given in ppm (δ) relative to tetra-
methylsilane (TMS) as internal standard. Combustion analyses
were performed by the Analytical Department of Novartis AG,
and values are within 0.4% of theoretical values unless other-
wise specified.
Method A: General Procedure for the Acylation of rac-5b with
a Carboxylic Acid Chloride. To a 10% solution of rac-5b in a 1:1
mixture of dry pyridine and dichloromethane at -10 °C a
solution of 1.3 equiv of the acid chloride in dichloromethane
was slowly added. After being stirred for 45 min at -10 °C, the
mixture was poured onto ice-water and extracted with ethyl
acetate. The combined organic phases were washed with aqu-
eous citric acid, dried over Na2SO4, and evaporated. The residue
was purified by MPLC with tert-butyl methyl ether as eluent.
Method B: General Procedure for the Condensation of Car-
boxylic Acids with rac-5b. To a suspension of the carboxylic acid
in pyridine (1-2%) an amount of 1.1 equiv of carbonyldiimi-
dazole was added. The mixture turned to a yellow solution and
was stirred overnight. After addition of 1 equiv of rac-5b stirring
was continued for 2 days, the mixture evaporated to dryness,
and the residue crystallized from hot ethanol.
Method C: General Procedure for the Saponification of rac-3-
Acylamino-2-benzyloxycarbonylaminopropionic Esters. To a 3%
solution of the substrate in THF 1 equiv of 1 M aqueous NaOH
solution was added. The mixture was stirred at room tempera-
ture until the starting material had disappeared (TLC). After
dilution with H2O, the THF was evaporated, the aqueous phase
was acidified with 1 M aqueous HCl solution and extracted with
ethyl acetate. The combined organic phases were washed with
brine, dried over Na2SO4, and concentrated to dryness.
Method D: General Procedure for Removal of the Z-Group by
Hydrogenation. A 5% solution of the Z-protected amino acid in
1:1 THF/H2O was treated with about 20% (w/w) of Pd on
carbon (5%) and hydrogenated until no hydrogen was taken up.
After filtration the catalyst was washed with 1:1 THF/H2O and
the filtrate evaporated to dryness. The residue was crystallized
from the specified solvent.
Method E: General Procedure for Removal of the Z-Group by
Iodotrimethylsilane. A 5% solution of the substrate in dichlor-
omethane was treated with 2.2 equiv of iodotrimethylsilane at
0 °C, and the mixture was stirred for 1 h. The reaction mixture
was concentrated and the residue distributed between water and

5104 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 16
Urwyler et al.
ethyl acetate. The pH of the water phase was adjusted to 6 with
0.1 M NaOH solution, and the precipitate was collected and
recrystallized from water/isopropanol.
with the enantiomer R-14 was observed. [R]20 -1.8° (c 1.00,
D
1 M NaOH). 1H NMR (D2O þ DCl): 7.75 d, J = 7 Hz, 1H; 7.58
d, J = 8 Hz, 1H; 7.38 t ꢀ d, J = 7 and 1 Hz, 1H; 7.22 t ꢀ d, J =
7 and 1 Hz, 1H; 7.16 s, 1H; 4.40-4.35 m, 1H; 4.10-3.90 m, 2H.
Method F: General Procedure for the Pd-Catalyzed Coupling
Reaction between Arylboronic Acids and Aryl Halogenides. An
amount of 1 equiv of the aryl halogenide, 1.1-1.7 equiv of the
arylboronic acid, and 3 mol % of tetrakistriphenylphosphine-
palladium were refluxed in a 10:3:2 mixture of toluene, ethanol,
and 2 M Na2CO3 solution until the aryl halogenide disappeared
(TLC). After filtration the phases were separated and the
aqueous phase was extracted with diethyl ether and acidified
with 6 M HCl solution. Extraction of the acidic water phase with
ethyl acetate, drying of the organic phase over Na2SO4, and
evaporation yielded the coupling product, in most cases suffi-
ciently pure for the next step.
rac-3-Amino-2-benzyloxycarbonylaminopropionic Acid rac-
5a. rac-5a was prepared (69%) using the literature procedure
for R-5a26 but starting from rac-Z-asparagin. Mp 235-239 °C
(dec). 1H NMR (D2O þ DCl): 7.40 br s, 5H; 5.15 s, 2H; 4.60-
4.45 m, 1H; 3.60-3.45 m, 1H; 3.40-3.20 m, 1H.
Anal. (C12H13N3O3 0.3H2O) C, H, N, O.
3
rac-2-Amino-3-[(benzo[b]thiophene-2-carbonyl)amino]propio-
nic Acid 17. Condensation of benzo[b]thiophene-2-carboxylic
acid with rac-5b (method B, 88%) and removal of the protecting
groups (method C, 86%; method E, 74%) gave 17, mp 230 °C
(dec). 1H NMR (D2O þ DCl): 7.7-7.6 m, 2H; 7.53 s, 1H; 7.25-
7.10 m, 2H; 4.15-4.05 m, 1H; 3.75-3.50 m, 2H. Anal.
(C12H12N2O3S) C, H, N, O, S.
rac-2-Amino-3-[(naphthalene-2-carbonyl)amino]propionic Acid
20. Condensation of naphthalene-2-carboxylic acid with rac-5b
(method B, 87%) and removal of the protecting groups (method
1
C, 95%; method E, 63%) gave 20, mp 220 °C (dec). H NMR
([(CD3)2SO]): 8.9 br s, 1H; 8.45 s, 1H; 8.1-7.9 m, 4H; 7.65-
7.55 m, 2H; 3.8-3.6 m, 2H; 3.5-3.4 m, 1H. Anal. (C14H14N2O3)
C, H, N, O.
rac-2-Amino-3-[(4-bromo-1H-indole-2-carbonyl)amino]propio-
nic Acid rac-22. Condensation of 2.00 g of 4-bromo-1H-indole-2-
carboxylic acid with rac-5b according to method B gave 3.15 g
(79%) of 2-benzyloxycarbonylamino-3-[(4-bromo-1H-indole-2-
carbonyl)amino]propionic acid methyl ester 22a as a slightly
rac-3-Amino-2-benzyloxycarbonylaminopropionic Acid Meth-
yl Ester Hydrochloride rac-5b. rac-5b was prepared (71%) using
the literature procedure for R-5b26 but starting from rac-5a. Mp
136-138 °C. 1H NMR (CD3OD): 7.40-7.25 m, 5H; 5.10 s, 2H;
4.55-4.45 m, 1H; 3.75 s, 3H; 3.50-3.40 m, 1H; 3.30-3.15 m, 1H.
(R)-2-Amino-3-[(1H-indole-2-carbonyl)amino]propionic Acid
R-14. Acylation of 4.33 g of R-5b with indole-2-carbonyl
chloride according to method A gave 5.71 g (96%) of (R)-2-
benzyloxycarbonylamino-3-[(1H-indole-2-carbonyl)amino]-
propionic acid methyl ester R-14a, mp 178 °C. 1H NMR
(CDCl3): 9.50 s, 1H; 7.60 d, J = 9 Hz, 1H; 7.40 d, J =
9 Hz, 1H; 7.35-7.20 m, 6H; 7.15-7.05 m, 1H; 7.05-6.95 m,
1H; 6.85 s, 1H; 6.5 d, J = 7 Hz, 1H; 5.10 s, 2H; 4.65-4.50 m, 1H;
3.95-3.80 m, 2H; 3.75 s, 3H. A solution of 3.95 g (10 mmol) of
R-14a in 50 mL of THF was added to a slurry of 11.3 mL of
Dowex OH- (capacity, 0.974 mmol OH-/mL, prepared by
elution of Dowex Cl- with 2 M NaOH solution until the eluent
gave no precipitation with diluted AgNO3 solution) in 30 mL of
H2O. The mixture was stirred for 20 h at room temperature,
during which time small portions of Dowex OH- were added
until no starting material was detected. The Dowex OH- was
filtered and, after washing with 300 mL of 1:1 THF/H2O, filled
into a glass column and eluted with 2 M HCl solution/THF, 1:1.
The product fractions were combined and concentrated and the
residue was dissolved in hot acetone, resulting in a deeply dark
solution that was decolorized by addition of carbon. After
filtration, the product was precipitated by addition of H2O,
yielding 2.20 g (57.6%) of (R)-2-benzyloxycarbonylamino-
3-[(1H-indole-2-carbonyl)amino]propionic acid R-14b as a col-
1
yellow solid, mp 187-193 °C. H NMR ([(CD3)2SO]): 12.05 s,
1H; 8.75 t, J = 6 Hz, 1H; 7.75 d, J = 6 Hz, 1H; 7.45, d, J = 6 Hz,
1H; 7.35 br s, 5H; 7.25 d, J = 6 Hz, 1H; 7.18-7.10 m, 2H; 5.05 s,
2H; 4.35 q, J = 6 Hz, 1H; 3.70-3.55 m, 5H. Deprotection of 22a
(method C, 97%; method E, 60%) gave rac-22, mp 252 °C (dec).
1H NMR (D2O þ DCl): 7.38 d, J = 8.4 Hz, 1H; 7.25 d, J = 6 Hz,
1H; 7.12-7.08 m, 1H; 6.98 s, 1H; 4.32-4.30 m, 1H; 4.0-3.8 m,
2H. Anal. (C12H12BrN3O3) C, H, Br, N, O.
(R)-2-Amino-3-[(4-bromo-1H-indole-2-carbonyl)amino]propio-
nicAcid R-22. Condensation of4-bromo-1H-indole-2-carboxylic
acid with R-5b (method B, 66%) and removal of the protecting
groups (method C, 96.4%; method E, 73%) gave R-22, mp
248 °C (dec). Thin layer chromatography on Chiralplate
(Machery-Nagel)49 with CH3CN/CH3OH/H2O, 4:1:2, showed
a single spot, Rf = 0.39 (ninhydrine). As a control, the racemate
rac-22 showed two spots with Rf = 0.39 and 0.48. [R]20D -0.5°
(c 1.00, 1 M NaOH). 1H NMR ([(CD3)2SO] þ DCl): 7.52 d, J =
7 Hz, 1H; 7.32 d, J = 7 Hz, 1H; 7.22 s, 1H; 7.18 t, J = 7 Hz, 1H;
4.10-4.05 m, 1H; 3.85-3.80 m, 2H. Anal. (C12H12BrN3O3) C,
H, Br, N, O.
rac-2-Amino-3-[(4-bromobenzo[b]thiophene-2-carbonyl)amino]-
propionic Acid 24. Condensation of 4-bromobenzo[b]thiophene-
2-carboxylic acid with rac-5b (method B, 42.9%) and removal of
the protecting groups (method C, 100%; method E, 69.5%) gave
1
24, mp 198 °C (dec). H NMR (D2O þ NaOD): 7.75 d, J =
orless powder, mp 196-206 °C. [R]20 6.90 (c 0.990, acetone).
7.2 Hz, 1H; 7.00 s, 1H; 7.55 d, J = 7.2 Hz, 1H; 7.3-7.2 m, 1H;
D
1H NMR ([(CD3)2SO]): 11.60 s, 1H; 8.55 br t, 1H; 7.70-7.55 m,
2H; 7.40 d, J = 9 Hz, 1H; 7.35-7.25 m, 5H; 7.15 t, J = 9 Hz, 1H;
7.10-6.95 m, 3H; 5.00 s, 2H; 4.35-4.20 m, 1H; 3.75-3.50 m,
2H. Anal. (C20H19N3O5) C, H, N, O. Deprotection of R-14b
according to method D yielded (95%) R-14 as a white powder,
mp 242 °C (dec; H2O/THF, 1:1). Thin layer chromatography on
Chiralplate (Machery-Nagel)49 with CH3CN/CH3OH/H2
O, 4/1/1, showed a single spot, Rf = 0.46 (ninhydrine). No con-
tamination with the enantiomer S-14 was observed. [R]20D 2.5°
(c 1.00, 1 M NaOH). 1H NMR (D2O þ DCl): 7.75 d, J = 7 Hz,
1H; 7.58 d, J = 8 Hz, 1H; 7.38 t ꢀ d, J = 7 and 1 Hz, 1H; 7.20 t,
J = 7 Hz, 1H; 7.16 s, 1H; 4.40-4.35 m, 1H; 4.06-3.90 m, 2H.
3.8-3.6 m, 1H; 3.6-3.4 m, 2H. Anal. (C12H11BrN2O3S 0.5H2O)
3
C, H, Br, N, O, S.
rac-2-Amino-3-[(5-bromonaphthalene-2-carbonyl)amino]propio-
nic Acid 25. Condensation of 5-bromonaphthalene-2-carboxylic
acid with rac-5b (method B, 88%) and removal of the protecting
groups (method C, 99%; method E, 63%) gave 25, mp 207 °C
(dec). 1H NMR ([(CD3)2SO]): 9.1 br s, 1H; 8.55 s, 1H; 8.17 d, J =
9Hz, 1H;8.12dꢀ d, J= 9 and 2 Hz, 1H; 8.08 d, J=7.2Hz;7.95d,
J = 7.2 Hz, 1H; 7.50 d ꢀ d, J = 7.8 and 7.8 Hz, 1H; 3.9-3.7 m, 1H;
3.7-3.5 m, 2H. Anal. (C14H13BrN2O3 0.4H2O) C, H, Br, N, O.
3
rac-2-Amino-3-{[4-(2-ethylphenyl)-1H-indole-2-carbonyl]amino}-
propionic Acid rac-26. An amount of 7.65 g (31.9 mmol) of
4-bromo-1H-indole-2-carboxylic acid and 8.20 g (54.7 mmol) of
2-ethylphenylboronic acid were coupled according to method F,
yielding 6.11 g (72%) of 4-(2-ethylphenyl)-1H-indole-2-carboxylic
acid 26a as a beige powder, mp 216 °C (dec). 1H NMR
([(CD3)2SO]): 12.9 br s, 1H; 11.9 s, 1H; 7.50-7.20 m, 5H; 6.95 d,
J = 7 Hz, 1H; 6.60 d, J = 0.2 Hz, 1H; 2.50 q, J = 7 Hz, 2H; 0.90 t,
J = 7 Hz, 3H. Condensation of 4-(2-ethyl-phenyl)-1H-indole-2-
carboxylic acid 26a with rac-5b (method B, 45%) and removal of
Anal. (C12H13N3O3 0.3H2O) C, H, N, O.
3
(S)-2-Amino-3-[(1H-indole-2-carbonyl)amino]propionic Acid
S-14. Acylation of S-5b27 with indole-2-carbonyl chloride
(method A, 87%) and removal of the protecting groups
(method C, 86%; method D, 79%) gave S-14, mp 225-250 °C
(dec; H2O/THF, 1:1). Thin layer chromatography on Chiral-
plate (Machery-Nagel)49 with CH3CN/CH3OH/H2O, 4:1:1,
showed a single spot, Rf = 0.52 (ninhydrine). No contamination

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 16 5105
the protecting groups (method C, 100%; method D, 83.5%) gave
rac-26, mp 258 °C (dec). 1H NMR ([(CD3)2SO]): 7.45 d, J = 8 Hz,
1H; 7.40-7.35 m, 2H; 7.30-7.25 m, 2H; 7.20 d, J = 8 Hz, 1H; 6.90
d, J = 7 Hz, 1H; 6.80 s, 1H; 4.1-4.0 m, 1H; 3.8-3.6 m, 2H; 2.5-2.4
m, 2H; 0.95 t, J = 6 Hz, 3H. Anal. (C20H21N3O3) C, H, N, O.
(R)-2-Amino-3-{[4-(2-ethylphenyl)-1H-indole-2-carbonyl]amino}-
propionic Acid R-26. Procedure I. Condensation of 4-(2-
ethylphenyl)-1H-indole-2-carboxylic acid 26a with R-5b (method
B, 96%) and removal of the protecting groups (method C, 82%;
method E, 15.4%) gave R-26, mp 240 °C (dec). Thin layer
chromatography on Chiralplate (Machery-Nagel)49 with CH3CN/
MeOH/H2O, 4:1:1) showed one single spot with Rf = 0.31
41 mL of concentrated H2SO4, 69 mL of H2O, and 103 mL of
ethanol. After the mixture was stirred for 1 h, H2O was added and
the mixture extracted with tert-butyl methyl ether. The organic
phase was washed with brine, dried over Na2SO4, and evaporated
to yield 5.54 g (63%) of 2-(phenylhydrazono)hexanoic acid ethyl
1
ester 33a as a red oil. H NMR (CDCl3): 12.0 br s, 1H (NH of
E-isomer); 7.90 br s, 1H (NH of Z-isomer); 7.30-7.20 m, 2H; 7.20-
7.10 m, 2H; 7.00-6.85 m, 1H; 4.30 q, J = 8 Hz, 2H; 2.50 pent, J =
8 Hz, 2H; 1.60-1.30 m, 7H; 0.90 t, J = 8 Hz, 3H. A mixture of
5.54 g of 33a, 5.6 mL of concentrated H2SO4, and 56 mL of dry
ethanol was heated at 100 °C for 90 min. After cooling and dilution
with water, the reaction mixture was extracted with toluene and the
organic phase washed with 1 M NaHCO3 solution and brine and
evaporated to dryness, leaving 5.52 g of a yellow oil. From another
reaction with 6.95 g of the starting hydrazone, 6.94 g of this oil was
obtained. Both oils were combined and purified by MPLC on silica
gel with toluene/cyclohexane, 4:1, yielding 6.50 g (56%) of 3-propyl-
1H-indole-2-carboxylic acid ethyl ester 33b as a slightly yellow solid.
1H NMR (CDCl3): 8.65 br s, 1H; 7.70 d, J = 8 Hz, 1H; 7.40-
7.25 m, 2H; 7.13 t ꢀ d, J = 8 and 2 Hz, 1H; 4.40 q, J = 7 Hz, 2H;
3.05 t, J = 7 Hz, 2H; 1.70 hex, J = 7 Hz, 2H; 1.40 t, J = 7 Hz, 3H;
0.95 t, J = 7 Hz, 3H. A mixture of 6.5 g (28.12 mmol) of 33b,
32.4 mL of 1 M NaOH solution, and 140 mL of ethanol was heated
at 70 °C for 2 h. After evaporation of the ethanol the residual
aqueous solution was diluted with H2O, washed with diethyl ether,
acidified with 2 M HCl solution, and extracted with diethyl ether.
Drying of the organic phase over Na2SO4 and evaporation yielded
5.27 g (92%) of 3-propyl-1H-indole-2-carboxylic acid 33c as slightly
(ninhydrine). A racemic sample gave two spots with Rf = 0.31
1
and 0.43, respectively. [R]20 -6.5° (c 1.00, DMSO). H NMR
D
([(CD3)2SO]): 11.90 s, 1H; 8.62 t, J = 6 Hz, 1H; 7.9-7.4 br, ∼3H;
7.45 d, J = 7 Hz, 1H; 7.40-7.32 m, 2H; 7.30-7.18 m, 3H; 6.90 d,
J= 7 Hz, 1H; 6.72 s, 1H; 3.78-3.70 m, 1H; 3.52-3.38 m, 2H; 2.55-
2.40 m, 2H; 0.95 t, J = 7 Hz, 3H. Anal. (C20H21N3O3) C, H, N, O.
Procedure II. : An analytically pure sample with the same
optical rotation and 1H NMR data was obtained in the following
way. Condensation of 5.64 g (21.26 mmol) of 4-(2-ethylphenyl)-
1H-indole-2-carboxylic acid 26a with 7.76 g (21.26 mmol) of
R-5c according to method B gave 9.99 g (81.6%) of (R)-2-ben-
zyloxycarbonylamino-3-{[4-(2-ethyl-phenyl)-1H-indole-2-carbonyl]-
amino}propionic acid benzyl ester 26b as a yellow foam. An amount
of 5.00 g of this foam was dissolved in 87 mL of glacial acetic acid, an
amount of 500 mg of 5% Pd on carbon was added, and the mixture
was hydrogenated at normal pressure for 3 h. After filtration from the
catalyst and evaporation of the solvent, the residue was triturated
with water, filtered, and recrystallized from isopropanol/water,
yielding 2.73 g (89.4%) of R-26, identical in all properties with the
sample from procedure (I).
1
yellow crystals, mp 114-119 °C. H NMR (CDCl3): 8.77 s, 1H;
7.70 d, J = 8 Hz, 1H; 7.45-7.30 m, 2H; 7.20-7.10 m, 1H; 6.5-
5.5 br, 1H; 3.10 t, J = 8 Hz, 2H; 1.75, hex, J = 8 Hz, 2H; 1.00 t, J =
8, 3H. Acylation of rac-5b with 3-propyl-1H-indole-2-carbonyl
chloride (prepared in 90% yield by refluxing 33c with 10 equiv
of thionyl chloride in dichloromethane) (method A, 96%) and
removal of the protecting groups (method C, 94%; method E, 82%)
gave 33, mp 225 °C (dec). 1H NMR ([(CD3)2SO]): 11.9 br s, 1H;
8.6 br t, 1H; 8.0-7.8 br, ∼3H; 7.58 d, J = 7.2 Hz, 1H; 7.45 d, J =
7.2 Hz, 1H; 7.15 t, J=7.2 Hz, 1H; 7.00 t, J = 7.2 Hz, 1H; 3.80-
3.65m, 2H; 3.6-3.5 m, 1H;3.1-3.0 t, J =9 Hz, 2H;1.60 hex, J =
rac-2-Amino-3-{[4-(2-ethylphenyl)benzo[b]thiophene-2-carbo-
nyl]amino}propionic Acid 27. An amount of 5.00 g (19.5 mmol)
of 4-bromobenzo[b]thiophene-2-carboxylic acid and 5.00 g
(33.1 mmol) of 2-ethylphenylboronic acid was coupled accord-
ing to method F, yielding 4.87 g (88%) of 4-(2-ethylphenyl)-
benzo[b]thiophene-2-carboxylic acid 27a as beige powder, mp
207-210 °C (H2O/isopropanol). 1H NMR ([(CD3)2SO]): 13.5 s,
1H; 8.07 d, J=9 Hz, 1H; 7,62-7.58 m, 1H; 7.45-7.42 m, 2H;
7.40 s, 1H; 7.35-7.30 m, 2H; 7.20 d, J=6 Hz, 1H; 2.50-2.25 m,
2H; 1.40 t, J = 6 Hz, 3H. Condensation of 27a with rac-5b
(method B, 97%) and removal of the protecting groups (method
C, 81%; method E, 33.1%) gave 27, mp 212 °C (dec). 1H NMR
([(CD3)2SO]): 9.0-8.9 br m, 1H; 8.05 d, J=7.4 Hz, 1H; 7.60 s,
1H; 7.50 t, J = 6 Hz, 1H; 7.45-7.50 m, 2H; 7.35-7.3 m, 1H; 7.25
d, J=6 Hz, 1H; 7.20 d, J = 6 Hz, 1H; 3.7-3.6 m, 1H; 3.4-3.3 m,
2H; 2.50-2.25 m, 2H; 0.95 t, J=6 Hz, 3H. Anal. (C20H20N2-
9 Hz, 2H; 0.90 t, J = 9 Hz, 3H. Anal. (C15H19N3O3 0.1H2O) C,
3
H, N, O.
rac-2-Amino-3-[(4-phenyl-1H-indole-2-carbonyl)amino]propio-
nic Acid 39. Amounts of 4.74 g (10 mmol) of rac 2-benzyloxy-
carbonylamino-3-[(4-bromo-1H-indole-2-carbonyl)amino]pro-
pionic acid methyl ester 22a and 1.34 g (11 mmol) of phenyl-
boronic acid were coupled according to method F, yielding 3.14 g
(68.6%) of 2-benzyloxycarbonylamino-3-[(4-phenyl-1H-indole-
2-carbonyl)amino]propionic acid 39a as a brownish, highly
1
O3S 0.6H2O) C, H, N, O, S.
viscous oil. H NMR ([(CD3)2SO]): 12.5 br, 2H; 11.90 s, 1H;
3
rac-2-Amino-3-{[5-(2-ethylphenyl)naphthalene-2-carbonyl]amino}-
propionic Acid 28. Amounts of 1.60 g (6.37 mmol) of 5-bromo-
naphthalene-2-carboxylic acid and 1.05 g (7.0 mmol) of 2-ethylphe-
nylboronic acid were coupled according to method F, yielding 1.53 g
(87%) of 5-(2-ethylphenyl)naphthalene-2-carboxylic acid 28a as a
beige powder, mp 154-168 °C. 1H NMR ([(CD3)2SO]): 13.1 s, 1H;
8.70 s, 1H; 8.20 d, J = 7 Hz, 1H; 7.95 d, J=7 Hz, 1H; 7.70 t, J =
7 Hz, 1H; 7.55-7.25 m, 5H; 7.20 d, J=7 Hz, 1H; 2.40-2.10 m, 2H;
0.90 t, J = 7 Hz, 3H. Condensation of 5-(2-ethylphenyl)naphth-
alene-2-carboxylic acid 28a with rac-5b (method B, 80%) and
removal of the protecting groups (method C, 96%; method D
8.65 br t, 1H; 7.70-7.40 m, 7H; 7.38-7.20 m, 7H; 7.10 d, J =
7 Hz, 1H; 5.10-4.95 m, 2H; 4.25-4.20 m, 1H; 3.75-3.50 m, 2H.
Removal of the Z-group of 39a (method E, 44%) gave 39, mp231
1
°C (dec). H NMR ([(CD3)2SO]): 12.0 br s, 1H; 8.82 br t, 1H;
8.1-7.5 br, ∼3H; 7.65 d, J = 7 Hz, 2H; 7.52 t, J = 8 Hz, 2H;
7.47 d, J = 8 Hz, 1H; 7.39 t, J = 7 Hz, 1H; 7.28 s, 1H; 7.27 t, J =
8 Hz, 1H; 7.12 d, J = 7 Hz, 1H; 3.85-3.75 m, 1H; 3.60-3.45 m,
2H. Anal. (C18H17N3O3 0.2H2O) C, H, N, O.
3
rac-2-Amino-3-[(4-bromo-1-methyl-1H-indole-2-carbonyl)amino]-
propionic Acid 52. To a suspension of 1.41 g of NaH (55% in oil,
washed with pentane; 32.3 mmol) in 79 mL of DMF, an amount of
7.4 g (29.3 mmol) of 4-bromo-1H-indole-2-carboxylic acid methyl
ester was added, and the mixture was stirred for 20 min. This was
followed by the addition of 4.2 g (29.6 mmol) of methyl iodide in
10 mL of DMF and stirring for 3 h at room temperature. The
reaction mixture was diluted with 250 mL of H2O, and the formed
precipitate was filtered and separated by MPLC on silica gel with
toluene/ethyl acetate, 4:1, yielding 5.73 g (73%) of 4-bromo-1-
methyl-1H-indole-2-carboxylic acid methyl ester 52a as white pow-
der, mp 83-84 °C. 1H NMR ([(CD3)2SO]): 7.65 d, J = 11 Hz, 1H;
7.40 d, J = 11 Hz, 1H; 7.30 t, J = 11 Hz, 1H; 7.10 s, 1H; 4.05 s, 3H;
1
54%) gave 28, mp 178-189 °C (dec; THF/H2O, 2:1). H NMR
([(CD3)2SO]): 9.0 br t, 1H; 8.56 s, 1H; 8.05 d, J = 9 Hz, 1H; 7.75 d ꢀ
d, J = 9 and 1 Hz, 1H; 7.65-7.60 m, 1H; 7.45-7.40 m, 3H; 7.40-
7.10 m, 2H; 7.18 d, J = 7.8 Hz, 1H; 3.85-3.70 m, 1H; 3.65-3.45 m,
2H; 2.40-2.10 m, 2H; 0.90 t, J = 6 Hz, 3H. Anal. (C22H22N2O3
0.4H2O) C, H, N, O.
3
rac-2-Amino-3-[(3-propyl-1H-indole-2-carbonyl)amino]propio-
nic Acid 33. A solution of 5.64 g (35.65 mmol) of 2-oxohexanoic
acid ethyl ester and 55 mL of ethanol was added at room tempera-
ture to a mixture of 3.55 mL (36.08 mmol) of phenylhydrazine,

5106 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 16
Urwyler et al.
3.90 s, 3H. Saponification was accomplished by heating 1.93 g of 52a
in a mixture of 15 mL of 1 M NaOH solution and 30 mL of
methanol for 30 min. After acidification with 2 M HCl solution and
filtration, an amount of 1.74 g (95%) of 4-bromo-1-methyl-1H-
indole-2-carboxylic acid 52b was obtained as a white powder, mp
240-245 °C. 1H NMR ([(CD3)2SO]): 13.5-12.9 br s, 1H; 7.65 d,
J = 11 Hz, 1H; 7.40 d, J = 11 Hz, 1H; 7.25 t, J = 11 Hz, 1H; 7.10 s,
1H; 4.10 s, 3H. Condensation of 52b with rac-5b (method B, 95%)
and removal of the protecting groups (method C, 93%; method
E 78%) gave 52, mp 212 °C (dec). 1H NMR ([(CD3)2SO]): 8.80 br t,
1H; 8.0-7.4 br, ∼3H; 7.58 d, J = 9 Hz, 1H; 7.32 d, J = 7.8 Hz, 1H;
7.20 t, J = 7.2 Hz, 1H; 7.12 s, 1H; 4.00 s, 3H; 3.75-3.65 m, 1H;
(8) Balsamini, C.; Bedini, A.; Diamantini, G.; Spadoni, G.; Tontini, A.;
Tarzia, G.; Di Fabio, R.; Feriani, A.; Reggiani, A.; Tedesco, G.;
Valigi, R. (E)-3-(2-(N-Phenylcarbamoyl)vinyl)pyrrole-2-carboxylic
acid derivatives. A novel class of glycine site antagonists. J. Med.
Chem. 1998, 41, 808–820.
(9) Donati, D.; Di Fabio, R. Synthesis and pharmacological properties
of novel glycine antagonists. Pharm. Acta Helv. 2000, 74, 239–245.
(10) Poschenrieder, H.; Hofner, G.; Stachel, H. D. Grevellin analogs
with affinity to the N-methyl-D-aspartate (glycine site) receptor, a
novel lead structure. Arch. Pharm. (Weinheim, Ger.) 2000, 333,
211–216.
(11) Snell, L. D.; Claffey, D. J.; Ruth, J. A.; Valenzuela, C. F.; Cardoso,
R.; Wang, Z. J.; Levinson, S. R.; Sather, W. A.; Williamson, A. V.;
Ingersoll, N. C.; Ovchinnikova, L.; Bhave, S. V.; Hoffman, P. L.;
Tabakoff, B. Novel structure having antagonist actions at both the
glycine site of the N-methyl-D-aspartate receptor and neuronal
voltage-sensitive sodium channels: biochemical, electrophysiologi-
cal, and behavioral characterization. J. Pharmacol. Exp. Ther.
2000, 292, 215–227.
(12) Jansen, M.; Dannhardt, G. Antagonists and agonists at the glycine
site of the NMDA receptor for therapeutic interventions. Eur. J.
Med. Chem. 2003, 38, 661–670.
(13) Buchstaller, H.-P.; Siebert, C. D.; Steinmetz, R.; Frank, I.; Berger,
M. L.; Gottschlich, R.; Leibrock, J.; Krug, M.; Steinhilber, D.;
Noe, C. R. Synthesis of thieno[2,3-b]pyridinones acting as cycto-
protectants and as inhibitors of [3H]glycine binding to the
N-methyl-D-aspartate (NMDA) receptor. J. Med. Chem. 2006,
49, 864–871.
(14) Cai, S. X. Glycine/NMDA receptor antagonists as potential CNS
therapeutic agents: ACEA-1021 and related compounds. Curr.
Top. Med. Chem. 2006, 6, 651–662.
(15) Nagata, R.; Katayama, S.; Ohtani, K.; Tanaka, H.; Shimago, K.
Tricyclic quinoxalinediones, aza-kynurenic acids, and indole-2
carboxylic acids as in vivo active NMDA-glycine antagonists.
Curr. Top. Med. Chem. 2006, 6, 733–745.
(16) Catarzi, D.; Colotta, V.; Varano, F. Competitive Gly/NMDA
receptor antagonists. Curr. Top. Med. Chem. 2006, 6, 809–821.
(17) Bare, T. M.; Brown, D. G.; Horchler, C. L.; Murphy, M.; Urbanek,
R. A.; Alford, V.; Barlaam, C.; Dyroff, M. C.; Empfield, J. B.;
Forst, J. M.; Herzog, K. J.; Keith, R. A.; Kirschner, A. S.; Lee, C.
M. C.; Lewis, J.; McLaren, F. M.; Neilson, K. L.; Steelman, G. B.;
Trivedi, S.; Vacek, E. P.; Xiao, W. Pyridazinoquinolinetriones as
NMDA glycine-site antagonists with oral antinociceptive activity
in a model of neuropathic pain. J. Med. Chem. 2007, 50, 3113–3131.
(18) Newcomer, J. W.; Krystal, J. H. NMDA receptor regulation of
memory and behavior in humans. Hippocampus 2001, 11, 529–542.
(19) Goff, D. C.; Coyle, J. T. The emerging role of glutamate in the
pathophysiology and treatment of schizophrenia. Am. J. Psychia-
try 2001, 158, 1367–1377.
(20) Shim, S. S.; Hammonds, M. D.; Kee, B. S. Potentiation of the
NMDA receptor in the treatment of schizophrenia: focused on the
glycine site. Eur. Arch. Psychiatry Clin. Neurosci. 2008, 258, 16–27.
(21) Snell, L. D.; Morter, R. S.; Johnson, K. M. Structural requirements
for activation of the glycine receptor that modulates the N-methyl-
D-aspartate operated ion channel. Eur. J. Pharmacol. 1988, 156,
105–110.
(22) McBain, C. J.; Kleckner, N. W.; Wyrick, S.; Dingledine, R.
Structural requirements for activation of the glycine coagonist site
of N-methyl-D-aspartate receptors expressed in Xenopus oocytes.
Mol. Pharmacol. 1989, 36, 556–565.
(23) Watson, G. B.; Lanthorn, T. H. Pharmacological characteristics of
cyclic homologues of glycine at the N-methyl-D-aspartate receptor-
associated glycine site. Neuropharmacology 1990, 29, 727–730.
(24) Monahan, J. B.; Hood, W. F.; Compton, R. P.; Cordi, A. A.;
Williams, R. M. Identification of a novel structural class of positive
modulators of the N-methyl-D-aspartate receptor, with actions
mediated through the glycine recognition site. Eur. J. Pharmacol.,
Mol. Pharmacol. Sect. 1990, 189, 373–379.
3.60-3.40 m, 2H. Anal. (C13H14BrN3O3 0.4H2O) C, H, Br, N, O.
3
rac-2-Amino-3-{[4-(2-ethylphenyl)-1-methyl-1H-indole-2-car-
bonyl]amino}propionic Acid 53. Amounts of 3.5 g (13.05 mmol)
of 4-bromo-1-methyl-1H-indole-2-carboxylic acid methyl ester
52a and 2.15 g (14.3 mmol) of 2-ethylphenylboronic acid were
coupled according to method F, yielding a dark residue which
was purified by MPLC on silica gel. Elution with hexane/tert-
butyl methyl ether, 9:1, afforded 4.2 g (96%) of 4-(2-ethyl-
phenyl)-1-methyl-1H-indole-2-carboxylic acid methyl ester
53a as a colorless oil. 1H NMR (CDCl3): 7.45-7.30 m, 6H;
7.08-7.02 m, 1H; 7.00 s, 1H; 4.13 s, 3H; 3.85 s, 3H; 2.60-2.40 m,
2H; 1.00 t, J = 8 Hz, 3H. A mixture of 4 g of 53a, 27 mL of 1 M
NaOH solution, and 50 mL of methanol was refluxed for
30 min. After dilution with ice the reaction mixture was extracted
with diethyl ether and the aqueous phase acidified with 2 M HCl
solution. The precipitate was filtered and dried yielding 3.56 g
(93%) of 4-(2-ethylphenyl)-1-methyl-1H-indole-2-carboxylic
acid 53b as a white powder, mp 210-212 °C. 1H NMR
([(CD3)2SO]): 12.9 br s, 1H; 7.60 d, J = 7 Hz, 1H; 7.50-7.25
m, 4H; 7.20 d, J = 7 Hz, 1H; 7.00 d, J = 7 Hz, 1H; 4.10 s, 3H;
2.6-2.4 m, 2H; 0.95 t, J = 8 Hz, 3H. Condensation of 53b with
rac-5b (method B, 88%) and removal of the protecting groups
(method C, 100%; method D, 76%) gave 53, mp 220 °C (dec;
AcOH/THF/H2O, 2:2:1). 1H NMR ([(CD3)2SO]): 8.52 br t, 1H;
7.52 d, J = 8.4 Hz, 1H; 7.40-7.30 m, 3H; 7.25 d ꢀ t, J = 7 and 1
Hz, 1H; 7.20 d, J = 7 Hz, 1H; 6.95 d, J = 6 Hz, 1H; 6.75 s, 1H;
4.02 s, 3H; 3.68-3.58 m, 1H; 3.50-3.40 m, 2H; 2.60-2.40 m, 2H;
0.95 t, J = 7.2 Hz, 3H. Anal. (C21H23N3O3 0.1H2O) C, H, N, O.
3
Supporting Information Available: Syntheses and spectral
data of the target compounds and intermediates not described
in the Experimental Section; elemental analysis results of all
target compounds. This material is available free of charge via
the Internet at http://pubs.acs.org.
References
(1) Kemp, J. A.; McKernan, R. M. NMDA receptor pathways as drug
targets. Nat. Neurosci. 2002, 5, 1039–1042.
(2) Johnson, J. W.; Ascher, P. Glycine potentiates the NMDA
response in cultured mouse brain neurons. Nature 1987, 325,
529–531.
(3) Thomson, A. M. Glycine is a coagonist at the NMDA receptor/
channel complex. Prog. Neurobiol. 1990, 35, 53–74.
(4) Leeson, P. D.; Iversen, L. L. The glycine site on the NMDA
receptor: structure-activity relationships and therapeutic poten-
tial. J. Med. Chem. 1994, 37, 4053–4067.
(5) Danysz, W.; Parsons, C. G. Glycine and N-methyl-D-aspartate
receptors: physiological significance and possible therapeutic
applications. Pharmacol. Rev. 1998, 50, 597–664.
(6) Hood, W. F.; Gray, N. M.; Dappen, M. S.; Watson, G. B.;
Compton, R. P.; Cordi, A. A.; Lanthorn, T. H.; Monahan, J. B.
Characterization of indole-2-carboxylate derivatives as antago-
nists of N-methyl-D-aspartate receptor activity at the associated
glycine recognition site. J. Pharmacol. Exp. Ther. 1992, 262, 654–
660.
(7) Carling, R. W.; Leeson, P. D.; Moore, K. W.; Moyes, C. R.;
Duncton, M.; Hudson, M. L.; Baker, R.; Foster, A. C.; Grimwood,
S.; Kemp, J. A.; Marshall, G. R.; Tricklebank, M. D.; Saywell, K.
L. 4-Substituted-3-phenylquinolin-2(1H)-ones: acidic and nonaci-
dic glycine site N-methyl-D-aspartate antagonists with in vivo
activity. J. Med. Chem. 1997, 40, 754–765.
(25) Furukawa, H.; Gouaux, E. Mechanisms of activation, inhibition
and specificity: crystal structures of the NMDA receptor NR1
ligand-binding core. EMBO J. 2003, 22, 2873–2885.
(26) Aebischer, B.; Frey, P.; Haerter, H.; Herrling, P. L.; Mueller, W.
Synthesis and NMDA antagonistic properties of the enantiomers
of 4-(3-phosphonopropyl)piperazine-2-carboxylic acid (CPP) and
for the unsaturated analogue (E)-4-(3-phosphonoprop-2-enyl)-
piperazine-2-carboxylic acid (CPP-ene). Helv. Chim. Acta 1989,
72, 1043–1051.
(27) Moore, S.; Patel, R. P.; Atherton, E.; Kondo, M.; Meienhofer, J.;
Blau, L.; Bittman, R.; Johnson, R. K. Synthesis of actinomycin and
analogs. J. Med. Chem. 1976, 19, 766–772.
(28) Ishii, H.; Murakami, Y.; Ishikawa, T. Abnormal Fischer indoliza-
tion and its related compounds. XXII. Fischer indolization of ethyl

Article
pyruvate 2-(2-chloro- and 2,6-dichlorophenyl)methylhydrazones.
Chem. Pharm. Bull. 1990, 38, 597–604.
(29) Bergmann, J. Reaction of N-acylpyridinium salts with indole.
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 16 5107
(40) Foster, A. C.; Kemp, J. A. HA-966 antagonizes N-methyl-
D-aspartate receptors through a selective interaction with the
glycine modulatory site. J. Neurosci. 1989, 9, 2191–2196.
J. Heterocycl. Chem. 1970, 7, 1071–1076.
(41) Singh, L.; Donald, A. E.; Foster, A. C.; Hutson, P. H.; Iversen, L.
L.; Iversen, S. D.; Kemp, J. A.; Leeson, P. D.; Marshall, G. R.;
Oles, R. J.; Priestley, T.; Thorn, L.; Tricklebank, M. D.; Vass, C.
A.; Williams, B. J. Enantiomers of HA-966 (3-amino-1-hydroxy-
pyrrolid-2-one) exhibit distinct central nervous system effects: (þ)-
(30) Grimwood, S.; Wilde, G. J. C.; Foster, A. C. Interactions between
the glutamate and glycine recognition sites of the N-methyl-
D-aspartate receptor from rat brain, as revealed from radioligand
binding studies. J. Neurochem. 1993, 60, 1729–1738.
(31) Mugnaini, M.; Giberti, A.; Ratti, E.; Van Amsterdam, F. T. M.
Allosteric modulation of [3H]CGP39653 binding by glycine in rat
brain. J. Neurochem. 1993, 61, 1492–1497.
HA-966 is a selective glycine/N-methyl-D-aspartate receptor
antagonist, but (-)HA-966 is a potent γ-butyrolactone-like seda-
tive. Proc. Natl. Acad. Sci. U.S.A. 1990, 87, 347–351.
(32) Bonhaus, D. W.; Yeh, G. C.; Skaryak, L.; McNamara, J. O. Glycine
regulation of the N-methyl-D-aspartate receptor-gated ion channel
in hippocampal membranes. Mol. Pharmacol. 1989, 36, 273–279.
(33) Kloog, Y.; Nadler, V.; Sokolovsky, M. Mode of binding of [3H]di-
benzocycloalkenimine [MK-801] to the N-methyl-D-aspartate
(NMDA) receptor and its therapeutic implication. FEBS Lett.
1988, 230, 167–170.
(34) Hood, W. F.; Compton, R. P.; Monahan, J. B. D-Cycloserine, a
ligand for the N-methyl-D-aspartate coupled glycine receptor has
partial agonist characteristics. Neurosci. Lett. 1989, 98, 91–95.
(35) Berger, M. L. On the true affinity of glycine for its binding site at
the NMDA receptor complex. J. Pharmacol. Toxicol. Methods
1995, 34, 79–88.
(36) Foster, A. C.; Kemp, J. A.; Tricklebank, M. D.; Leeson, P. D.;
Carling, R. W.; Williams, B. J.; Grimwood, S.; Donald, A. E.;
Priestley, T.; Marshall, G. R.; Saywell, K.; Gill, R. Selective
antagonists for the glycine site of the N-methyl-D-aspartate recep-
tor. Mol. Neuropharmacol. 1992, 2, 97–102.
(37) Priestley, T.; Marshall, G. R.; Hill, R. G.; Kemp, J. A. L-687,414, a
low efficacy NMDA receptor glycine site partial agonist in vitro,
does not prevent hippocampal LTP in vivo at plasma levels known
to be neuroprotective. Br. J. Pharmacol. 1998, 124, 1767–1773.
(38) Arunlakshana, O.; Schild, H. O. Some quantitative uses of drug
antagonists. Br. J. Pharmacol. 1959, 14, 48–58.
(42) Inanobe, A.; Furukawa, H.; Gouaux, E. Mechanism of partial
agonist action at the NR1 subunit of NMDA receptors. Neuron
2005, 47, 71–84.
(43) Schneider, C. I.; Urwyler, S. Biochemical and thermodynamic
aspects of the binding of [3H]glycine to its strychnine-insensitive
recognition site associated with the N-methyl-D-aspartate receptor
complex. Biochem. Pharmacol. 1992, 43, 1693–1699.
(44) Sills, M. A.; Fagg, G.; Pozza, M.; Angst, C.; Brundish, D. E.; Hurt,
S. D.; Wilusz, E. J.; Williams, M. [3H]CGP39653: a new N-methyl-
D-aspartate antagonist radioligand with low nanomolar affinity in
rat brain. Eur. J. Pharmacol. 1991, 192, 19–24.
(45) Urwyler, S.; Laurie, D.; Lowe, D. A.; Meier, C. L.; Muller, W.
Biphenyl-derivatives of 2-amino-7-phosphonoheptanoic acid, a
novel class of potent competitive N-methyl-D-aspartate receptor
antagonists. 1. Pharmacological characterization in vitro. Neuro-
pharmacology 1996, 35, 643–654.
(46) Yoneda, Y.; Ogita, K.; Enomoto, R.; Suzuki, T.; Kito, S. Identi-
fication and characterization of specific binding sites of
[3H]spermidine in synaptic membranes of rat brain. Brain Res.
1991, 563, 17–27.
(47) Ahlrichs, R.; Bar, M.; Haser, M.; Horn, H.; Kolmel, C. Electronic
structure calculations on workstation computers: the program
system turbomole. Chem. Phys. Lett. 1989, 162, 165–169.
(48) Still, W. C.; Kahn, M.; Mitra, A. Rapid chromatographic techni-
que for preparative separations with moderate resolution. J. Org.
Chem. 1978, 43, 2923–2925.
(39) Watson, G. B.; Bolanowski, M. A.; Baganoff, M. P.; Deppeler, C.
L.; Lanthorn, T. H. D-Cycloserine acts as a partial agonist at the
glycine modulatory site of the NMDA receptor expressed in
Xenopus oocytes. Brain Res. 1990, 510, 158–160.
(49) Gunther, K. Thin-layer chromatographic enantiomeric resolution
via ligand exchange. J. Chromatogr. 1988, 448, 11–30.