Synthesis and benzodiazepine receptor (ω receptor) affinities of 3-substituted derivatives of pyrrolo[2,3-c]pyridine-5-carboxylate, a novel class of ω1 selective ligands

Article abstract of DOI:10.1016/S0968-0896(99)00053-X

Based on the structure of ZK91296 (4d), a high affinity partial agonist of the central benzodiazepine (ω) receptor, a series of pyrrolo[2,3-c]pyridine-5-carboxylate derivatives having mainly aralkyl and aralkyloxy substituents at C-3 was synthesized. The

Full text of DOI:10.1016/S0968-0896(99)00053-X

Bioorganic & Medicinal Chemistry 7 (1999) 921±932
Synthesis and Benzodiazepine Receptor (! Receptor) Anities of
3-Substituted Derivatives of Pyrrolo[2,3-c]pyridine-5-carboxylate,
a Novel Class of !1 Selective Ligandsy
Xavier Doisy,a Mouloud Dekhane,a Mireille Le Hyaric,a Jean-Francois Rousseau,a
c
Sunil K. Singh,a Suan Tan,b Valerie Guilleminot,b Hans Schoemaker,b Mireille Sevrin,b
Pascal George,b Pierre Potiera and Robert H. Dodda,*
aInstitut de Chimie des Substances Naturelles, Centre National de la Recherche Scienti®que, 91198 Gif-sur-Yvette Cedex, France
bSynthelabo Research (L.E.R.S.), 31, Avenue Paul Vaillant Couturier, 92220 Bagneux, France
Received 28 September 1998; accepted 30 December 1998
AbstractÐBased on the structure of ZK91296 (4d), a high anity partial agonist of the central benzodiazepine (o) receptor, a series
of pyrrolo[2,3-c]pyridine-5-carboxylate derivatives having mainly aralkyl and aralkyloxy substituents at C-3 was synthesized. The
in vitro binding anities of these compounds for three subclasses of the o receptor (o1, o2, o5) were determined using rat brain
tissue. Practically all of these compounds (except the diethyl ester derivative 22c) showed an approximately twofold selectivity for
o1 (IC50's in the 200±500 nM range) compared to o2 receptors and practically no anity for o5 receptors. Compound 22c showed
the highest anity of all the compounds synthesized (IC50=70 nM for o1 receptors) as well as a ®vefold selectivity for o1 versus o2
receptors but also displayed signi®cant binding to o5 receptors (IC50=250 nM). The absence of appreciable binding of 4-methyl and
4-methoxymethyl derivatives to o receptors, in contrast to b-carbolines having these similarly located substituents, suggests that the
pyrrolo[2,3-c]pyridine-5-carboxylates may be considered an entirely novel class of selective o receptor ligands. # 1999 Elsevier
Science Ltd. All rights reserved.
Introduction
groups.4 It also became clear that this recognition site,
now commonly referred to as the benzodiazepine
receptor (o receptor or BZR), in fact resided on the
GABAergic receptor complex itself, modulating the
activity of the latter via allosteric interactions.5 The use
of radioligand binding techniques (mainly with tritiated
diazepam) subsequently led to two major advances. The
®rst was the discovery of the benzodiazepine receptor
antagonist Ro 15-1788 (¯umazenil, 3).6 The second
advance was the observation that many classes of com-
pounds structurally unrelated to 1,4-benzodiazepines
could also bind with high anity to the benzodiazepine
receptor.2b While some of these compounds (e.g. the
triazolopyridazine CL 218,872,7 the pyrazoloquinoline
CGS 9896)8 exhibited to varying degrees the same
pharmacological properties as 1,4-benzodiazepines,
others, such as the methyl and ethyl esters of b-carbo-
line-3-carboxylates (b-CCM, 4a and b-CCE, 4b) actu-
ally produced behavioral e€ects opposite to those of
benzodiazepines. Thus, these b-carbolines were shown
in rodents to be respectively convulsant and pro-
convulsant as well as anxiogenic.9 These types of ligands
were termed inverse agonists to distinguish them from
the agonist benzodiazepines. At ®rst sight the pharma-
cological pro®le of inverse agonists would seem to
exclude them from possible therapeutic applications.
It has now been almost 40 years since the introduction
of the prototypic 1,4-benzodiazepines, chlordiazepoxide
(Librium1, 1) and diazepam (Valium1, 2) for clinical use
in the treatment of convulsions, anxiety and sleep dis-
orders.1 This class of compounds, now represented by a
myriad of structural analogues, is still today one of the
most highly prescribed drugs in the world.2 Since their
initial discovery, the ensuing study of the mechanism of
action of 1,4-benzodiazepines has been marked by a
number of key discoveries which have fueled further
intense interest in the area and, signi®cantly, led to the
development of more ecacious and selective drugs.
Thus, while an intimate connection between benzodia-
zepines and the g-aminobutyric acid (GABA) inhibitory
neurotransmitter system was convincingly demon-
strated in the early 1970s,3 it was not until 1977 that a
pharmacologically relevant recognition site for benzo-
diazepines was discovered independently by several
Key words: Pyrrolo[2,3-c]pyridine-5-carboxylate; benzodiazepine recep-
tor; selectivity; synthesis; modeling.
*Corresponding author. Tel.: 33-1-69-82-45-94; fax: 01-69-07-72-47;
e-mail: robert.dodd@icsn.cnrs-gif.fr
This article is dedicated to the memory of our friend and colleague,
Sir Derek Barton.
y
0968-0896/99/$ - see front matter # 1999 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(99)00053-X

922
X. Doisy et al. / Bioorg. Med. Chem. 7 (1999) 921±932
However, the demonstration that b-CCM10 facilitates
memorization in animal models of learning, again in
contrast to the well documented amnesic e€ects of ben-
zodiazepines, suggested that inverse agonists of the
BZR may in fact have a therapeutic future. The b-car-
bolines (as well as other compounds such as CL
218,872)7 were also of paramount importance in show-
ing that the benzodiazepine receptor, and by extension
the GABA receptor, exists in multiple forms.11 Various
in vitro and in vivo techniques convincingly showed that
b-carbolines bound to at least two di€erent subclasses
of benzodiazepine receptor, the BZ1 (or o1) and the
BZ2 (or o2) receptors, having di€erent molecular
weights, ligand anities and brain distributions.12 The
extremely powerful techniques of molecular biology and
receptor cloning led, toward the end of the 1980s, to a
much more complicated vision of the multiplicity of the
GABA/benzodiazepine receptor system.13 In the pre-
sently accepted model, the GABAA receptor is a supra-
molecular complex formed by ®ve protein subunits of
which at least 15 have so far been identi®ed by cloning
techniques. Based on the sequence homology of these
proteins, they have been classi®ed as a (6 isoforms), b (3
isoforms), g (3 isoforms), d and r (2 isoforms) subunits.
GABA/benzodiazepine receptor subtypes would then
result from the possibility of combining any ®ve of these
subunits into a functional entity.14 In reality, using
cloned subunits, it has been shown that the association
of a, b, and g subunits is the minimum requirement to
produce benzodiazepine-responsive recombinant recep-
tors.15 The anities, ecacies and presumably even the
intrinsic activities (agonist or inverse agonist) of BZR
ligands would then be controlled by the speci®c subunit
compositions of each GABAA receptor subtype.16 For
instance, in the case of the BZ1 (o1) subtype, the pre-
dominant subtype found in mammalian cerebellum, the
subunit composition corresponds to a1b2g2.14,17 How-
ever, little is known for the moment concerning either
the regional composition and stoichiometry of di€erent
GABAA receptor subtypes in the brain or their precise
physiological roles.
In order to obtain information concerning the con-
tribution of each of the major subtypes of GABAA
receptors to the multiple pharmacological properties of
BZR ligands (e.g. anticonvulsant, anxiolytic, sedative±
hypnotic, muscle relaxant and cognitive e€ects) the dis-
covery of subtype-selective ligands remains today a
major goal. Such selective ligands could also perhaps
alleviate some of the major inconveniences of benzo-
diazepines such as dependence, ethanol potentiation
and over-sedation. Some progress in this area has been
made over the years. Thus, zolpidem (5) shows good
selectivity for the o1 subtype of benzodiazepine recep-
tors displaying only intermediate anity for the o2
subtype (thought to be associated with GABAA recep-
tors containing the a2 or a3 subunits) and very low a-
nity for o5 subtype (thought to be associated with the
a5 subunit).18 Recently, ligands of type 6 selective for
a5-subunit containing receptors, were also described.19
With the objective of designing novel ligands of the
benzodiazepine receptor possessing, ideally, good a-
nities, subtype selectivities and partial agonist proper-
ties, we have concentrated our e€orts over the past
years on the study of compounds having the b-carboline
nucleus as a common motif.20 This seemed reasonable
in that it has been shown by others that in the form of
the 6- and 5-benzyloxy 4-methoxymethyl derivatives
ZK93423 and ZK91296 (4c and 4d, respectively), all
these qualities are united.21 High benzodiazepine recep-
tor anities for b-carbolines in general are assured by
(1) a completely aromatic nucleus, (2) a free indolic NH
function, (3) a pyridinyl nitrogen at the 2 position and
(4) a carbonyl group at C-3. Agonist, as opposed to
inverse agonist activity, is conferred by a combination
of substituents at C-4 (methoxymethyl) and C-5 or C-6
(benzyloxy).22 Interestingly, while the analogue of 4c
devoid of the C-4 methoxymethyl chain maintains a
good anity for the benzodiazepine receptor, the
resulting compound no longer displays agonist activity
in vivo.21b On the other hand, nothing is known about
the contribution of the A-ring of the b-carboline nucleus
to receptor anity and selectivity. In this paper, then,
we describe the synthesis of compounds having the 6-
azaindole-5-carboxylate nucleus (i.e., 7) as a common
structural unit. While this nucleus represents the b-car-
boline skeleton devoid of the A-ring, it still maintains
the minimal structural characteristics enumerated above
which allow binding of b-carbolines to the benzodiaze-
pine receptor. In order to provide further anchoring
points for the 6-azaindole derivatives on the binding
site, substituents were incorporated at C-3 mimicking
the benzyloxy moiety of ZK91296 (4d). Molecular
modeling studies de®ning the target molecules, synthesis

X. Doisy et al. / Bioorg. Med. Chem. 7 (1999) 921±932
923
of the latter and the determination of their binding
anities for three di€erent subclasses of the benzo-
diazepine receptor (o1, o2, o5) are described below.
Molecular modeling studies
By analogy with the active b-carboline ZK91296 (4d),
our intention was to prepare 6-azaindole-5-carboxylate
derivatives incorporating a phenyl group tethered to the
C-3 position. While the exocyclic phenyl group of the
rigid ZK91296 is relatively restricted in its movements, a
phenyl group attached to the azaindole nucleus by a
chain of three atoms would have considerably more
¯exibility. We thus decided to resort to simple computer
assisted molecular modeling to ensure that any pro-
posed structures had a reasonable chance, energetically-
speaking, of adopting a conformation close to that of
ZK91296, at least as far as the proposed phenyl group
was concerned. Figure 1 shows minimized structures for
ZK91296 obtained using Macromodel1 software as well
as for two candidate 3-substituted azaindole structures,
one having a benzyloxymethyl group at C-3 (8), the
other having a phenylpropene moiety at this position
(9). Superposition of either 8 or 9 with ZK91296 showed
surprisingly high positional concordance of the phenyl
groups of both azaindoles with that of the b-carboline,
especially for 9 (Fig. 2). Moreover, the tether functions
of each azaindole also adopt a spatial geometry which
mimics that of the C5-O-CH2 component of ZK91296.
Each of the proposed azaindole structures presents
certain characteristics which could be potentially
advantageous to benzodiazepine receptor binding.
Thus, azaindole 8 maintains the entire benzyloxy motif
and although it has never been demonstrated in
ZK91296 that the ether oxygen atom of the benzyloxy
group actually takes part in receptor interactions, it is
quite obvious that it by no means interferes with bind-
ing. On the other hand, azaindole 9, unlike 8, has a p-
electron system which, as originally planned, was meant
to partially recreate the p environment of the A-ring of
the b-carbolines, no doubt an integral part of the
receptor interactions of this class of high anity com-
pound. As it happens, though, our modeling studies
show that the exocyclic double bond of azaindole 9
adopts a position which precludes superposition with
the A ring of ZK91296. However, since it was felt that
Figure 1. Minimized structures of (A) azaindole 8, (B) ZK91296 (4d) and (C) azaindole 9.
Figure 2. (A) Superposition of ZK91296 (4d) and azaindole 8. (B) Superposition of ZK91296 (4d) and azaindole 9.

924
X. Doisy et al. / Bioorg. Med. Chem. 7 (1999) 921±932
both a p-electron system at C-3 and an ether oxygen
atom could be bene®cial to receptor binding, both types
of azaindole 8 and 9 were synthesized and tested for
benzodiazepine receptor anity.
phenethyl- and (carbethoxymethylene)triphenyl-phos-
phoranes provided the vinylic derivatives 21a±c. While
the diester 21c was obtained exclusively in the E form,
21a and 21b were isolated as inseparable mixtures of the
E and Z geometrical isomers, with the former pre-
dominating. Removal of the benzenesulfonyl protecting
group of 21a±c by treatment with sodium ethoxide in
ethanol a€orded azaindoles 22a±c. Both 22a and 22b
showed a tendency to isomerize to the E form when left
in solution (e.g. in CDCl3). Catalytic hydrogenation
(palladium on carbon) of the exocyclic double bond of
compounds 22a±c provided the corresponding saturated
derivatives 23a±c in good yield.
Chemistry
Two general routes were used to introduce substituents at
C-3 of the starting ethyl 6-azaindole-3-carboxylate (7b).23
In the ®rst, azaindole 7b was eciently transformed into
the azagramine derivative 10 by treatment with N,N-
dimethylmethyleneammonium iodide (Eschenmoser's
salt) in re¯uxing acetonitrile (Scheme 1). Compound 10
was then quantitatively converted into the N,O-diacetyl
derivative 11 by the action of sodium acetate in acetic
anhydride.24 Compound 11 subsequently served as the
reactive intermediate for the introduction at C-3 of
phenyl ring-containing side-chains of varying lengths.
Thus, 11 was dissolved in the given aralkyl alcohol
(benzyl, phenethyl alcohols and ¯uorinated derivatives
thereof as well as cinnamyl alcohol) and allowed to stir
for 2±18 h in the presence of excess solid potassium
carbonate and a trace of water, forming ethers 12±17.
N-Deacylation occurred concomitantly. The major
side-products in these reactions arose from transester-
i®cation of the ethyl ester of 7b by the aralkyl alcohol.
Reduction of the 3-formyl functionality of compound
19 using sodium borohydride in ethanol followed by N-
deprotection of the product (EtONa/EtOH) a€orded
the 3-hydroxymethyl azaindole 20 in 33% overall yield.
The synthesis of 4-methyl 6-azaindole-5-carboxylate 28
was accomplished using a methodology previously
developed by us in the b-carboline area.26 Thus, N-
benzenesulfonylpyrrole-2-carboxaldehyde (24) reacted
with methyl 2-amino-3,3-diethoxybutyrate in dichloro-
methane in the presence of triethylamine and molecular
sieves to give an imine which was directly reduced to
amine 26 with sodium cyanoborohydride (Scheme 3).
Compound 26 was then cyclized using titanium(IV)
chloride in re¯uxing benzene, a€ording the 4-methyl
azaindole derivative 27 in 65% overall yield (from 24).
N-Deprotection of 27 was e€ected by the action of
sodium methoxide in methanol, giving compound 28.
Treatment of the 4-methyl derivative 27 with N-bromo-
succinimide and AIBN in CCl4 produced the 4-bromo-
methyl compound 29 as the major product (61%).
Treatment of 29 with sodium methoxide in methanol
then resulted in simultaneous nucleophilic displacement
of the bromide and N-deprotection, forming the 4-
methoxymethyl azaindole derivative 30.
In the second approach, the 3-formyl-6-azaindole deri-
vative 18 was ®rst prepared from 7b by reaction with
1,1-dichloromethyl methyl ether in the presence of alu-
minum chloride25 (Scheme 2). Use of the more tradi-
tional Vilsmeier conditions for formylating indoles
(POCl3, DMF) failed to give any trace of 18 when
applied to azaindole 7b. Protection of the nitrogen
function of compound 18 with a benzenesulfonyl group
then gave 19. Wittig condensations of 19 with benzyl-,
In vitro binding assays
The concentrations of test compounds inhibiting 50%
of tritiated ¯umazenil binding (IC50) to selected rat
brain tissues were determined in vitro using previously
described techniques.18c Tissues were taken from the
cerebellum, containing mainly the BZ1 or o1-subtype of
benzodiazepine receptor (i.e., a1-subunit combining
receptors), from the spinal cord, containing BZ2 or o2
receptor subtypes (and which is actually a mixture of
a2-, a3- and a5-subunit combining receptors) and,
®nally, from the hippocampus, a source of mainly
a5-subunit containing receptors.17a,c,27 Results of the
binding assays are given in Table 1.
Results and Discussion
As shown in Table 1, ethyl 6-azaindole-5-carboxylate
(7b) itself has very mediocre anity for the BZR,
showing IC50s greater than 1 mM for all three subtypes
of receptor (o1, o2, o5). A simple substitution at C-3, in
the form of the formyl group, provides a compound (18)
displaying signi®cant binding to o1 and o2 receptors,
with a twofold selectivity for the former (0.17 versus
Scheme 1.

X. Doisy et al. / Bioorg. Med. Chem. 7 (1999) 921±932
925
Scheme 2.
0.39 mM, respectively). Interestingly, compound 18
showed practically no anity for o5 receptors (i.e.,
IC50>1 mM). Reduction of the formyl group of 18 to
the primary alcohol 20 abolished binding to o1 and o2
receptors, suggesting that a p-electron system at C-3 of
azaindoles may be favorable to receptor binding. Both
compounds 18 and 20 have oxygen atoms at C-3 though
in the formyl derivative 18, it is more likely to act as a
hydrogen-bond acceptor in any eventual interactions
with the receptor, in the same way as the equidistant
ether oxygen of ZK91296 (4d). The primary alcohol of
20 would, however, most probably behave as a hydro-
gen bond donor, an apparently receptor-disfavored
situation. This assumes, of course that the azaindole
occupies the same receptor pocket as a b-carboline such
as ZK91296 and that, moreover, the structural features
common to both types of compounds are oriented in the
same fashion with respect to the receptor. This assump-
tion can, however, be disputed based on results to be
discussed later. The binding pro®les of the aralkyl ether
series of compounds 12±15 were somewhat similar to
that of the 3-formyl derivative, showing an approxi-
mately twofold selectivity for the o1 versus the o2
receptor subclass and negligible binding to o5 receptors.
Thus, the simplest member of this group, the benzyloxy-
methyl derivative 12, displayed o1 and o2 binding
anities of 0.31 and 0.76 mM, respectively. Incor-
poration of an o-¯uoro substituent on the phenyl ring
(compound 13) produced a slight increase in anity for
the o1 receptor (IC50=0.24 mM) but a twofold increase
for o2 (IC50=0.32 mM) with the overall e€ect of a
decreased selectivity for the two receptor subclasses. On
the other hand, substitution of the phenyl ring with a
¯uorine atom at the meta position (compound 14) had
only a mild e€ect on anity for o1 and o2 receptors
compared to the unsubstituted analogue 12. Lengthen-
ing of the chain separating the phenyl group from the
azaindole nucleus led to substantial loss of binding a-
nity, as demonstrated by results with the phenethyl
derivative 15 (IC50=0.56 mM for o1) and the cinnamyl
derivative 17 (IC50=2.23 mM for o1). Surprisingly, an
attempt to improve the anity of the phenethyl com-
pound 15 by adding an o-¯uoro substituent (analogous
to the e€ect obtained by o-¯uorination of 12) led to
quite the opposite e€ect, the product, compound 16,
being almost completely inactive. The combined steric
e€ects of the extra carbon atom (relative to 12) and the
¯uorine atom apparently override any positive electro-
nic e€ects which the latter can o€er.
Comparison of the latter binding results with those in
which the ether linkage is absent (i.e., compounds 22a,b,
23a,b) shows that for the same tether length between the
aromatic and heteroaromatic rings, equivalent or even
superior anities are obtained in the absence of the
ether oxygen function, while the rank order of selectiv-
ities (o1>o2 ꢀo5) is maintained. This e€ect is particu-
larly ¯agrant in compound 22b, in which the oxygen
atom of the ether derivative 12 is replaced by an sp2
carbon atom. Not only is the anity of compound 22b

926
X. Doisy et al. / Bioorg. Med. Chem. 7 (1999) 921±932
for o1 receptors improved compared to compound 12
(0.21 mM versus 0.31 mM), but its selectivity for this
receptor subclass is also enhanced. Reduction of the
double bond of 22b (to give 23b) maintains the same
relative selectivity but produces some loss in o1 receptor
anity (IC50=0.35 mM). Although this result would
appear to lend credence to our working hypothesis that
p-electron systems in this part of the molecule may tend
to enhance receptor binding, a paradoxically di€erent
e€ect is observed in the case of the one-carbon shorter
analogues 22a and 23a. In this case, saturation of the
double bond leads to a minor gain in o1 receptor anity,
but an equally small loss in o2 receptor anity. The e€ects
of ¯uoro substituents on the aromatic rings of compounds
22a,b and 23a,b have not yet been investigated.
Two results were obtained, however, which have caused
us to question our original premise that the 6-azaindole-
5-carboxylate nucleus is recognized by the BZR in the
same manner as b-carboline-3-carboxylates. Firstly, the
most active of the 3-substituted azaindoles prepared in
this study was actually the diethyl ester derivative 22c
(IC50=70 nM for o1), which is devoid of any phenyl
group on the C-3 side-chain and is thus structurally very
di€erent from ZK91296 (4d), the molecular template
serving as the starting point in our study. While it can
be argued that the p-electron system of the ester func-
tion at C-3 of compound 22c can mimic that of the C-5
benzyloxy group of ZK91296, interactions of these two
functions with the receptor would probably be of a very
di€erent nature. Interestingly, while compound 22c also
shows the highest o1 to o2 selectivity of all the com-
pounds synthesized (greater than ®vefold), it is also the
only azaindole derivative displaying any signi®cant
Scheme 3.
Table 1. In vitro o receptor binding anities of the synthesized 3-substituted 6-azaindole-5-carboxylate derivatives
Compound
R1
R2
R3
IC50 (mM)a
o1
o2
o5
7b
12
13
14
15
16
17
18
Et
Et
Et
Et
Et
Et
Et
Et
Et
Et
Et
Et
Et
Et
Et
Me
Me
(5)
(4b)
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
>1
0.31
0.24
0.39
0.56
>1
2.23
0.17
>1
0.37
0.21
0.07
0.32
0.35
0.13
>1
>10
0.014
0.0007
>1
0.76
0.32
0.60
1.08
>1
5.28
0.39
>1
0.55
>1
0.37
0.62
>1
>1
>10
ꢁ10
>10
ꢁ10
>1
>10
>1
>1
>1
>1
0.25
>1
>1
>1
>1
PhCH2OCH2
o-FPhCH2OCH2
m-FPhCH2OCH2
PhCH2CH2OCH2
o-FPhCH2CH2OCH2
PhCH=CHCH2OCH2
CHO
b
20
HOCH2
22a
22b
22c
23a
23b
23c
28
30
Zolpidem
b-CCE
PhCH=CHb
PhCH2CH=CHb
(E)-EtOCOCH=CH
PhCH2CH2
PhCH2CH2CH2
EtOCOCH2CH2
H
H
CH3
CH2OCH3
0.46
>1
>10
0.130
0.0061
H
>10
>10
0.13
aDetermined by the method of refs 18c and 29, see Experimental.
bPredominantly the E geometrical isomer.

X. Doisy et al. / Bioorg. Med. Chem. 7 (1999) 921±932
927
binding to o5 receptors. Saturation of the double bond
of 22c, to give 23c, abolishes anity for o5 receptors
while anity for o1 receptors is diminished only twofold.
Electron impact and chemical ionization mass spectra
were recorded on an AEI MS-50 and AEI MS-9 spec-
trometer, respectively. High-resolution mass spectra
were obtained using a Kratos MS-80 spectrometer.
Thin-layer chromatography was performed on Merck
silica gel plates with ¯uorescent indicator. The plates
were visualized with UV light (254 nm) or with a 3.5%
solution of phosphomolybdic acid in ethanol. All col-
umn chromatography was conducted on Merck 60
silica gel (230±400 mesh) at medium pressure (200
mbar). All reagents were purchased from the Aldrich
Chemical Co. and were used without further puri®ca-
tion. Elemental analyses were performed at the ICSN,
CNRS, Gif-sur-Yvette.
Finally, very unexpected results were obtained with the
4-methyl and 4-methoxymethyl azaindoles 28 and 30,
respectively. Neither of these compounds demonstrated
any signi®cant binding to any of the three BZR sub-
classes studied. This is a major di€erence with the well-
documented structure±activity relationships of b-carbo-
line-3-carboxylates in which such groups are known not
to interfere with binding to the BZR and in fact can
actually enhance binding.20c,21b,28
Conclusion
Ethyl 3-N,N-dimethylaminomethyl-1H-pyrrolo[2,3-c]pyr-
idine-5-carboxylate (10). A solution of compound 7b
(410 mg, 2.15 mmol) and N,N-dimethylmethylene-
ammonium iodide (439 mg, 2.37 mmol) in acetonitrile
(25 mL) was re¯uxed for 12 h under argon. The reaction
mixture was cooled, concentrated to near dryness under
reduced pressure and the residue was taken up in
CH2Cl2 and EtOH (25 mL of a 9/1 mixture). Saturated
aqueous NaHCO3 (10 mL) was added, the organic layer
was separated and the aqueous layer was extracted with
9/1 CH2Cl2/EtOH (3Â10 mL). The organic extracts
were combined, dried (Na2SO4) and evaporated under
reduced pressure, leaving a white solid (430 mg, 81%)
which crystallized in CHCl3: mp 128±130 C. H NMR
(200 MHz, CDCl3) d 1.46 (t, J=7.2, 3H), 2.34 (s, 6H),
4.03 (s, 2H), 4.49 (q, J=7.2, 2H), 7.67 (s, 1H), 8.57 (s,
1H), 9.12 (s, 1H), 12.81 (s, 1H, exchangeable with D2O);
13C NMR (62.5 MHz, CDCl3) d 17.0, 47.5, 54.0, 69.4,
113.1, 133.8, 135.2, 136.7, 137.8, 139.0, 166.7; IR (®lm)
1714 cm 1. HREIMS calcd for C13H17N3O2 m/z
247.1319, found 247.1320.
The 3-substituted 6-azaindole-5-carboxylate derivatives
described in this study represent a new class of benzo-
diazepine (o) receptor ligands showing selectivity for the
o1 subclass of receptors. While none of these com-
pounds simultaneously shows as high an anity and an
o1 receptor selectivity as zolpidem (5), certain deriva-
tives do demonstrate separately one or the other of
these properties. Thus, the diethyl ester derivative 22c
possesses a good anity for o1 receptors (IC50=70 nM,
compared to 14 nM for zolpidem) as well as a ®vefold
selectivity for this receptor subclass with respect to o2
receptors (compared to a ninefold selectivity for zolpi-
dem), but also shows appreciable binding to o5 recep-
tors. On the other hand, both the 3-phenylpropenyl and
3-phenylpropyl derivatives (22b and 23b, respectively)
exhibit only modest anities for o1 receptors
(IC50=210 and 350 nM, respectively) but very good
selectivity for this receptor subclass. Further structure±
activity studies in this series should help pinpoint the
structural characteristics which in¯uence both of these
properties. In this regard, while the high anity b-car-
boline partial agonist ZK91296 (4d) originally served as
our model in choosing the aralkyl substituents intro-
duced at C-3 of the azaindole nucleus, it is now appar-
ent, in view of some of the results obtained in this study
(i.e., in particular, the unexpected lack of receptor
binding by the 4-substituted azaindoles 28 and 30) that
the structural elements common to both the azaindoles
and the b-carbolines are recognized by the o receptor in
a di€erent manner. This implies that, in future work,
any proposed structural modi®cations of these azaindoles,
aimed at improving their anities and selectivities, will
not necessarily be guided by the well-documented struc-
ture±activity relationships in the b-carboline family of o
receptor ligands.
ꢂ
1
Ethyl 3-acetoxymethyl-1-acetylpyrrolo[2,3-c]pyridine-5-
carboxylate (11). A solution of the azagramine 10
(100 mg, 0.4 mmol) in anhydrous acetic anhydride
(1.5 mL) containing dry sodium acetate (76.5 mg,
0.93mmol) was re¯uxed for 4 h under argon. The reaction
mixture was cooled, saturated aqueous NaHCO3 (5 mL)
was added and the solution was extracted with ethyl
acetate (3Â5 mL). The organic extracts were combined,
dried (Na2SO4) and evaporated under reduced pressure.
Traces of acetic anhydride were removed by repeated
co-evaporation of the residue with toluene. The residual
white solid (101 mg, 82%) could be used in the follow-
ing steps without further puri®cation: 1H NMR
(250 MHz, CDCl3) d 1.48 (t, J=7.2, 3H), 2.09 (s, 3H),
2.70 (s, 3H), 4.51 (q, J=7.2, 2H), 5.28 (s, 2H), 7.73 (s,
1H), 8.44 (s, 1H), 9.75 (s, 1H); 13C NMR (62.5 MHz,
CDCl3) d 14.5, 20.9, 23.6, 56.9, 62.0, 116.6, 117.3, 129.0,
134.0, 135.3, 138.4, 142.1, 165.6, 167.9, 170.8; IR (®lm)
1735, 1725, 1709 cm 1; EIMS m/z 304 (M)+. Anal.
(C15H16N2O5) C, H, N.
Experimental
Chemistry
General procedures. Melting points were determined on
a Buchi apparatus and are uncorrected. IR spectra of
samples were obtained either as KBr pellets or as ®lms
Ethyl 3-(benzyloxymethyl)-1H-pyrrolo[2,3-c]pyridine-5-
carboxylate (12). A mixture of compound 11 (100 mg,
0.32 mmol), benzyl alcohol (1.5 mL), solid potassium
carbonate (45 mg) and water (0.5 mL) was stirred for
10 h at rt. The reaction mixture was diluted with
1
with a Nicolet 205 FT-IR spectrometer. H NMR and
13C-NMR were determined on a Bruker 200, 250, or
300 MHz instrument. Chemical shifts are given as d
values with reference to Me4Si as internal standard.

928
X. Doisy et al. / Bioorg. Med. Chem. 7 (1999) 921±932
dichloromethane (25 mL) and the solution was washed
with water (2Â5 mL). The organic layer was dried
(Na2SO4), the solvents were removed under reduced
pressure and the residue was puri®ed by column chro-
matography on silica gel (CH2Cl2/EtOH, 95/5) a€ord-
ing compound 12 as a white solid (50 mg, 50%): mp
187±189 ꢂC. 1H NMR (200 MHz, CDCl3) d 1.48 (t,
J=7.2, 3H), 4.52 (q, J=7.2, 2H), 4.60 (s, 2H), 4.84 (s,
2H), 7.37 (m, 5H), 7.78 (s, 1H), 8.61 (s, 1H), 9.23 (s,
1H), 12.12 (br s, 1H, exchangeable with D2O); 13C
NMR (62.5 MHz, CDCl3) d 14.7, 61.5, 63.7, 72.0, 114.2,
118.0, 127.8, 128.1, 128.6, 131.1, 132.4, 134.0, 135.6,
137.1, 179.3; IR (®lm) 2870, 1719 cm 1; EIMS m/z 310
(M)+. Anal. (C18H18N2O3.1/6 H2O) C, H, N.
a€orded compound 15 as a white solid (65 mg, 40%):
mp 162±164 ꢂC. H NMR (250 MHz, CDCl3) d 1.50 (t,
1
J=7.2, 3H), 2.94 (t, J=7.5, 2H), 3.74 (t, J=7.5, 2H),
4.55 (q, J=7.2, 2H), 4.81 (s, 2H), 7.23 (m, 5H), 7.77 (s,
1H), 8.59 (s, 1H), 9.24 (s, 1H), 12.73 (br s, 1H,
exchangeable with D2O); 13C NMR (62.5 MHz, CDCl3)
d 14.6, 36.3, 61.4, 64.4, 70.7, 113.9, 117.8, 126.1, 128.3,
128.9, 131.5, 132.2, 133.7, 135.5, 136.4, 138.9, 166.5; IR
(KBr) 3217, 1721 cm 1; EIMS m/z 324 (M)+. Anal.
(C19H20N2O3.H2O) C, H, N.
Ethyl 3-[2-(o-¯uorophenyl)ethyloxymethyl]-1H-pyrrolo-
[2,3-c]pyridine-5-carboxylate (16). A mixture of com-
pound 11 (186 mg, 0.61 mmol), o-¯uorophenethyl alco-
hol (1.5 mL), K2CO3 (50 mg) and water (50 mL) was
stirred overnight at rt. The reaction mixture was diluted
with ethyl acetate (10 mL) and the solution was washed
with water (3Â5 mL). The organic phase was dried
(Na2SO4) and evaporated under reduced pressure. The
residue was puri®ed by column chromatography on
silica gel using CH2Cl2 as developer until complete elu-
tion of the excess o-¯uorophenethyl alcohol. Further
elution of the column with CH2Cl2/EtOH (95/5) a€or-
ded compound 16 as a white solid (65 mg, 31%) which
was crystallized from CH2Cl2/pentane: mp 170±171 ꢂC.
1H NMR (250 MHz, CDCl3) d 1.48 (t, J=7.2, 3H), 2.97
(t, J=6.9, 2H), 3.75 (t, J=6.9, 2H), 4.54 (q, J=7.2,
2H), 4.78 (s, 2H), 7.01±7.25 (m, 4H), 7.75 (s, 1H), 8.57
(s, 1H), 9.21 (s, 1H), 12.56 (br s, 1H, exchangeable with
D2O); IR (KBr) 3061, 1717 cm 1; CIMS m/z 343
(MH)+. Anal. (C19H19N2O3F) C, H, N, F.
Ethyl 3-(o-¯uorobenzyloxymethyl)-1H-pyrrolo[2,3-c]pyr-
idine-5-carboxylate (13). A mixture of compound 11
(250 mg, 0.82 mmol), o-¯uorobenzyl alcohol (1 mL),
K2CO3 (80 mg) and water (100 mL) was stirred at rt for
4 h. The reaction mixture was then applied directly to a
column of silica gel which was ®rst developed with hep-
tane until complete elution of the excess o-¯uorobenzyl
alcohol. Further elution of the column with CH2Cl2/
EtOH (95/5) a€orded compound 13 as a white solid
(91 mg, 34%) which was crystallized in EtOH: mp 195±
ꢂ
1
197 C. H NMR (250 MHz, CDCl3) d 1.49 (t, J=7.2,
3H), 4.53 (q, J=7.2, 2H), 4.68 (s, 2H), 4.87 (s, 2H), 7.09
(t, J=9.5, 1H), 7.14 (t, J=7.4, 1H), 7.28 (m, 1H, par-
tially hidden by CHCl3), 7.45 (dt, J=1.5 and 7.4, 1H),
7.74 (d, J=1.5, 1H), 8.62 (s, 1H), 9.16 (s, 1H), 11.71 (br
1
s, 1H, exchangeable with D2O); IR (KBr) 1720 cm
;
EIMS m/z 328 (M)+. Anal (C18H17N2O3F.1/3 H2O) C,
H, N.
Ethyl 3-[3-(phenyl)-2,3-propenoxymethyl]-1H-pyrrolo-
[2,3-c]pyridine-5-carboxylate (17). A mixture of com-
pound 11 (170 mg, 0.55 mmol), cinnamyl alcohol
(2 mL), potassium carbonate (60 mg) and water (50 mL)
was stirred for 2 h at 40 ꢂC. The reaction mixture was
cooled to rt and applied to a column of silica gel which
was developed with heptane/ethyl acetate (8/2) until
complete elution of the excess cinnamyl alcohol. Further
elution of the column with dichloromethane/ethanol
(95/5) provided compound 17 as an o€-white solid
(101 mg, 55%): mp 169±171 ꢂC. 1H NMR (250 MHz,
CDCl3) d 1.45 (t, J=7.0, 3H), 4.25 (d, J=6.0, 2H), 4.51
(q, J=7.0, 2H), 4.86 (s, 2H), 6.35 (sex, J=6.0 and 15.6,
1H), 6.65 (d, J=15.6, 1H), 7.23±7.41 (m, 5H), 7.78 (s,
Ethyl 3-(m-¯uorobenzyloxymethyl)-1H-pyrrolo[2,3-c]pyr-
idine-5-carboxylate (14). A mixture of compound 11
(250 mg, 0.82 mmol), m-¯uorobenzyl alcohol (1 mL),
K2CO3 (80 mg) and water (100 mL) was stirred at rt for
18 h. The reaction mixture was then passed through a
column of silica gel, which was developed ®rst with
hexane until complete elution of the excess m-¯uoro-
benzyl alcohol and afterwards with CH2Cl2/EtOH (95/5),
a€ording compound 14 (134 mg, 50%). The latter was
obtained as white crystals from ethanol: mp 194±195 ꢂC.
1H NMR (250 MHz, CDCl3) d 1.49 (t, J=7.2, 3H), 4.53
(q, J=7.2, 2H), 4.58 (s, 2H), 4.86 (s, 2H), 6.98±7.14 (m,
4H), 7.80 (s, 1H), 8.61 (s, 1H), 9.24 (s, 1H), 12.36 (br s,
1
1H), 8.64 (s, 1H), 9.21 (s, 1H), 12.13 (br s, 1H,
1
1H, exchangeable with D2O); IR (KBr) 1711 cm
;
exchangeable with D2O); IR (®lm) 3386, 1722 cm
;
EIMS m/z 328 (M)+. Anal. (C18H17 N2O3F.1/3 H2O) C,
H, N.
EIMS m/z 336 (M)+. Anal. (C20H20N2O3.0.5 H2O) C,
H, N.
Ethyl 3-[2-(phenyl)ethyloxymethyl]-1H-pyrrolo[2,3-c]pyr-
idine-5-carboxylate (15). A mixture of compound 11
(151 mg, 0.5 mmol), phenethyl alcohol (2.5 mL), K2CO3
(41 mg) and water (50 mL) was stirred for 7 h at rt.
Water (10 mL) was added to the reaction mixture and
the latter was extracted with CH2Cl2/EtOH (3Â5 mL of
a 95/5 mixture). The organic extracts were combined,
dried over Na2SO4 and evaporated under reduced pres-
sure. The residue was puri®ed by column chromato-
graphy on silica gel using heptane as developer until
complete elution of the excess phenethyl alcohol. Fur-
ther elution of the column with CH2Cl2/EtOH (9/1)
Ethyl 3-formyl-1H-pyrrolo[2,3-c]pyridine-5-carboxylate
(18). To a stirring solution of compound 7b (600 mg,
3.16 mmol) in nitromethane and 1,2-dichloroethane
(25 mL of a 1:1 mixture) held at 0 ꢂC under argon was
added 1,1-dichloromethyl methyl ether (1.44 mL,
15.8 mmol) and, in small portions, aluminum chloride
(1.42 g). The addition of the same quantities of dichloro-
methyl methyl ether and aluminum chloride was repe-
ated twice at 30 min intervals and the reaction mixture
was left to stir for a further 30 min. Water (20 mL) was
added followed by saturated aqueous NaHCO3
(100 mL). The mixture was extracted with a solution of

X. Doisy et al. / Bioorg. Med. Chem. 7 (1999) 921±932
929
CH2Cl2/EtOH (4Â100 mL of a 9/1 mixture), the organic
extracts were combined, washed with aqueous NaCl
(50 mL) and dried over Na2SO4. The solvents were
removed under reduced pressure and the residue was
crystallized in ethanol, a€ording compound 18 as tan
crystals (507 mg, 74%): mp 150 ꢂC. 1H NMR (200 MHz,
DMSO-d6) d 1.35 (t, J=7.1, 3H), 4.35 (q, J=7.1, 2H),
8.75 (s, 1H), 8.87 (s, 1H), 9.07 (s, 1H), 10.05 (s, 1H),
12.80 (br s, 1H, exchangeable with D2O); 13C NMR
(62.5 MHz, DMSO-d6) d 14.2, 60.7, 117.7, 118.0, 128.9,
134.4, 135.1, 139.6, 141.7, 165.2, 185.7; IR (KBr) 3345,
1714, 1676 cm 1; CIMS m/z 219 (MH)+. Anal. (C11H10
N2O3.0.5 H2O) C, H, N.
51.7, 54.7, 117.1, 117.4, 128.4, 134.0, 134.7, 135.9, 148.4,
165.2; IR (KBr) 3365, 1713 cm 1; CIMS m/z 221
(MH)+. Anal. (C11H12N2O3.0.2H2O) C, H, N.
(E,Z)-Ethyl 1-benzenesulfonyl-3-(2-phenylethylene)pyrrolo-
[2,3-c]pyridine-5-carboxylate (21a). A solution of com-
pound 19 (20 mg, 0.06 mmol), benzyltriphenylphos-
phonium bromide and sodium amide (42 mg of a
commercial packaged mixture; 0.08 mmol) in anhydrous
THF (2 mL) was stirred for 2 h at rt. Fresh benzyl-
triphenylphosphonium bromide/sodium amide mixture
(14 mg, 0.03 mmol) was then added and stirring was
continued for 2 h. Water (100 mL) was added and the
solution was extracted with CH2Cl2 (3Â100 mL). The
combined organic extracts were dried (Na2SO4), the
solvents were removed under reduced pressure and the
residue was puri®ed by preparative TLC on silica gel
(heptane/ethyl acetate, 1/1) a€ording compound 19, a
reddish solid, as an inseparable mixture of E and Z iso-
mers (3/1) (17 mg, 70%): mp 80 ꢂC. 1H NMR (200 MHz,
CDCl3) d 1.42 (t, J=7.1, 0.75H), 1.46 (t, J=7.1, 2.25H),
4.43 (q, J=7.1, 0.5 H), 4.55 (q, J=7.1, 1.5H), 6.56 (d,
J=11.8, 0.25H), 6.86 (d, J=11.8, 0.25H), 7.16±7.60 (m,
ꢁ10.75 H), 7.90 (d, J=15.1, 0.75 H), 7.98 (s, 0.75H),
8.10 (s, 0.25H), 8.64 (s, 1H), 9.36 (s, 0.25H), 9.43 (s,
0.75H); 13C NMR (62.5 MHz, DMSO-d6) d 14.4, 61.8
(E), 62.0 (Z), 117.0 (E), 117.7 (Z), 120.2, 120.9, 126.4,
127.0, 127.9, 128.8, 129.6, 129.7, 131.8, 133.4, 133.9,
134.6, 134.7, 134.9, 135.7, 137.2, 141.0, 141.7, 165.4; IR
(KBr) 3205, 1717 cm 1; CIMS m/z 433 (MH)+. Anal.
(C24H20N2O4S.0.5 H2O) C, H, N.
Ethyl 1-benzenesulfonyl-3-formylpyrrolo[2,3-c]pyridine-
5-carboxylate (19).
A solution of compound 18
(300 mg, 1.4 mmol) in anhydrous THF (70 mL) was
treated at rt with sodium hydride (70 mg of a 60% dis-
persion in oil; 1.6 mmol). The mixture was stirred for
1.5 h and a solution of benzenesulfonyl chloride
(0.21 mL, 1.6 mmol)) in THF (5 mL) was added drop-
wise. The reaction mixture was then stirred for 2 h,
water (60 mL) was added and the solution was extracted
with CH2Cl2/EtOH (5Â100 mL of a 9/1 mixture). The
organic extracts were combined, washed with saturated
aqueous NaCl (100 mL) and dried over Na2SO4.
Removal of the solvents under reduced pressure left a
crude product which was puri®ed by column chroma-
tography on silica gel (CH2Cl2/EtOH, 98/2), a€ording
compound 19 as a white powder (390 mg, 78%): mp
160 ꢂC (heptane). 1H NMR (300 MHz, CDCl3) d 1.42 (t,
J=7.1, 3H), 4.47 (q, J=7.1, 2H), 7.54 (t, J=7.7, 2H),
7.65 (t, J=7.7, 1H), 7.99 (d, J=7.7, 2H), 8.42 (s, 1H),
8.93 (s, 1H), 9.39 (s, 1H), 10.14 (s, 1H); 13C NMR
(62.5 MHz, DMSO-d6) d 14.4, 62.2, 119.3, 121.7, 127.5,
130.2, 132.8, 133.1, 135.5, 135.7, 136.5, 138.5, 143.2,
165.2, 184.4; IR (®lm) 3119, 1713, 1676 cm 1; CIMS m/z
359 (MH)+. Anal. (C17H14N2O5S.0.3H2O) C, H, N, S.
(E,Z)-Ethyl 1-benzenesulfonyl-3-(3-phenyl-1,2-propenyl)-
pyrrolo[2,3-c]pyridine-5-carboxylate (21b). A solution of
phenethyltriphenylphosphonium bromide (37.5 mg,
0.08 mmol) in anhydrous THF was treated at 20 ꢂC
under argon with a solution of n-butyllithium in hexane
(52 mL of a 1.6 M solution, 0.08 mmol). The reaction
mixture was stirred for 1.5 h and a solution of com-
pound 19 (20 mg, 0.06 mmol) in THF (1 mL) was added
dropwise. The reaction mixture was allowed to come to
rt and stirring was continued for 3 h. Ethanol (0.5 mL)
was added, the solution was evaporated to dryness
under reduced pressure and the residue was puri®ed by
preparative TLC on silica gel (heptane/ethyl acetate, 3/
2). Compound 21b was isolated as a yellowish solid
(20 mg, 82%) in the form of a mixture of E and Z iso-
Ethyl 3-hydroxymethyl-1H-pyrrolo[2,3-c]pyridine-5-car-
boxylate (20). A solution of compound 19 (50 mg,
0.14 mmol) in ethanol (10 mL) was treated at 0 ꢂC under
argon with sodium borohydride (5.3 mg, 0.14 mmol).
The reaction mixture was stirred for 1 h, acetic acid
(0.2 mL) and water (10 mL) were added and the solution
was extracted with CH2Cl2/EtOH (4Â10 mL of a 9/1
mixture). The combined organic extracts were dried
(Na2SO4), the solvents were removed under reduced
pressure and the residue was dissolved in anhydrous
ethanol (10 mL). The solution was cooled to 0 ꢂC and a
0.4 M solution of sodium in ethanol (1 mL) was added
dropwise. The reaction mixture was stirred for 1 h, ace-
tic acid (0.2 mL) and water (20 mL) were added and the
solution was extracted with CH2Cl2/EtOH (6Â50 mL of
a 9/1 mixture). The combined organic extracts were
dried (Na2SO4), the solvents were removed in vacuo and
the residue was puri®ed by preparative TLC on silica gel
(dichloromethane/ethanol, 9/1) a€ording compound 20
as a white powder (10 mg, 33%): mp 250 ꢂC (decomp).
1H NMR (200 MHz, DMSO-d6) d 1.46 (t, J=7.1, 3H),
4.45 (q, J=7.1, 2H), 4.80 (d, J=5.6, 2H), 7.74 (s, 1H),
8.55 (s, 1H), 8.90 (s, 1H), 11.89 (br s, 1H, exchangeable
with D2O); 13C NMR (62.5 MHz, DMSO-d6) d 14.2,
mers (4/1): mp 91 ꢂC. H NMR (200 MHz, CDCl3) d
1
1.45 (t, J=7.2, 3H), 3.60 (d, J=5.0, 0.4H), 3.65 (d,
J=7.6, 1.6H), 4.51 (q, J=7.1, 2H), 6.14 (m, 1H), 6.55
(m, 1H), 7.20±7.60 (m, 8H), 7.70 (s, 0.2H), 7.73 (s,
0.8H), 7.87 (d, J=7.7, 2H), 8.42 (s, 0.8H), 8.51 (s,
0.2H), 9.40 (s, 0.2H), 9.41 (s, 0.8H); 13C NMR (75 MHz,
CDCl3) d 14.5 (E), 14.7 (Z), 35.7, 62.0, 117.2 (E), 117.4
(E), 118.8 (Z), 119.8 (Z), 126.6, 127.0, 127.1, 127.3,
128.2, 128.3, 128.6, 128.7, 128.8, 129.0, 129.2, 129.3,
129.8, 132.0, 133.2, 133.8, 134.3, 134.8, 135.6, 135.7,
1
136.9, 137.4, 139.2, 165.5; IR (®lm) 2982, 1716 cm
HRCIMS calcd for C25H23N2O4S m/z 447.1378, found
;
447.1365.
(E)-Ethyl 1-benzenesulfonyl-3-[2-(ethoxycarbonyl)ethyl-
ene]pyrrolo[2,3-c]pyridine-5-carboxylate (21c). A solution

930
X. Doisy et al. / Bioorg. Med. Chem. 7 (1999) 921±932
of compound 19 (50 mg, 0.14 mmol) in anhydrous THF
(10 mL) was treated at rt under argon with (car-
(E)-Ethyl 3-[2-(ethoxycarbonyl)ethylene]-1H-pyrrolo[2,3-
c]pyridine-5-carboxylate (22c). Compound 22c was
obtained as white crystals (116 mg, 72%) from com-
pound 21c (200 mg, 0.56 mmol): mp 290 ꢂC (ethanol)
(decomp). 1H NMR (250 MHz, CDCl3) d 1.36 (t,
J=7.1, 3H), 1.47 (t, J=7.1, 3H), 4.31 (q, J=7.1, 2H),
4.48 (q, J=7.1, 2H), 6.58 (d, J=16.0, 1H), 8.03 (d,
J=16.0, 1H), 8.43 (s, 1H), 8.66 (s, 1H), 9.00 (s, 1H),
12.77 (br s, 1H, exchangeable with D2O); IR (KBr)
2983, 1716, 1634 cm 1; CIMS m/z 289 (MH)+. Anal.
(C15H16N2O4.0.3 H2O) C, H, N.
(carbethoxymethylene)triphenylphosphorane
(73 mg,
0.21 mmol). The reaction mixture was stirred for 30 min,
acetic acid (0.5 mL) was added and the solution was
evaporated to dryness under reduced pressure. The
residual oil was crystallized in CCl4, a€ording com-
pound 21c as colorless crystals (54 mg, 90%): mp
174 ꢂC. H NMR (300 MHz, CDCl3) d 1.35 (t, J=7.1,
1
3H), 1.44 (t, J=7.1, 3H), 4.28 (q, J=7.1, 2H), 4.48 (q,
J=7.1, 2H), 6.57 (d, J=16.2, 1H), 7.52 (t, J=7.7, 2H),
7.63 (t, J=7.7, 1H), 7.75 (d, J=16.2, 1H), 7.90 (d,
J=7.7, 2H), 7.99 (s, 1H), 8.57 (s, 1H), 9.40 (s, 1H); 13C
NMR (62.5 MHz, DMSO-d6) d 14.4, 14.5, 35.3, 60.9,
62.2, 117.8, 118.2, 120.6, 127.2, 130.0, 131.2, 133.5,
134.3, 135.1, 135.9, 137.1, 142.5, 165.3, 166.5; IR (KBr)
1712, 1641 cm 1; CIMS m/z 429 (MH)+. Anal. (C21H20
N2O6S) C, H. N.
General procedure for the reduction of the double bonds
of compounds 22a±c. A mixture of the unsaturated deri-
vatives 22a, 22b, or 22c (40, 50 and 20 mg, respectively)
and an equivalent mass of 10% palladium on carbon in
ethanol (5 mL) (22a, 22b) or ethanol (8 mL) and dichloro-
methane (6 mL) (22c) was hydrogenated at atmospheric
pressure for 2 h. The reaction mixture was ®ltered on
Celite1, the ®ltrate was evaporated to dryness under
reduced pressure and the residue was puri®ed by pre-
parative TLC on silica gel (CH2Cl2/EtOH, 9/1). The
following compounds were prepared in this manner:
General procedure for the deprotection of 1-benzene-
sulfonylpyrrolo[2,3-c]pyridine derivatives 21a±c. A solu-
tion of 21a, 21b, or 21c (50±60 mg) in absolute ethanol
(5 mL) was treated at rt with a solution of sodium eth-
oxide in ethanol (1 mL of a 0.4 M solution). The reac-
tion mixture was stirred for 2 h, acetic acid (0.5 mL) was
added and the solution was evaporated to dryness under
reduced pressure. The residue was puri®ed by column
chromatography on silica gel using CH2Cl2/EtOH (95/5)
as eluent. The following compounds were prepared in
this manner:
Ethyl 3-(2-phenylethyl)-1H-pyrrolo[2,3-c]pyridine-5-car-
boxylate (23a). Compound 23a was obtained as a white
powder (24 mg, 58%) from compound 22a: mp 179 ꢂC. 1H
NMR (250 MHz, DMSO-d6) d 1.35 (t, J=7.1, 3H), 3.08
(t, J=7.2, 2H), 3.17 (t, J=7.2, 2H), 4.45 (q, J=7.1, 2H),
7.35 (s, 5H), 7.60 (s, 1H), 8.38 (s, 1H), 8.87 (s, 1H), 12.82
(br s, 1H, exchangeable with D2O); 13C NMR (62.5 MHz,
CD3OD) d 14.4, 27.8, 38.0, 62.0, 118.1, 127.0, 129.3, 129.6,
129.8, 134.7, 165.5; IR (KBr) 3182, 1708 cm 1; CIMS m/z
295 (MH)+. Anal. (C18H18N2O2. 0.3 H2O) C, H, N.
(E,Z)-Ethyl 3-(2-phenylethylene)-1H-pyrrolo[2,3-c]pyr-
idine-5-carboxylate (22a). Compound 22a was obtained
as a tan powder (33 mg, 81%) and as a mixture of E
and Z isomers (3/1) from compound 21a (60 mg,
ꢂ
1
0.14 mmol): mp 247 C (decomp). H NMR (250 MHz,
DMSO-d6) d 1.39 (t, J=7.1, 0.75H), 1.48 (t, J=7.1,
2.25H), 4.35 (q, J=7.1, 0.5 H), 4.50 (q, J=7.1, 1.5H),
6.75 (d, J=12.1, 0.25H), 6.93 (d, J=12.1, 0.25H), 7.35
(d, J=16.6, 0.75H), 7.40 (m, 2H), 7.49 (t, J=7.7, 1H),
7.63 (d, J=16.6, 0.75H), 7.73 (d, J=7.7, 2H), 7.95 (s,
0.25H), 8.18 (s, 0.75H), 8.82 (s, 1H), 8.88 (s, 0.25H),
8.95 (s, 0.75H), 12.10 (br s, 0.25H, exchangeable with
D2O), 12.30 (br s, 0.75H, exchangeable with D2O); 13C
NMR (50 MHz, DMSO-d6) d 14.1, 60.3, 116.8, 120.3,
125.7, 125.8, 126.6, 128.4, 129.5, 134.4, 134.8, 137.3,
137.6, 165.0; IR (KBr) 3428, 1721 cm 1; CIMS m/z 293
(MH)+. Anal. (C18H16N2O2.C2H5OH) C, H, N.
Ethyl 3-(3-phenylpropyl)-1H-pyrrolo[2,3-c]pyridine-5-car-
boxylate (23b). Compound 23b was obtained as a light
brown powder (30 mg, 42%) from compound 22b: mp
159 ꢂC. 1H NMR (250 MHz, DMSO-d6) d 1.45 (t,
J=7.1, 3H), 2.07 (t, J=7.1, 2H), 2.75 (t, J=7.1, 2H),
2.84 (t, J=7.1, 2H), 4.42 (q, J=7.1, 2H), 7.36 (m, 5H),
7.65 (s, 1H), 8.39 (s, 1H), 8.88 (s, 1H), 12.02 (br s, 1H,
exchangeable with D2O); 13C NMR (62.5 MHz, DMSO-
d6) d 14.2, 23.6, 31.7, 34.8, 60.3, 115.8, 116.2, 125.6,
127.8, 128.2, 130.9, 134.0, 134.5, 135.9, 141.9, 165.9; IR
(KBr) 2930, 1716 cm 1; CIMS m/z 309 (MH)+. Anal.
(C19H20N2O2.0.5 H2O) C, H, N.
Ethyl 3-[2(ethoxycarbonyl)ethyl]-1H-pyrrolo[2,3-c]pyr-
idine-5-carboxylate (23c). Compound 23c was obtained
as a pale-yellow solid (20 mg, 100%) from compound
(E,Z)-Ethyl 3-(3-phenyl-1,2-propenyl)-1H-pyrrolo[2,3-c]-
pyridine-5-carboxylate (22b). Compound 22b was
obtained as a white powder (28 mg, 82%) and as a
mixture of E and Z isomers (4/1) from compound 21b
(50 mg, 0.11 mmol): mp 162 ꢂC (decomp). 1H NMR
(250 MHz, DMSO-d6) d 1.46 (t, J=7.1, 3H), 3.68 (d,
J=7.0, 0.4H), 3.80 (d, J=7.0, 1.6H), 4.46 (q, J=7.1,
2H), 5.94 (m, 1H), 6.90 (d, J=11.3, 0.2H), 7.40 (m,
5.8H), 7.90 (s, 1H), 8.55 (s, 0.8H), 8.60 (s, 0.2H), 8.90 (s,
0.2H), 8.96 (s, 0.8H), 12.27 (br s, 1H, exchangeable with
D2O); 13C NMR (50 MHz, DMSO-d6) d14.7, 35.9, 61.6,
117.7, 119.7, 126.3, 128.5, 128.8, 130.0, 134.0, 167.9; IR
(KBr) 3026, 1724 cm 1; CIMS m/z 307 (MH)+. Anal.
(C19H18N2O2.0.6H2O) C, H, N.
ꢂ
1
22c: mp 190 C. H NMR (250 MHz, CDCl3) d 1.25 (t,
J=7.1, 3H), 1.50 (t, J=7.1, 3H), 2.82 (t, J=7.2, 2H),
3.20 (t, J=7.2, 2H), 4.15 (q, J=7.1, 2H), 4.48 (q,
J=7.1, 2H), 8.03 (s, 1H), 8.67 (s, 1H), 9.04 (s, 1H),
12.77 (br s, 1H, exchangeable with D2O); 13C NMR
(62.5 MHz, DMSO-d6) d 13.9, 14.1, 19.2, 34.4, 59.8,
61.8, 116.9, 117.2, 130.9, 132.5, 133.0, 134.2, 141.9,
165.9, 172.0; IR (KBr) 2861, 1727, 1716 cm 1; HRCIMS
calcd for C15H19N2O4 m/z 291.1344, found 291.1340.
Methyl 2-[1-(benzenesulfonylpyrrol-2-yl)methyl]amino-
3,3-diethoxybutyrate (26). To a solution of aldehyde 24

X. Doisy et al. / Bioorg. Med. Chem. 7 (1999) 921±932
931
(176 mg, 0.75 mmol) in anhydrous CH2Cl2 (30 mL)
was added freshly activated 4 A molecular sieves (2 g)
and methyl 2-amino-3,3-diethoxybutyrate26 (384 mg,
1.87 mmol). The reaction mixture was stirred for 36 h at
rt and then ®ltered on Celite1. The ®lter pad was rinsed
with 5% methanolic CH2Cl2 (60 mL) and the ®ltrate
was washed with water (3Â20 mL), dried (Na2SO4) and
mixture was treated as before and the crude product
was puri®ed by preparative TLC (ethyl acetate/methanol
9/1), a€ording pure 28 as a white powder (17 mg, 65%):
mp 202±203 ꢂC. H NMR (200 MHz, CDCl3) d 2.87 (s,
1
3H), 4.00 (s, 3H), 6.77 (d, J=3.1, 1H), 7.46 (d, J=3.1,
1H), 8.79 (s, 1H), 9.22 (br s, 1H, exchangeable with
D2O); IR (KBr) 3416, 1720 cm 1; HRCIMS calcd for
C10H11N2O2 m/z 191.0791, found 191.0828.
1
evaporated leaving the crude imine (210 mg, 66%). H
NMR (300 MHz, CDCl3) d 1.15 (2t, J=7.0, 6H), 1.51
(s, 3H), 3.41±3.65 (m, 4H), 3.75 (s, 3H), 4.35 (s, 1H),
6.31 (t, J=3.5, 1H), 7.40±7.56 (m, 6H), 7.81 (d, J=7.1,
1H), 8.63 (s, 1H); IR (®lm) 1743, 1675 cm 1; CIMS m/z
423 (MH)+.
Methyl 1-benzenesulfonyl-4-bromomethylpyrrolo[2,3-c]-
pyridine-5-carboxylate (29). A solution of compound 27
(326 mg, 0.98 mmol) in CCl4 (15 mL) containing N-bro-
mosuccinimide (210 mg, 1.18 mmol) and AIBN (4 mg,
0.02 mmol) was re¯uxed for 4 h. The reaction mixture
was ®ltered, the ®ltrate was washed with water
(3Â10 mL) and the organic phase was dried (Na2SO4).
Evaporation of the solvent under reduced pressure left a
residue which was puri®ed by column chromatography
on silica gel (heptane/ethyl acetate, 6/4). Compound 29
was obtained as a white powder (250 mg, 61%): mp
104±105 ꢂC. 1H NMR (300 MHz, CDCl3) d 4.06 (s, 3H),
5.19 (s, 2H), 6.98 (d, J=3.2, 1H), 7.54 (t, J=7.9, 2H),
7.66 (d, J=7.2, 1H), 7.87 (d, J=3.2, 1H), 7.97 (d, J=
7.9, 2H), 9.35 (s, 1H); IR (®lm) 1731 cm 1; CIMS m/z
411 (MH)+ with 81Br, 409 (MH)+ with 79Br. Anal.
(C16H13BrN2O4S) C, H, N.
This crude material, more than 95% pure and contain-
1
ing no trace of starting material (as evaluated by H
NMR) was used without further puri®cation in the fol-
lowing step. Thus, a solution of the imine (160 mg,
0.38 mmol) in ethanol (10 mL) was treated with sodium
cyanoborohydride (25 mg, 0.40 mmol). The solution was
stirred for 50 min at rt, the pH being maintained at 4 by
the intermittent addition of alcoholic HCl. The reaction
mixture was diluted with water (10 mL), made basic by
the addition of saturated aqueous NaHCO3 and extrac-
ted with CH2Cl2 (4Â20 mL). The combined organic
extracts were washed with water (15 mL), dried
(Na2SO4) and evaporated under reduced pressure,
1
a€ording amine 26 as a brown oil (157 mg, 98%). H
Methyl 4-methoxymethyl-1H-pyrrolo[2,3-c]pyridine-5-
carboxylate (30). Following the same procedure as for
the preparation of 28, a solution of compound 29
(140 mg, 0.34 mmol) in methanol (7 mL) was treated
with a solution of sodium methoxide in methanol
(5.01 mL of a 0.17 M solution; 0.85 mmol). The reaction
mixture was stirred at rt for 2 h and then worked up in
the usual manner. The crude product was puri®ed by
column chromatography on silica gel (ethyl acetate/
methanol, 8/2), a€ording compound 30 as a white solid
(60 mg, 80%): mp 134±135 ꢂC. 1H NMR (300 MHz,
CDCl3) d 3.46 (s, 3H), 4.01 (s, 3H), 5.22 (s, 2H), 6.95 (d,
J=3.1, 1H), 7.54 (d, J=3.1, 1H), 8.88 (s, 1H), 9.46 (br
s, 1H, exchangeable with D2O); IR (®lm) 3412, 1731
cm 1; HRCIMS calcd for C11H13N2O3 m/z 221.0953,
found 221.0918.
NMR (300 MHz, CDCl3) d 1.09 (2t, J=7.1, 6H), 1.39
(s, 3H), 1.75 (br s, 1H, exchangeable with D2O), 3.45±
3.60 (m, 3H), 3.65 (s, 3H), 3.67±3.80 (m, 3H), 4.65 (s,
1H), 6.42 (m, 1H), 7.15 (m, 1H), 7.45±7.70 (m, 3H),
7.82±8.01 (m, 3H); IR (®lm) 3156, 1743 cm 1; CIMS
m/z 425 (MH)+.
Methyl 1-benzenesulfonyl-4-methylpyrrolo[2,3-c]-pyridine-
5-carboxylate (27). To a vigorously stirring solution of
amine 26 (147 mg, 0.35 mmol) in anhydrous benzene
(6 mL) was added at rt under argon titanium (IV) chlo-
ride (170 mL, 1.56 mmol). The reaction mixture was
re¯uxed for 2 h, cooled to rt and ethyl acetate (10 mL)
followed by saturated aqueous NaHCO3 (10 mL) were
added. The organic phase was separated and the aqu-
eous phase was extracted with ethyl acetate (3Â10 mL).
The organic extracts were combined, washed with water
(10 mL), dried (Na2SO4) and the solvents were removed
under reduced pressure. The solid residue was puri®ed
by column chromatography on silica gel (heptane/ethyl
acetate, 6/4) a€ording compound 27 as a pale-yellow
Molecular modeling. Molecular modeling was per-
formed using Macromodel (version 5.5) on a Silicon
Graphics workstation. For each molecule (4d, 8, and 9),
5000 conformations were sampled using the Monte
Carlo procedure and minimized by use of MM3 force
®elds. Only the ®ve lowest energy conformers which
were generated most frequently and whose ÁE did not
exceed 4 kcal/mol were retained. Of these, the ones
whose side-chain alignments showed the best super-
position are shown in Figure 1.
powder (78 mg, 68%): mp 130 ꢂC. H NMR (200 MHz,
1
CDCl3) d 2.74 (s, 3H), 3.98 (s, 3H), 6.84 (d, J=3.7, 1H),
7.43±7.65 (m, 3H), 7.75 (d, J=3.7, 1H,), 7.90 (d, J=7.2,
2H), 9.22 (s, 1H); 13C NMR (62.5 MHz, CDCl3) d 15.8,
52.5, 107.5, 126.7, 127.0, 128.1, 129.4, 129.7, 132.8,
134.6, 166.7; IR (®lm) 1724 cm 1; CIMS m/z 331
(MH)+. Anal. (C16H14N2O4S) C, H, N.
Receptor binding studies. Inhibition of [3H]¯umazenil
binding to native benzodiazepine (o) receptor subtypes
in vitro was studied as described by Schoemaker et al.29
Brie¯y, the cerebellum, spinal cord or hippocampus of
male Sprague-Dawley rats (180±300 g, OFA, I€a Credo,
St. Germain sur l'arbresle, France) were homogenized
in 50 mM Tris±HCl bu€er (pH 7.4) containing 120 mM
NaCl and 5 mM KCl. The binding of [3H]¯umazenil
Methyl 4-methyl-1H-pyrrolo[2,3-c]pyridine-5-carboxy-
late (28). Using the same general procedure as for the
preparation of compounds 22a±c, compound 27 (46 mg,
0.14 mmol) in methanol (5 mL) was treated with a
solution of sodium methoxide in methanol (1.22 mL of
a 0.17 M solution; 0.2 mmol). After 2 h, the reaction

932
X. Doisy et al. / Bioorg. Med. Chem. 7 (1999) 921±932
(1 nM; speci®c activity 70±87 Ci/mmol, NEN Life Sci-
ence Products) to the o1 receptor was studied in mem-
branes from the rat cerebellum, a region enriched in this
receptor subtype,30 using a 45 min incubation at 0±4 ꢂC
and 1 mM diazepam to de®ne nonspeci®c binding.
[3H]Flumazenil binding to the o2 receptor was studied
using membranes from the rat spinal cord, where a
majority of the expressed o receptors appears to be of
the o2 subtype,31 under otherwise identical conditions.
The native o5 receptor was studied using [3H]¯umazenil
binding to membranes from the rat hippocampus in the
presence of 5 mM zolpidem in order to mask the o1 and
o2 receptor subtypes,17c under otherwise identical con-
ditions except for the use of 1 mM ¯unitrazepam to
de®ne nonspeci®c binding. Following incubation,
membranes were recovered by vacuum ®ltration over
Whatman GF/B ®lters, washed and the amount of
radioactivity retained on the ®lter was quanti®ed by
liquid scintillation spectrometry. Data are presented in
Table 1 as the compound concentration required to
inhibit 50% of speci®c radioligand binding (IC50) and
were obtained from a single experiment performed in
duplicate.
Pharmacol. 1988, 34, 124. (d) Maziere, M.; Hantraye, P.;
Kajima, M.; Dodd, R. H.; Guibert, B.; Prenant, C.; Sastre, J.;
Crouzel, M.; Comar, D.; Naquet, R. Life Sci. 1985, 36, 1609.
13. Nayeem, N.; Green, T. P.; Martin, I. L.; Barnard, E. A. J.
Neurochem. 1994, 62, 815.
14. For reviews, see: (a) Luddens, H.; Wisden, W. Trends
Pharm. Sci. 1991, 12, 49. (b) Sieghart, W. Pharmacol. Rev.
1995, 47, 181.
15. Pritchett, D. B.; Sontheimer, H.; Shivers, B. D.; Ymer, S.;
Kattenmann, H.; Scho®eld, P. R.; Seeburg, P. H. Nature 1989,
338, 582.
16. (a) Von Blankenfeld, G.; Ymer, S.; Pritchett, D. B.; Son-
theimer, H.; Ewert, M.; Seeburg, P. H.; Kettenmann, H. Neu-
rosci. Lett. 1990, 115, 269. (b) Kleingoor, C.; Ewert, M.; Von
Blankenfeld, G.; Seeburg, P. H.; Kettenmann, H. Neurosci.
Lett. 1991, 130, 169. (c) Puia, G.; Vicini, S.; Seeburg, P. H.;
Costa, E. Mol. Pharmacol. 1991, 39, 691. (d) Wa€ord, K. A.;
Whiting, P. J.; Kemp, J. A. Mol. Pharmacol. 1993, 43, 240.
17. (a) Maquire, P. A.; Davies, M. F.; Villar, H. O.; Loew, G.
Eur. J. Pharmacol. 1992, 214, 85. (b) Pritchett, D. B.; Luddens,
H.; Seeburg, P. H. Science 1989, 245, 1389. (c) Tan, S.;
Schoemaker, H. Can. J. Physiol. Pharmacol. 1994, 72, 448.
18. (a) Benavides, J.; Peny, B.; Scatton, B. Eur. J. Neu-
ropsychopharmacol. 1991, 3, 398. (b) Dennis, T.; Dubois, A.;
Benavides, J.; Scatton, B. J. Pharmacol. Exp. Ther. 1988, 247,
309. (c) Benavides, J.; Peny, B.; Durand, A.; Arbilla, S.; Scat-
ton, B. J. Pharmacol. Exp. Ther. 1992, 263, 884. (d) Pritchett,
D. B.; Seeburg, P. H. J. Neurochem. 1990, 54, 1802.
19. Liu, R.; Hu, R. J.; Zhang, P.; Skolnick, P.; Cook, J. M. J.
Med. Chem. 1996, 39, 1928.
20. (a) Dodd, R. H.; Ouannes, C.; Prado de Carvalho, L.;
Valin, A.; Venault, P.; Chapouthier, G.; Rossier, J.; Potier, P.
J. Med. Chem. 1985, 28, 824. (b) Dodd, R. H.; Ouannes, C.;
Potier, M.-C.; Prado de Carvalho, L.; Rossier, J.; Potier, P. J.
Med. Chem. 1987, 30, 1248. (c) Dorey, G.; Poissonnet, G.;
Potier, M.-C.; Prado de Carvalho, L.; Venault, P.; Chapou-
thier, G.; Rossier, J.; Potier, P.; Dodd, R. H. J. Med. Chem.
1989, 32, 1799. (d) Dorey, G.; Dubois, L.; Prado de Carvalho,
L.; Potier, P.; Dodd, R. H. J. Med. Chem. 1995, 38, 189.
21. (a) Stephens, D. N.; Kehr, W.; Wachtel, H.; Schmiechen,
R. Pharmacopsychiatry 1985, 18, 167. (b) Hollinshead, S. P.;
Trudell, M. L.; Skolnick, P.; Cook, J. M. J. Med. Chem. 1990,
33, 1062. (c) Petersen, E. N.; Jensen, L. H.; Honore, T.;
Braestrup, C.; Kehr, W.; Stephens, D. N.; Wachtel, H.; Sei-
delman, D.; Schmiechen, R. Psychopharmacologia, 1984, 83,
240.
References and Notes
1. Sternbach, L. H. J. Med. Chem. 1979, 22, 1.
2. For reviews, see: (a) Haefely, W.; Kyburz, E.; Gerecke, M.;
Mohler, H. Advances in Drug Research; Testa, B., Ed.; Aca-
demic: London, 1985; Vol. 14, pp 165±322. (b) Gardner, C. R.;
Tully, W. R.; Hedgecock, C. J. R. Prog. Neurobiol. 1993, 40, 1.
3. (a) Stratten, W. P.; Barnes, C. D. Neuropharmacology 1971,
10, 685. (b) Steiner, F. A.; Felix, D. Nature 1976, 260, 346.
4. (a) Squires, R. F.; Braestrup, C. Nature 1977, 266, 732. (b)
Mohler, H.; Okada, T. Science 1977, 198, 849.
5. (a) Tallman, J. F.; Thomas, J. W.; Gallager, D. W. Nature
1978, 274, 383. (b) Gavish, M.; Snyder, S. H. Life Sci. 1980,
26, 579. (c) Study, R. E.; Barker, J. L. Proc. Natl. Acad. Sci.
USA 1981, 78, 7180.
6. Hunkeler, W.; Mohler, H.; Pieri, L.; Polc, P.; Bonetti, E. P.;
Cumin, R.; Scha€ner, R.; Haefely, W. Nature 1981, 290, 514.
7. Klepner, C. A.; Lippa, A. S.; Benson, D. I.; Sano, M. C.;
Beer, B. Pharmacol. Biochem. Behav. 1979, 11, 457.
8. Yokoyama, N.; Ritter, B.; Neubert, A. D. J. Med. Chem.
1982, 25, 337.
9. (a) Braestrup, C.; Schmiechen, R.; Neef, G.; Nielsen, M.;
Petersen, E. N. Science 1982, 216, 1241. (b) Braestrup, C.;
Nielsen, M.; Olsen, C. E. Proc. Natl. Acad. Sci. USA 1980, 77,
2288. (c) Oakley, N. R.; Jones, B. J. Neuropharmacology 1982,
21, 587. (d) Prado de Carvalho, L.; Grecksch, G.; Cavalheiro,
E. A.; Dodd, R. H.; Chapouthier, G.; Rossier, J. Eur. J. Phar-
macol. 1984, 103, 287. (e) Prado de Carvalho, L.; Grecksch, G.;
Chapouthier, G.; Rossier, J. Nature 1983, 301, 64.
10. (a) Venault, P.; Chapouthier, G.; Prado de Carvalho, L.;
Simiand, J.; Morre, M.; Dodd, R. H.; Rossier, J. Nature, 1986,
321, 864. (b) Ra€alli-Sebille, M. J.; Chapouthier, G.; Venault,
P.; Dodd, R. H. Pharmacol. Biochem. Behav. 1990, 35, 281. (c)
For a review of the memory-enhancing activities of b-carbo-
lines, see: Dodd, R. H. Eur. Bull. Cog. Psychol. 1992, 12, 484.
11. (a) Nielsen, M.; Braestrup, C. Nature 1980, 286, 606. (b)
Olsen, R. W.; McCabe, R. T.; Wamsley, J. K. J. Chem. Neu-
roanat. 1990, 3, 59. (c) Sieghart, W.; Schlerka, W. Eur. J.
Pharmacol. 1991, 197, 103.
22. Zhang, W.; Koehler, K. F.; Zhang, P.; Cook, J. M. Drug
Design Discovery 1995, 12, 193.
23. (a) Rousseau, J.-F.; Dodd, R. H. J. Org. Chem. 1998, 63,
2731. (b) Dekhane, M.; Potier, P.; Dodd, R. H. Tetrahedron
1993, 49, 8139. (c) Dodd, R. H.; Doisy, X.; Potier, P.; Potier,
M.-C.; Rossier, J. Heterocycles 1989, 28, 1101.
24. Coker, J. N.; Mathre, O. B.; Todd, W. H. J. Org. Chem.
1963, 28, 589.
25. (a) Sonnet, P. E. J. Med. Chem. 1972, 15, 97. (b) Loader,
C. E.; Anderson, H. J. Tetrahedron 1969, 25, 3879.
26. Dekhane, M.; Dodd, R. H. Tetrahedron 1994, 50, 6299.
27. Olivier, A.; Sevrin, M.; Durant, F.; George, P. Bioorg.
Med. Chem. Lett. 1997, 7, 2277.
28. Cox, E. D.; Diaz-Arauzo, H.; Huang, Q.; Reddy, M. S.;
Ma, C.; Harris, B.; McKernan, R.; Skolnick, P.; Cook, J. M.
J. Med. Chem. 1998, 41, 2537.
29. Schoemaker, H.; Claustre, Y.; Fage, D.; Rouquier, L.;
Chergui, K.; Curet, O.; Oblin, A.; Gonon, F.; Carter, C.;
Benavides, J.; Scatton, B. J. Pharmacol. Exp. Ther. 1997, 280,
83.
30. Braestrup, C.; Nielsen, M. Trends Neurosci. 1980, 3, 301.
31. Ruano, D.; Vizuete, M.; Cano, J.; Machado, A.; Vitorica,
J. J. Neurochem. 1992, 58, 485.
12. (a) Sieghart, W. Trends Pharmacol. Sci. 1989, 10, 407. (b)
Zezula, J.; Karall, S.; Dodd, R. H.; Sieghart, W. Eur. J.
Pharmacol. 1995, 281, 93. (c) Potier, M.-C.; Prado de Car-
valho, L.; Dodd, R. H.; Besselievre, R.; Rossier, J. Mol.