2,3-Disubstituted quinuclidines as a novel class of dopamine transporter inhibitors

Article abstract of DOI:10.1016/S0968-0896(02)00450-9

There is considerable interest in developing dopamine transporter (DAT) inhibitors as potential therapies for the treatment of cocaine abuse. We report herein our pharmacophore-based discovery and molecular modeling-assisted rational design of 2,3-disubst

Full text of DOI:10.1016/S0968-0896(02)00450-9

Bioorganic & Medicinal Chemistry 11 (2003) 1123–1136
2,3-Disubstituted Quinuclidines as a Novel Class of
Dopamine Transporter Inhibitors
Sukumar Sakamuri,a,{ Istvan J. Enyedy,b,c,y Wahiduz A. Zaman,d Srihari R. Tella,e
Alan P. Kozikowski,a Judith L. Flippen-Anderson,f Tivadar Farkas,g
Kenneth M. Johnsond and Shaomeng Wangb,c,
*
aDepartment of Neurology, Georgetown University Medical Center, 3900 Reservoir Rd., Washington, DC 20007, USA
bDepartment of Oncology, Georgetown University Medical Center, 3900 Reservoir Rd., Washington, DC 20007, USA
cDepartment of Neuroscience, Georgetown University Medical Center, 3900 Reservoir Rd., Washington, DC 20007, USA
dDepartment of Pharmacology and Toxicology, University of Texas Medical Branch, Galveston, TX 77555-1031, USA
eDepartment of Pharmacology, Georgetown University Medical Center, 3900 Reservoir Rd., Washington, DC 20007, USA
fLaboratory for the Structure of Matter, Naval Research Laboratory, 4555 Overlook Avenue, Washington, DC 20375-5341, USA
gPhenomenex, Inc., 2320 W. 205th Street, Torrance, CA 90501, USA
Received 21 June 2002; accepted 18 August 2002
Abstract—There is considerable interest in developing dopamine transporter (DAT) inhibitors as potential therapies for the treat-
ment of cocaine abuse. We report herein our pharmacophore-based discovery and molecular modeling-assisted rational design of
2,3-disubstituted quinuclidines as potent DAT inhibitors with a novel chemical scaffold. Through 3-D-database pharmacophore
searching, compound 12 was identified as a very weak DAT inhibitor with Ki values of 7.3 and 8.9 mM in [3H]mazindol binding and
in inhibition of dopamine reuptake, respectively. Molecular modeling-assisted rational design and chemical modifications led to
identification of potent analogues (ꢀ)-29 and 34 with Ki values of 14 and 32 nM for both compounds in binding affinity and inhi-
bition of dopamine reuptake, respectively. Behavioral pharmacological evaluations in rodents showed that 34 has a profile very
different from cocaine. While 34 is substantially more potent than cocaine as a DAT inhibitor, it is approximately four times less
potent than cocaine in mimicking the discriminative stimulus properties of cocaine in rat. On the other hand, 34 (3–30 mg/kg) lacks
either the locomotor stimulant or stereotypic properties of cocaine in mice. Importantly, 34 blocks locomotor stimulant activity
induced by 20 mg/kg cocaine in mice, with an estimated ED50 of 19 mg/kg. Taken together, our data suggest that 34 represents a
class of potent DAT inhibitors with a novel chemical scaffold and a behavioral pharmacological profile different from that of
cocaine in rodents. Thus, 34 may serve as a novel lead compound in the ultimate development of therapeutic entities for cocaine
abuse and/or addiction.
# 2002 Elsevier Science Ltd. All rights reserved.
Introduction
Although cocaine (1) potently inhibits the reuptake of
both norepinephrine (NE) and serotonin (SERT), many
lines of evidence indicate that its reinforcing property
stems from its ability to inhibit the reuptake of dop-
amine (DA) into dopaminergic neurons.2ꢀ6 Cocaine’s
ability toinhibit DA reuptake and tocnosequently
increase dopaminergic transmission in the reward
mediating mesolimbic system of the brain is the essence
of the dopamine hypothesis of reinforcement advanced
by Wise6 and later elaborated for cocaine by Kuhar.4
Cocaine (1) abuse is one of the greatest concerns of the
American public today and therefore has become a
focus of medical, social, and political leaders. Despite
intensive research efforts, no effective pharmacotherapy
for the treatment of cocaine abuse is currently available.1
*Corresponding author at CCGC 3316, Departments of Internal
Medicine and Medicinal Chemistry, University of Michigan, Ann
Arbor, MI 48109-0934, USA. Tel.: +1-734-615-0362; fax: +1-734-
647-9647; e-mail:
E-mail address:shaomeng@umich.edu ().
yPresent address: Bayer Pharmaceutical Division, 400 Morgan Lane,
West Haven, CT 06516, USA.
{Present address: Morphochem Inc., 11 Deer Park Drive, Suite 116,
Monmouth Junction, NJ 08852, USA.
There is considerable interest in the development of
DAT inhibitors as potential therapies for the treatment
of cocaine abuse.1,7 TwoDAT inhibitors, RTI 113 ( 2)8
and GBR 12909 (3),9 were shown to effectively decrease
cocaine self-administration behavior in monkeys, and
0968-0896/03/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(02)00450-9

1124
S. Sakamuri et al. / Bioorg. Med. Chem. 11 (2003) 1123–1136
One major advantage of 3-D-pharmacophore database
searching is its ability toidentify many structurally
diverse ‘lead’ compounds, thus providing great chemical
diversity to the design of novel DAT inhibitors.10
However, many such compounds only have moderate
or weak activity, typically with a Ki value in the micro-
molar range. Inhibitors with weak activity are not useful
pharmacological agents or potential therapeutic agents
unless their activity can be significantly improved. A big
challenge we face is how to effectively optimize com-
pounds with weak activity. Herein, we wish to report
the discovery, molecular modeling-assisted design and
rapid optimization of 2,3-disubstituted quinuclidines as
a novel class of dopamine transporter inhibitors. Beha-
vioral pharmacological evaluation of analogue 34 sug-
gests that this compound is distinct from cocaine and
other classes of known DAT inhibitors and that it may
be useful in the ultimate development of therapeutic
entities for cocaine abuse and/or addiction.
are currently being considered for clinical development
for the treatment of cocaine abuse by the National
Institute on Drug Abuse. In view of the urgency and
complexity in the development of an effective cocaine
therapy, we believe that discovery of DAT inhibitors
that have truly novel chemical scaffolds would have
considerable value. Novel DAT inhibitors may have
pharmacological and behavioral profiles different from
cocaine and other known DAT inhibitors. DAT inhibi-
tors with truly novel chemical scaffolds will also provide
new insights intothe 3-D structure of DAT and its
interactions with cocaine and dopamine.
Results and Discussion
Identification of novel DAT inhibitors through 3-D-
database pharmacophore searching
Based upon the extensive structure–activity relation-
ships (SARs) of cocaine (1) and its analogues, and
molecular modeling studies, a simple pharmacophore
model was proposed.10 This pharmacophore model
includes three chemical groups that are important for
binding tothe DAT and their three-dimensional (3-D)
geometrical parameters (Fig. 1).10 A 3-D-database
pharmacophore searching was then performed to iden-
tify potential inhibitors of dopamine transporter
(DAT), whose 3-D structures meet the requirements
specified in the model. To date, testing of compounds
identified from this approach has led to the discovery of
several classes of novel DAT inhibitors, some of which
display fairly potent activity in binding and uptake
assays.10ꢀ16 In addition to those DAT inhibitors repor-
ted in our studies,10ꢀ15 several other compounds showed
moderate activity. The chemical structures and the Ki
values of these compounds as inhibitors of DA uptake
are provided in Chart 1. Because of their moderate
potency, the DAT inhibitors shown in Chart 1 prob-
ably would not be useful therapeutic agents unless
their potency can be significantly improved. Although
each of these compounds could be considered as a
potential lead, we first carried out optimization of
compound 12.
The approach we have been employing in our labora-
tory for the discovery of potent DAT inhibitors with
novel chemical scaffolds is to first identify potential
DAT inhibitors through 3-D-pharmacophore searching
of large chemical databases, followed by confirmation in
DAT binding and DA uptake assays. Todate, using this
approach, we have discovered several classes of novel
DAT inhibitors (4, 5, and 6).10ꢀ15 Interestingly, some of
these novel DAT inhibitors have a significant functional
antagonism against cocaine in vitro.10ꢀ12 It was found
that a potent DAT inhibitor, 3,4-dichlorophenyl 4-(3,4-
dichlorophenyl)-4-hydroxy-1-methyl-3-piperidyl ketone
(4), mimics some of cocaine’s effects in tests of loco-
motor activity and drug discrimination in rodents.10
Further behavioral pharmacological evaluations
showed that 4 attenuates locomotor activity induced by
20 mg/kg cocaine in mice and appears to be a mild sti-
mulant with a slow onset and long duration of action.16
Collectively, our results support our approach that
potent DAT inhibitors with novel chemical scaffolds
may have the therapeutic potential for the treatment of
cocaine abuse, functioning either as cocaine antagonists
or ‘partial agonists’.
Molecular modeling-assisted design
Compound 12 can be classified as a 2,3-disubstituted
quinuclidine. This compound has a moderate potency
with Ki values of 7.3 and 8.9 mM in inhibition of [3H]-
mazindol binding and DA reuptake, respectively (Table
1). Despite its very weak activity, this compound has a
structural scaffold different from other classes of known
DAT inhibitors. Thus, we proposed that 12 may repre-
sent a ‘lead’ compound for a novel class of DAT in-
hibitors if its potency could be significantly improved.

S. Sakamuri et al. / Bioorg. Med. Chem. 11 (2003) 1123–1136
Table 1. Binding and DA reuptake activity at the DAT site
1125
Compd
Ki (nM)a
Mazindol binding
DA reuptake
R-Cocaine (1)
12
17
18
274ꢁ22
7.3ꢁ0.4 (mM)
n.t.
274ꢁ20
8.9ꢁ0.4 (mM)
>10 (mM)
31 3 (mM)
461ꢁ18
237ꢁ7
n.t.
23
260ꢁ4
210ꢁ25
20ꢁ1
(ꢁ)-28
(ꢁ)-29
(+)-29
(ꢀ)-29
30
31
32
33
34
49ꢁ1
343ꢁ16
14ꢁ2
354ꢁ1
32ꢁ2
155ꢁ21
30ꢁ1
186ꢁ16
57ꢁ4
n.t.
n.t.
119ꢁ10
198ꢁ10
32ꢁ5
Figure 1. The pharmacophore model used in 3-D-database pharma-
cophore searching, which led to the identification of the initial com-
pound 12.
14ꢁ1
n.t., Not tested.
aMeanꢁstandard error or range of 2–3 experiments, each conducted
using six concentrations of drug in triplicate.
Toward this end, we carried out molecular modeling-
assisted design and chemical modifications of 12.
its receptor. Thus, two compounds binding to the same
binding site with similar binding modes often have a
minimal exclusion volume especially if the binding site is
not on the receptor surface. Mutagenesis analysis showed
that the binding site of cocaine at the DAT is not located
on a surface.17,18 Therefore, the two overlaps shown in
Figure 2 may not represent the ‘true’ binding mode of the
compound 12 in comparison to that of cocaine.
Compound 12 has two basic nitrogen atoms, one car-
bonyl group and two equivalent phenyl groups. Thus,
two different overlaps are possible between 12 and
cocaine using the three pharmacophore elements
defined in Figure 1 (i.e., a tertiary nitrogen, a carbonyl
group, and a phenyl ring as the reference points). It was
found that lead compound 12 has a fairly good overlap
with cocaine with respect to the three crucial pharma-
cophore elements. The lowest root-mean-square devia-
tion (RMSD) values in these two different overlaps (Fig.
2A and B) between low energy conformations of 12 and
the X-ray structure of cocaine (1) are 1.12 and 0.95 A,
respectively, using the three reference points (a nitrogen
atom, a carbonyl group and an aromatic ring center).
Although the lead compound 12 and cocaine have fairly
good overlap with respect to the pharmacophore elements
defined in Figure 1, close examination of the two over-
laps between 12 and cocaine (Fig. 2) showed that there is
a large exclusion volume between these two molecules.
While 12 and cocaine have an overlapping volume of
159 and 174 A3 with the superposition shown in Fig. 2A
and B, respectively, they have an exclusion volume of
212 and 198 A3, respectively. Van der Waals (steric)
interaction is perhaps the single most important factor
in determining the binding mode of a drug molecule to
Therefore, we investigated whether there was an alter-
native overlap between the compound 12 and cocaine. It
was shown that replacement of the ester group in
cocaine at position 2 with small alkyl groups resulted in
very potent DAT inhibitors. For example, compound
13 with a butyl group at the 2b position and a
p-Cl-phenyl group at the 3b position is a highly potent
DAT inhibitor with a low nanomolar potency in bind-
ing affinity and inhibition of DA reuptake.19 This sug-
gested that the carbonyl group defined in the
pharmacophore model in Figure 1 can be modified to
Chart 1. Novel DAT inhibitors discovered through 3-D-database pharmacophore searching that have not been reported previously.

1126
S. Sakamuri et al. / Bioorg. Med. Chem. 11 (2003) 1123–1136
Figure 3. An alternative overlap between the compound 12 (green) and
cocaine (1, yellow) using an augmented pharmacophore model.
the exclusion volume is only slightly better than that
shown in Figure 2, it was found that the bulky
2-hydroxyl-2,2-diphenylacetate group accounts for
much of this exclusion volume. It was shown that in
cocaine, replacement of its benzoate group at the 3b
position with a phenyl group resulted in compound 14
(WIN 35065-2) with a binding affinity 4 times better than
cocaine at the DAT site.20 Thus, the bulky 2-hydroxy-
2,2-diphenylacetate group at position 3 of the quinucli-
dine ring in 12 may be replaced with a simple phenyl
group to improve the overlapping volume and conse-
quently the activity. Since a small ester or a simple alkyl
group at the 2b position of cocaine is desirable for high
affinity at the DAT site,19,20,23 the N,N-dimethylmethyl-
amino group at position 2 of the quinuclidine ring in 12
may be replaced with a simple alkyl group for achieving
potent activity at the DAT site. The two substituents at
positions 2 and 3 of the quinuclidine ring can be in
either trans or cis configurations. Molecular modeling
showed that compounds with cis-configuration have a
better overlap with cocaine (1) and WIN 35065-2 (14).
Figure 2. Two possible overlaps between the compound 12 (green)
and cocaine (1, yellow) using the three pharmacophore elements
defined in Figure 1.
include simple alkyl groups. With this modified
pharmacophore model, it is possible that the small
N,N-dimethylmethylamino group of 12 may mimic the
ester group at the 2b position of cocaine and the
2-hydroxyl-2,2-diphenylacetate group at position 3 may
mimic the benzoate group at the 3b position of cocaine.
The lowest RMSD value obtained between the low
energy conformations of 12 and the X-ray structure of
cocaine is 0.50 A, using the five reference points shown
in Figure 3. As can be seen, a good overlap was found
between these twomolecules ( Fig. 3). The 2-hydroxy-
2,2-diphenylacetate group at position 3 of the quinuclidine
ring locates in the same region as the phenyl ester group at
the 3b position of cocaine, and the N,N-dimethylamino
group at position 2 of the quinuclidine ring overlaps
nicely with the methyl ester group at the 2b position of
cocaine. However, the 2-hydroxyl-2,2-diphenylacetate
group at position 3 of the quinuclidine ring appeared to
be too bulky for achieving optimal potency based upon
the structure–activity relationships (SARs) of cocaine
and its analogues. Molecular volume calculations
showed that with the overlap depicted in Fig. 3, com-
pound 12 and cocaine have an overlapping volume of
179 A3 and an exclusion volume of 194 A3. Although
Based upon these molecular modeling analyses, com-
pound 28 was designed, which has a simple butyl group
at position 2 and a phenyl group at position 3 with a cis-
configuration between them. A fairly good overlap was
found between 28 and cocaine as depicted in Figure 4A
and the lowest RMSD value was 1.07 A using the 5
reference points shown in Figure 4A with the low
energy conformations of 28 and the X-ray structure of
cocaine. But an excellent overlap was found between 28
and WIN 35065-2 (14), an analogue more potent than
cocaine (Fig. 4B). The lowest RMSD value was 0.30 A
between their low energy conformations using the 5
reference points shown in Figure 4B for superposition.

S. Sakamuri et al. / Bioorg. Med. Chem. 11 (2003) 1123–1136
1127
Compound 28 and WIN 35065-2 (14) have an over-
lapping volume of 179 A3 and an exclusion volume of
54 A3, indicating an excellent overlap in terms of their
overall shape. It is of interest to note that although the
locations of the nitrogen atom in 28 and WIN 35065-2
(14) (Fig. 4B) are within 0.1 A, the orientations of the
nitrogen lone pair in these two compounds differ by
approximately 60ꢂ. Taken together, our molecular
modeling results suggested that 28 should have a much
improved potency as compared to 12.
Compound 28 was synthesized (Scheme 1) and eval-
uated as a DAT inhibitor. The Ki values of 28 in [3H]-
mazindol binding and inhibition of DA reuptake are
210 and 237 nM (Table 1), respectively. Thus, 28 is
more than 30 times more potent than 12 both in binding
and in inhibition of DA reuptake and is as potent as
cocaine. These data provide unambiguous confirmation
of our molecular modeling-assisted design strategy. Two
intermediates 17 and 18 were alsotested tobotain
additional information about the SARs of this class of
compounds. Compound 17 did not show any measur-
able activity at 10 mM in inhibition of DA reuptake
(Table 1), suggesting an important role of the phenyl
group and/or a detrimental effect of the ketone group at
position 3. Compound 18 had a Ki value of 31 mM in
inhibition of DA reuptake (Table 1), 131-fold less
potent than 28. Since solvation effect plays a major role
for the binding of DAT inhibitors, the weak binding
affinity of 18 is probably due to its unfavorable solva-
tion energy as compared to 28.37
Chemistry
Synthesis of 28 and other 2,3-disubstituted quinuclidines
in racemic form was accomplished using a synthetic
procedure as shown in Scheme 1.15 Briefly, starting from
Figure 4. Overlaps between the designed analogue 28 (green) and
cocaine (1, yellow) (A), and between 28 (green) and WIN 35065-2 (14,
yellow) (B).
Scheme 1. Synthesis of 2,3-disubtituted quinuclidines.

1128
S. Sakamuri et al. / Bioorg. Med. Chem. 11 (2003) 1123–1136
3-quinuclidinone (15), 2-methylene-3-quinuclidinone
(16) was prepared by using Mannich reaction.21 Reac-
tion of 15 with aq dimethylamine and aq formaldehyde
in ethanol–water mixture at 70 ꢂC gave the Mannich
base, which on deamination under distillation gave
compound 16 in 86% yield. Reaction of 16 with allyl-
formation of 28 as shown in Figure 4. Molecular mod-
eling analysis confirmed that 29 has an excellent overlap
with WIN 35065-2 (14) and 28. Compound 29 was syn-
thesized in racemic form using the same procedure as
for 28 (shown in Scheme 1) and evaluated as a DAT
inhibitor. It was found that 29 has Ki values of 20 and
49 nM in binding affinity and inhibition of DA reup-
take, respectively. Thus, 29 is 11 and 5 times more
potent than 28 in its binding and uptake activities,
respectively.
.
magnesium bromide in the presence of CuI Me2S and
Me3SiCl at ꢀ78 ꢂC furnished the conjugate addition
product 17 in 47% yield along with the 1,2-addition
product in 12% yield (structure not shown). Aryl
Grignard addition was carried out using arylmagnesium
bromide in THF at 0 ꢂC togive compounds
18–22 in
good yield. Dehydration using a 1:1 mixture of EtOH
and 6 N HCl under reflux conditions gave the dehy-
drated compounds 23–27. Reduction of the double
bonds in compounds 23 and 24 was carried out using
standard hydrogenation conditions (Pd/C, H2, EtOH,
60 psi) to provide compounds 28 and 29 in near quan-
titative yield. Selective reduction of terminal double
bond in compounds 23–27 using Wilkinson catalyst
under hydrogen atmosphere in toluene at 70 ꢂC over-
night gave compounds 30–34 in good yield.
Toconfirm the cis-configuration between substituents at
positions 2 and 3 in 29 and our molecular modeling
results, the X-ray structure of 29 was obtained (Fig. 5).
As can be seen, the butyl group at position 2 and the
p-methylphenyl at position 3 indeed have the desired
cis-configuration. Since the binding of cocaine to DAT
is stereospecific, it was thus interesting to investigate the
stereospecificity of compound 29 in binding tothe
DAT. It was found that (ꢀ)-29 has Ki values of 14 and
32 nM, while (+)-29 has Ki values of 343 and 354 nM in
binding affinity and inhibition of DA reuptake, respec-
tively. Hence, (ꢀ)-29 is approximately 2-fold more
potent than (ꢁ)-29 and is 11-fold more potent than its
enantiomer (+)-29.
The enantiomers (+)-29 and (ꢀ)-29 were obtained
using a semi-preparative chiral HPLC column (Chirex
3018), in which chiral stationary phase (CPS) consists of
(S)-proline and (R)-1-a-naphthylethylamine covalently
bound to a g-aminopropyl silanized silica gel, and hex-
ane/CH2Cl2/EtOH–TFA (20:1) in 83:15:2 ratioas the
eluent.22 The optical rotation of (+)-29 was found to be
[a]D=+104ꢂ (c 0.5, acetone) and that of (ꢀ)-29 was
[a]D=ꢀ104ꢂ (c 0.5, acetone).
Structure–activity relationship studies
Tofurther improve the potency of 28, we have carried out
focused structure–activity relationship studies. Previous
studies have shown that an additional substitution to
the phenyl ring such as a p-methyl group may further
improve the potency.23 Thus, compound 29 with an
additional p-methyl group should have an improved
activity if it can adopt the similar low energy con-
To overcome the tedious separation of the enantiomers
using chiral HPLC, we tested the hypothesis that com-
pounds with a double bond at the C2–C3 position, such
as 23, have reasonable activity. Molecular modeling
analysis showed that 23 has an excellent overlap with
28, suggesting that 23 may be a potent DAT inhibitor.
Indeed, 23 was found to have Ki values of 260 and 461
nM in binding affinity and inhibition of DA reuptake,
only slightly less potent than 28. Reduction of the
terminal double bond in 23 gave compound 30. Co m-
pound 30 was found to have Ki values of 155 and 186
nM in binding and inhibition of DA reuptake, respec-
tively. Thus, 30 is as potent as 28. Encouraged by the
good potency of 30, compound 31 with a para-methyl-
phenyl substitution at the C3 position was synthesized.
It was found that compound 31 has Ki values of 30 and
57 nM in binding and inhibition of DA reuptake,
respectively. Thus, 31 is as potent as (ꢁ)-29 and is only
2-fold less potent than its active enantiomer (ꢀ)-29.
We further investigated the effect of the aromatic sub-
stituent. First, compounds 32 and 33 with a meta- o r an
ortho-methyl group in the phenyl ring were synthesized
and tested. While 32 with a methyl group at meta-posi-
tion is 2-fold less potent than 31, 33 with a methyl group
at ortho-position is approximately 4-fold less potent
than 31 in inhibition of DA reuptake. These results
Figure 5. The structure of compound 29 drawn using the experimen-
tally determined X-ray coordinates with arbitrary thermal parameters.

S. Sakamuri et al. / Bioorg. Med. Chem. 11 (2003) 1123–1136
1129
clearly suggest that a methyl substitution at the para-
position is more favorable than that at either the ortho- o r
the meta-position for the DAT activity.
Because of its lack of locomotor stimulant activity, we
decided todetermine whether 34 was capable of blocking
the cocaine-induced locomotor stimulant activity. As can
be seen in Figure 8a, 34 significantly attenuated cocaine-
induced locomotor activity as measured by distance tra-
veled. At 30 mg/kg, 34 blocked the cocaine-induced loco-
motor stimulant activity by 80% and has an estimated
ED50 value of 19 mg/kg. But 34 had nosignificant effect
on cocaine-induced stereotypic movement (Fig. 8b).
In the WIN analogues, it is well known that replacing
hydrogen with a chlorine at para-position yielded more
potent DAT inhibitors than the unsubtituted WIN
compound.20,23 Accordingly, compound 34 with a para-
chloro substituent was synthesized and tested. Com-
pound 34 was found to have Ki values of 14 and 32 nM
in binding and inhibition of DA reuptake, respectively.
Thus, 34 is 2 times more potent than 31 and is as potent
as (ꢀ)-29 in both binding and inhibition of DA reup-
take. Although 34 is not the most potent DAT inhibitor
reported in the literature, this compound is approxi-
mately 16- and 8-fold more potent than cocaine in
binding and inhibition of DA reuptake, respectively
(Table 1), representing a promising lead compound for
further design and structure–activity relationship stud-
ies. Compound 34 is also potent enough for carrying
out behavioral pharmacological evaluations.
Behavioral pharmacological studies
Several observations have demonstrated that potent
DAT inhibitors may have promise as potential medica-
tions for cocaine addiction. First, although many potent
DAT inhibitors indeed behave exactly like cocaine in
behavioral pharmacological models, some DAT inhibi-
tors such as RTI-113 (2) and GBR 12909 (3) as well as
compound 4 are milder stimulants that are slower in
onset and have a longer duration of action than
cocaine.8,9 Second, some potent DAT inhibitors including
several benzotropine analogues24,25 and compound 4 were
found to mimic some of cocaine’s effects in behavioral
models.10 Third, recent studies showed that potent DAT
inhibitors antagonize locomotor stimulant effects induced
by cocaine in mice, while having no significant effect on
locomotor activity when tested alone.26 Because of its
novel chemical scaffold and its good potency, we deci-
ded to scale up the synthesis of compound 34 and eval-
uate it in behavioral pharmacological models in rodents.
In rats trained todiscriminate 10 mg/kg cocaine (ip)
from saline, both cocaine (1.56–10 mg/kg) and 34 (0.3–
30 mg/kg) produced dose-dependent, full substitution
(Fig. 6a). The doses of cocaine and 34 needed topro-
duce 50% generalization (ED50 value) to training dose of
cocaine (10 mg/kg) are 3 and 12 mg/kg, respectively. Thus,
34 is about 4-fold less potent than cocaine, although 34 is
16 and 8 times more potent than cocaine in mazindol
binding and DA reuptake assays, respectively. Neither
cocaine nor 34 altered the response rates (Fig. 6b).
In locomotor activity testing in mice, cocaine (3–30 mg/
kg) produced dose-dependent enhancements in the dis-
tance traveled and stereotypic movements (Fig. 7a and
b); 34 produced no significant increase in either loco-
motor activity or stereotypic behavior in mice in the
dose range (3–30 mg/kg) tested. It is of note that at 56
and 100 mg/kg, 34 produced convulsions in 50% (4 out
of 8 animals) and 100% (8 out of 8 animals) of the ani-
mals tested, respectively.
Figure 6. Discriminative stimulus effects of cocaine (circles) and 34
(squares) in Sprague–Dawley rats. Both cocaine and 34 produced
dose-dependent generalization to cocaine but 34 is less efficacious than
cocaine. Both drugs were administered ip 10 min prior to the testing.
The response requirement for food reinforcement was a fixed-ratio 10
(FR10). A maximum of 20 pellets/session was allowed. (a) Percent
cocaine-lever response. (b) Response rate as percentage of vehicle
control. **P<0.001 as compared to the saline control response.

1130
S. Sakamuri et al. / Bioorg. Med. Chem. 11 (2003) 1123–1136
The generalization to cocaine by 34 in drug discrimi-
nation experiments is consistent with the fact that 34 is a
potent DAT inhibitor. However, the differential corre-
lation between the discriminative stimulus effects and
the inhibition of DAT function for cocaine and 34 sug-
gest that these molecules differ radically from each other
in either pharmacokinetic properties and/or in basic
biologic mechanisms of action. The complete lack of
locomotor stimulant activity of 34 in mice is surprising
since a strong correlation was reported between the
relative dopamine transporter binding affinities of
cocaine analogues versus their locomotor stimulant
potencies.27ꢀ29 As evident from its ability to antagonize
the stimulant effect of cocaine, 34 is accessible tothe
central nervous system in mice. Thus, the lack of loco-
motor stimulant activity of 34 cannot be attributed to
its possible inability to pass through the blood–brain
barrier.
Since cocaine functions as an indirect dopamine recep-
tor agonist, we investigated the possibility that the lack
Figure 7. Locomotor stimulant effects of cocaine (circles) and com-
pound 34 (squares) in Swiss Webster mice. The locomotor activity was
recorded using Truscan activity monitors (Coulbourn Instruments,
Columbus, Ohio). Following 1 h of habituation to test environment,
cocaine, 34 or their vehicles were administered ip. The data were
recorded for the next 2 h. The original 12 blocks (10-min bins) of 2-h
data were converted to 30-min totals. The maximal 30-min total for
each dose was identified and presented as the response. Cocaine pro-
duced significant and dose-dependent locomotor activation. Unlike
cocaine, 34 did not produce locomotor stimulation. (a) Distance tra-
veled. (b) Stereotypic movement. *P<0.05; **P<0.001 as compared
to the saline control response.
Figure 8. Effect of 34 on locomotor activity induced by cocaine in
mice (pretreatment experiments). A group of 9–10 male Swiss Webster
mice were injected with 34, 20 min before the ip injection of 20 mg/kg
cocaine. The data were recorded for 2 h. The first 30-min period was
selected for analysis of dose–response data using standard analysis of
variance. (a) Distance traveled. (b) Stereotypic movement. *P<0.05 as
compared to the group pretreated with vehicle.

S. Sakamuri et al. / Bioorg. Med. Chem. 11 (2003) 1123–1136
1131
of locomotor stimulant activity of 34 when tested alone
and its ability to antagonize cocaine-induced locomotor
activity could be due to its interaction with dopamine
receptors. Compound 34 was tested for its binding to
the D1, D2, and D3 receptors by the Medication Devel-
opment Division at the National Institute on Drug
Abuse (NIDA). No significant inhibition of ligand
binding toeither D , D2, o r D receptors was found for
We have chosen one potent analogue 34 tocarry out
behavioral pharmacological evaluations in rodents. Our
results showed that 34 has a behavioral pharmacological
profile different from cocaine. While 34 lacks cocaine’s
locomotor stimulant activity in mice, it dose-depen-
dently generalizes tococaine in drug discrimination test
in rats. Importantly, 34 was found to antagonize
cocaine-induced locomotor activity induced with an
estimated ED50 of 19 mg/kg. Taken together, our data
suggest that 34 represents a class of potent DAT inhi-
bitors with a novel chemical scaffold and a behavioral
pharmacological profile different from that of cocaine in
rodents. Thus, 34 may serve as a novel lead compound
in the ultimate development of therapeutic entities for
cocaine abuse and/or addiction. The ability of 34 to
antagonize cocaine-induced locomotor activity by a
mechanism that is incompletely understood also
emphasizes the need to carry out additional neuro-
pharmacological studies on this class of DAT inhibitors.
1
3
34 at 10 mM (data not shown). We also evaluated 34 for
its binding to serotonin receptors (5-HT1A, 5-HT2A, and
5-HT2C). Compound 34 has no significant inhibition of
ligand binding to5-HT
and 5-HT2C receptors at 10
1A
mM and only has a weak activity to 5-HT2A receptor
(Ki=3.5 mM), 100 times less potent than its activity at
the DAT site. Therefore, the lack of locomotor stimu-
lant activity of 34 is not due to its possible interaction
with the D1, D2 and D3 receptors and three serotonin
receptors (5-HT1A, 5-HT2A, and 5-HT2C).
Consistent with the dopamine hypothesis, most potent
DAT inhibitors were found to stimulate locomotor
activity and mimic cocaine’s discriminative stimulus.
However, there are some exceptions to this rule. For
example, Newman’s group found that some novel ben-
zotropine analogues produce locomotor stimulation,
but lack cocaine-like discriminative stimulus effects.24
Previously, we found that a potent DAT inhibitor (4)
has a behavioral pharmacological profile similar to that
of the novel benzotropine analogues reported by New-
man.10 However, 34 lacks locomotor stimulant activity
but fully generalizes to cocaine in drug discrimination
experiments. Furthermore, compound 34 dose-depen-
dently antagonizes locomotor activity induced by
cocaine in mice. Taken together, these data show that
34 has a behavioral pharmacological profile that is very
different from that of cocaine in rodents.
Experimental
Molecular modeling
Conformational analysis was performed using the con-
formational analysis module in the QUANTA pro-
gram.30 Generally, if a compound had fewer than five
rotatable single bonds, the grid scan conformational
search protocol was employed. In this protocol, each
rotatable bond was systematically rotated to generate a
starting conformation, which was subsequently mini-
mized using the CHARMm program31 within
QUANTA. If a compound had more than five rotatable
bonds, a random sampling protocol was used to gen-
erate conformations. Up to 5000 conformations were
generated and minimized. Energy minimization of
each conformation was computed with 5000 iterations
or until convergence, defined as an energy gradient of
0.001 kcal molꢀ1 Aꢀ1 or less. An Adopted Basis Newton–
Raphson algorithm, implemented in the CHARMm
program, was used for energy minimization. A constant
dielectric constant (equal to 1) was used throughout all
the calculations. Upon the completion of conformation
generation and energy minimization, the most stable
conformation was identified (the global minimum in
vacuum). It is noted, however, that the lowest energy
conformation may not be the bio-active conformation,
as was shown previously.32 For this reason, other low
energy conformations, typically within 5 kcal/mol of
the global minimum, were identified. Cluster analysis
was performed to determine the number of truly
unique conformations (clusters), using the cluster ana-
lysis module available in the QUANTA program.
These low energy conformational clusters together are
likely to include the bio-active conformations for a
compound.
Summary
Novel DAT inhibitors have promise as effective thera-
pies for the treatment of cocaine abuse. Toward this
end, we have employed a 3-D-database pharmacophore
searching approach. This approach has identified many
classes of novel DAT inhibitors and has provided a
considerable chemical diversity to the design of DAT
inhibitors. A major challenge is how to effectively opti-
mize the DAT inhibitors with weak activity discovered
by this approach.
In the current study, we reported our molecular model-
ing-assisted design and chemical modifications toward
optimization of one class of DAT inhibitors, 2,3-di-
substituted quinuclidines. The initial compound (12) has
Ki values of 7.2 and 8.9 mM in binding and inhibition of
dopamine reuptake, respectively. Through molecular
modeling-assisted design and chemical modifications,
potent analogues such as (ꢀ)-29 and 34 with much
improved potency were quickly identified. It is of note
that our design and optimization efforts have benefited
enormously from previous extensive SAR studies on
cocaine and WIN analogues, pioneered by Carroll’s
group.5,20
3-D-Database search
The Chem-X program (version July 96),33 running on a
Silicon Graphics Indigo2 R10000, was used to carry out
3-D-database pharmacophore searching. The primary

1132
S. Sakamuri et al. / Bioorg. Med. Chem. 11 (2003) 1123–1136
reason for choosing this program was its ability to gen-
erate and search multiple conformations for flexible
compounds in the database. The problem of multiple
conformations for flexible compounds was found to be
important in building and searching a 3-D-database
because flexible compounds may be able to adopt a
number of different conformations depending on their
environment. It is often difficult to know precisely
which conformation is the biologically active one if a
compound can adopt multiple conformations with little
energy difference. The biologically active conformations
may be different for the same compound when it binds
to different receptors. Therefore, it was decided that the
best way tohandle this was togenerate and search
multiple conformations for flexible compounds. The
ability of the Chem-X program to generate and search a
large number of conformations for flexible compounds
was found to be one key factor to our success in identify-
ing a large number of structurally novel, diverse lead
compounds in several projects carried out so far. We have
found that if only single conformations for flexible com-
pounds are searched, many identified lead compounds
would be missed. Therefore, multiple conformations for
flexible compounds are necessary. However, for a flexible
compound with more than 10 single bonds, using a step
size of 60ꢂ in generating conformations, the total number
of possible conformations will exceed 60 million. In prac-
tice, we set 3 million conformations as the maximum
number to be examined for any single compound.
tetrahydrofuran; DCM=dichloromethane; ether=diethyl
ether.
2-Methylene-3-quinuclidinone (16). A solution of 3-qui-
nuclidinone (15, 6.0 g, 48.0 mmol), 40% aqueous dime-
thylamine (10.0 mL, 72.0 mmol), 37% aqueous
formaldehyde (6.0 mL, 72.0 mmol), 20.0 mL of ethanol
and 8.0 mL of water was stirred at reflux for 1 h, then at
70 ꢂC for 17 h and allowed to cool to room temperature.
The solvents and excess reagents were evaporated in
vacuo and the oily residue fractionally distilled to pro-
vide 5.7 g (86%) of title compound as a light yellow oil,
ꢂ
1
bp 91–92 /7 mm. H NMR (300 MHz, CDCl3) d 1.90–
1.98 (4H, m), 2.51–2.55 (1H, narrow m), 2.87–2.98 (2H,
m), 3.03–3.13 (2H, m), 5.18 (1H, s), 5.78 (1H, s); 13C
NMR (CDCl3) d 24.9, 40.3, 48.3, 113.3, 152.3, 204.1.
Anal. (C8H11NO) C, H, N.
2-But-3-enylquinuclidin-3-one (17). To a solution of
.
CuI Me2S complex [prepared by the addition of Me2S
(0.8 mL, 10.9 mmol) to CuI (1.4 g, 7.3 mmol) at 0 ꢂC] in
THF at ꢀ78 ꢂC was added 1 M solution of allylmagne-
sium bromide (9.5 mL) and HMPA (2.5 mL, 15.6
mmol) stirred for 20 min. To this, a mixture 2-methyl-
ene-3-quinuclidinone (16, 1.0 g, 7.3 mmol) and TMS-Cl
(1.02 mL, 8.0 mmol) in THF was slowly added and
stirred at the same temperature for 2 h, quenched with
aq NH4Cl solution. The organic layer was separated
and the aqueous layer extracted with ethyl acetate, and
the combined organic layers were dried over Na2SO4
and evaporated to get the crude compound. This was
purified by column chromatography using ether:-
acetone: triethylamine in 85:10:5 ratio to afford the title
The initial compound (12) was identified through the
search of the NCI 3-D-database. The version of the
NCI 3-D-database34 used in our 3-D-database search
was built using the July 94 version of the Chem-X pro-
gram. It consists of 206,876 ‘open’ compounds, whose
structures and related biological data can be accessed by
the public.34 Employing the Chem-X program, a total
of 4094 compounds were identified as ‘hits’, that is,
compounds that met the requirements specified in the
pharmacophore model (Fig. 1). A number of additional
criteria were used in the selection of compounds for
biological evaluation in order to achieve maximum effi-
ciency in the discovery of lead compounds. These cri-
teria include simple chemical structure, small molecular
weight, non-peptidic and chemical structure diversity.
1
compound as a colorless oil (610 mg, 47%). H NMR
(300 MHz, CDCl3) d 1.46–1.59 (1H, m), 1.79–1.93 (5H,
m), 2.05–2.23 (2H, m), 2.30–2.35 (1H, m), 2.71–3.11
(5H, twom), 4.90–5.02 (2H, m), 5.68–5.82 (1H, m); 13C
NMR (CDCl3) d 25.3, 26.0, 27.1, 30.4, 39.8, 40.6, 48.5,
68.8, 115.1, 137.5, 221.7; MS m/z (%) 179 (6), 110 (100).
.
Anal. (C11H17NO HCl) C, H, N.
General procedure for the aryl Grignard addition. Tothe
ketone in dry THF at 0 ꢂC was added the appropriate
Grignard reagent (1.1 equiv). The mixture was stirred at
the same temperature for 30 min, quenched with aq
NH4Cl, and extracted with ethyl acetate. The combined
organic extracts were dried (Na2SO4) and concentrated
under reduced pressure. The resulting crude compound
was purified by column chromatography using ether/
acetone/triethylamine as eluent to afford the following
compounds:
Chemistry
General methods. THF was freshly distilled under
nitrogen from sodium benzophenone. 1H and 13C NMR
spectra were obtained with a Varian Unity Inova
1
instrument at 300 and 75.46 MHz, respectively. H che-
mical shifts (d) are reported in ppm downfield from
internal TMS. 13C chemical shifts are referenced to
CDCl3 (central peak, d=77.0 ppm).
2-But-3-enyl-3-phenylquinuclidin-3-ol (18). Colorless
1
thick syrup; yield 70%; H NMR (300 MHz, CDCl3) d
1.34–1.46 (3H, m), 1.48–1.58 (1H, m), 1.81–1.94 (2H,
m), 2.05–2.33 (4H, m), 2.64–2.75 (1H, m), 2.87 (2H,
broad t, J=8.3 Hz), 3.10–3.20 (1H, m), 3.35–3.42 (1H,
m), 4.98–5.07 (2H, m), 5.79–5.93 (1H, m), 7.30 (1H, d,
J=7.3 Hz), 7.39 (2H, t, J=7.1 Hz), 7.58 (1H, d, J=7.5
Hz); 13C NMR (CDCl3) d 21.8, 23.2, 26.1, 31.4, 35.6,
41.1, 48.8, 61.8, 75.1, 114.8, 126.0, 127.2, 128.2, 138.7,
146.2; MS m/z (%) 257 (12), 124 (100). Anal.
Melting points were determined in Pyrex capillaries with
a Thomas-Hoover Unimelt apparatus and are uncor-
rected. Mass spectra were measured in the EI mode at
an ionization potential of 70 eV. TLC was performed on
Merck silica gel 60F254 glass plates; column chroma-
tography was performed using Merck silica gel (60–200
mesh). The following abbreviations are used: THF=
.
(C17H23NO HCl) C, H, N.

S. Sakamuri et al. / Bioorg. Med. Chem. 11 (2003) 1123–1136
1133
2-But-3-enyl-3-(4-methylphenyl)quinuclidin-3-ol (19).
Colorless syrup; yield 74%; 1H NMR (300MHz, CDCl3)
d 1.32–1.54 (4H, m), 1.78–1.89 (2H, m), 2.01–2.28 (4H,
twom), 2.34 (3H, s), 2.62–2.71 (1H, m), 2.80–2.86 (2H,
m), 3.06–3.17 (1H, m), 3.29–3.34 (1H, m), 4.94–5.05 (2H,
m), 5.76–5.89 (1H, m), 7.16 (2H, d, J=8.6 Hz), 7.42 (2H,
d, J=6.6 Hz); 13C NMR (CDCl3) d 21.0, 22.0, 23.5,
26.2, 31.6, 35.8, 41.3, 49.0, 62.1, 75.1, 114.9, 126.1, 129.1,
2.70 (2H, m), 2.81–2.86 (1H, m), 2.96–3.05 (2H, m),
4.90–5.05 (2H, m), 5.75–5.88 (1H, m), 7.18 (4H, s); 13C
NMR (CDCl3) d 21.2, 29.4, 31.2, 32.4, 34.0, 49.2, 114.6,
127.6, 129.0, 136.1, 136.9, 138.7, 140.1, 147.0. Anal.
.
(C18H23N HCl) C, H, N.
2-But-3-enyl-2,3-didehydro-3-(3-methylphenyl)quinuclidine
1
(25). Colorless syrup; yield 68%; H NMR (300 MHz,
.
136.9, 139.0, 143.5. Anal. (C18H25NO HCl) C, H, N.
CDCl3) d 1.58–1.75 (4H, m), 2.71–2.39 (7H, m), 2.60–
2.70 (2H, m), 2.83–2.87 (1H, m), 2.97–3.05 (2H, m),
4.91–5.06 (2H, m), 5.76–5.89 (1H, m), 6.79–7.05 (3H,
m), 7.19–7.26 (1H, m); 13C NMR (CDCl3) d 21.4, 29.3,
30.2, 31.0, 32.2, 33.8, 48.9, 114.5, 124.7, 127.1, 127.9,
128.3, 137.6, 138.6, 139.6, 140.1, 147.1. Anal.
2-But-3-enyl-3-(3-methylphenyl)quinuclidin-3-ol (20).
Colorless syrup; yield 72%; 1H NMR (300 MHz, CDCl3)
d 1.34–1.58 (4H, m), 1.80–1.93 (2H, m), 2.06–2.31 (4H, m),
2.40 (3H, s), 2.66–2.75 (1H, m), 2.84–2.93 (2H, m), 3.10–
3.20 (1H, m), 3.38 (1H, dd, J=4.1, 9.8 Hz), 4.98–5.09 (2H,
m), 5.80–5.94 (1H, m), 7.12 (1H, d, J=7.3 Hz), 7.24–7.40
(3H, m); 13C NMR (CDCl3) d 21.6, 21.8, 23.2, 26.0, 31.4,
35.6, 41.1, 48.8, 61.8, 75.1, 114.7, 122.9, 126.9, 127.9, 128.1,
.
(C18H23N HCl) C, H, N.
2-But-3-enyl-2,3-didehydro-3-(2-methylphenyl)quinuclidine
1
(26). Colorless syrup; yield 62%; H NMR (300 MHz,
.
137.9, 138.8, 146.2. Anal. (C18H25NO HCl) C, H, N.
CDCl3) d 1.62–1.75 (4H, m), 2.04–2.12 (2H, m), 2.20–
2.27 (5H, m), 2.52–2.59 (1H, m), 2.62–2.75 (2H, m)
2.95–3.14 (2H, m), 4.85–5.01 (2H, m), 5.62–5.76 (1H,
m), 7.02–7.21 (4H, m); 13C NMR (CDCl3) d 20.4, 28.6,
29.9, 31.1, 31.6, 33.6, 48.9, 49.6, 114.3, 125.4, 126.7,
128.9, 129.9, 135.3, 138.6, 139.9, 140.3, 147.2. Anal.
2-But-3-enyl-3-(2-methylphenyl)quinuclidin-3-ol (21).
Colorless syrup; yield 68%; 1H NMR (300 MHz,
CDCl3) d 1.23–1.43 (4H, m), 1.72–1.87 (2H, m), 2.11–2.23
(2H, m), 2.27–2.44 (2H, m), 2.62 (3H, s), 2.65–2.93 (2H,
m), 3.10–3.20 (1H, m), 3.49 (1H, dd, J=2.9, 10.2 Hz),
4.99–5.10 (2H, m), 5.80–5.93 (1H, m), 7.10–7.31 (4H, m);
13C NMR (CDCl3) d 21.5, 22.5, 22.9, 25.4, 31.3, 32.3,
41.8, 48.9, 60.4, 76.6, 114.9, 125.0, 125.8, 127.2, 133.8,
.
(C18H23N HCl) C, H, N.
2-But-3-enyl-2, 3-didehydro-3-(4-chlorophenyl)quinuclidine
1
(27). Colorless syrup; yield 63%; H NMR (300 MHz,
.
138.6, 138.8, 141.8. Anal. (C18H25NO HCl) C, H, N.
CDCl3) d 1.57–1.74 (5H, m), 1.91–2.01 (1H, m), 2.27–
2.37 (5H, m), 2.61–2.71 (2H, m), 2.81–2.84 (1H, m),
2.99–3.08 (2H, m), 4.94–5.07 (2H, m), 5.75–5.87 (1H,
m), 7.19 (2H, d, J=6.4 Hz), 7.31 (2H, d, J=6.4 Hz); 13C
NMR (CDCl3) d 29.5, 31.2, 32.3, 34.0, 49.2, 114.9,
128.5, 129.1, 132.3, 138.4, 138.6, 139.4, 148.3. Anal.
2-But-3-enyl-3-(4-chlorophenyl)quinuclidin-3-ol (22).
Light-yellow syrup, yield 70%; 1H NMR (300 MHz,
CDCl3) d 1.32–1.54 (4H, m), 1.77–2.26 (5H, m), 2.64–
2.73 (1H, m), 2.85 (2H, t, J=7.9 Hz), 3.06–3.16 (1H, m),
3.27 (1H, dd, J=4.4, 9.5 Hz), 4.96–5.04 (2H, m), 5.75–
5.89 (1H, m), 7.34 (2H, d, J=6.6 Hz), 7.47 (2H, d,
J=6.6 Hz); 13C NMR (CDCl3) d 21.6, 23.3, 26.2, 31.5,
35.9, 41.2, 48.9, 62.1, 75.0, 115.2, 127.8, 128.6, 133.2,
.
(C17H20ClN HCl) C, H, N.
General procedure for the hydrogenation. A mixture of
olefin and a catalytic amount of Pd/C in EtOH was
hydrogenated under 60 psi of H2 at 25 ꢂC fo r 24 h. The
catalyst was filtered off, and the filtrate was con-
centrated to give the crude compound as a yellow syrup,
which on purification by column chromatography with
ether/triethylamine afforded the saturated compound as
a colorless thick syrup in quantitative yield.
.
138.8, 145.1. Anal. (C17H22ClNO HCl) C, H, N.
General procedure for the dehydration. To a solution of
hydroxy compound in EtOH, 6 N HCl was added,
refluxed overnight and cooled to room temperature.
The reaction mixture was neutralized by slow addition
of solid NaHCO3 and extracted with ethyl acetate. The
combined organic layers were washed with satd NaCl
solution, dried (Na2SO4) and concentrated to get the
crude compound, which was purified by passing through
a silica gel column using acetone/ether as eluent.
2-Butyl-3-phenylquinuclidine (28). Colorless syrup; 1H
NMR (300 MHz, CDCl3) d 0.80 (3H, t, J=7.1 Hz),
1.07–1.37 (6H, twom), 1.46–1.54 (1H, m), 1.70–1.76
(2H, m), 2.01–2.10 (2H, m), 2.67–2.78 (1H, m), 2.96–
3.05 (1H, m), 3.09–3.29 (4H, m), 7.19–7.34 (5H, m); 13C
NMR (CDCl3) d 14.0, 22.3, 22.7, 26.8, 29.7, 30.2, 30.3,
40.7, 45.4, 49.4, 60.2, 125.5, 127.8, 128.9, 142.9; MS m/z
2-But-3-enyl-2,3-didehydro-3-phenylquinuclidine (23).
Colorless syrup; yield 61%; 1H NMR (300MHz,
CDCl3) d 1.62–1.79 (4H, m), 2.30–2.42 (4H, m), 2.64–2.73
(2H, m), 2.86–2.92 (1H, narrow m), 3.01–3.10 (2H, m),
4.95–5.08 (2H, m), 5.78–5.90 (1H, m), 7.24–7.29 (3H,
m), 7.37 (2H, t, J=7.6 Hz); 13C NMR (CDCl3) d 29.1,
30.8, 32.2, 38.8, 48.9, 114.5, 126.4, 127.6, 128.1, 138.4,
139.5, 140.2, 146.9; MS m/z (%) 239 (22), 82 (100),
.
(%) 243 (18), 42 (100). Anal. (C17H25N HCl) C, H, N.
2-Butyl-3-(4-methylphenyl)quinuclidine (29). Colorless
syrup; 1H NMR (300 MHz, CDCl3) d 0.77 (3H, t,
J=6.8 Hz), 1.02–1.32 (6H, twom), 1.40–1.49 (1H, m),
1.65–1.72 (2H, m), 1.96–2.06 (2H, m), 2.32 (3H, s),
2.64–2.74 (1H, m), 2.89–3.23 (5H, twom), 7.06–7.14
(4H, m); 13C NMR (CDCl3) d 14.2, 21.1, 22.5, 22.9,
27.2, 29.9, 30.4, 30.6, 40.9, 45.3, 49.7, 60.4, 128.8, 129.0,
135.1, 140.0; MS m/z (%) 257 (29), 42 (100). Anal.
.
Anal. (C17H21N HCl) C, H, N.
2-But-3-enyl-2,3-didehydro-3-(4-methylphenyl)quinuclidine
1
(24). Colorless syrup; yield 66%; H NMR (300 MHz,
CDCl3) d 1.56–1.75 (4H, m), 2.28–2.38 (7H, m), 2.60–
.
(C18H27N HCl) C, H, N.

1134
S. Sakamuri et al. / Bioorg. Med. Chem. 11 (2003) 1123–1136
General procedure for the terminal double bond reduction.
To a solution of olefin in toluene was added Chloro-
tris(triphenylphosphine)rhodium(1) and heated at 70 ꢂC
under hydrogen atmosphere for 20 h. Volatiles were
removed under vacuum and passed through a small bed
of silica gel using acetone/hexane as eluent.
mobile phase and the separation was carried out by
injecting 30 mL on a 250ꢃ10 mm chiral column.
X-ray crystallography. Crystal structure details: mono-
clinic crystal in space group P21/c, a=18.059(9),
b=9.162(2), and c=15.799(2) A, b=92.21(3)ꢂ, V=2612(1)
A3, Z=4. Dx=1.40
g
cmꢀ3
,
ꢀ=1.62 mmꢀ1
,
2-Butyl-2,3-didehydro-3-phenylquinuclidine (30). Color-
less liquid; yield 94%; 1H NMR (300 MHz, CDCl3) d 0.91
(3H, t, J=7.3 Hz), 1.28–1.38 (2H, m), 1.54–1.78 (6H, two
m), 2.36 (2H, broad d, J=8.1 Hz), 2.63–2.72 (2H, m),
2.86–2.89 (1H, narrow m), 3.00–3.09 (2H, m), 7.24–7.28
(3H, m), 7.36 (2H, broad t, J=7.5 Hz); 13C NMR (CDCl3)
d 14.0, 23.3, 29.3, 30.5, 31.5, 33.7, 49.0, 126.2, 127.6, 128.0,
F(000)=1160, T=293 K. Final agreement factor was
R(F)=0.099 for the 1623 observed data. The higher
than normal R-factor was due in part to the extended
alkane side chain being disordered over at least two dif-
ferent positions as well as to the possibility of some unre-
solved twinning. Atomic coordinates have been deposited
with the Cambridge Crystallographic Data Centre.35
.
139.2, 139.9, 148.4. Anal. (C17H23N HCl) C, H, N.
Pharmacology
2-Butyl-2,3-didehydro-3-(4-methylphenyl)quinuclidine
1
(31). Colorless liquid; yield 96%; H NMR (300 MHz,
[3H]Mazindol binding. Binding assays were conducted as
previously described.10 Briefly conventional P2 mem-
brane pellets were prepared by differential centrifu-
gation from rat striatum. The P2 pellet was resuspended
in Krebs–Ringer–HEPES (KRH) buffer consisting of
(in mM): NaCl (125), KCl (4.8), MgSO4 (1.2), CaCl2
(1.3), KH2PO4 (1.2), glucose (5.6), nialamide (0.01), and
HEPES (25) (pH 7.4) and centrifuged again. Finally, the
pellet was resuspended in 30 volumes of buffer, pelleted
at 15,000g and frozen at ꢀ80 ꢂC until used. The striatal
homogenates were thawed by resuspension in the buf-
fer described above at 75–125 mg protein/mL and
incubated with [3H]mazindol, with or without compet-
ing drugs, for 60 min in a 4 ꢂC cold room. Nonspecific
binding was determined with 30 mM cocaine. The bound
and free [3H]mazindol were separated by rapid vacuum
filtration over Whatman GF/C filters, using a Brandel
M24R cell harvester, followed by two washes with 5 mL
of cold buffer. Radioactivity on the filters was then
extracted by allowing to sit overnight with 5 mL of
scintillant. The vials were vortexed and counted. IC50
values were determined using the computer program
LIGAND.
CDCl3) d 0.88 (3H, t, J=7.4 Hz), 1.25–1.38 (2H, m),
1.51–1.75 (6H, m), 2.20 (2H, broad t, J=8.0 Hz), 2.35
(3H, s), 2.59–2.69 (2H, m), 2.80–2.86 (1H, narrow m),
2.96–3.05 (2H, m), 7.10–7.17 (4H, m); 13C NMR
(CDCl3) d 14.0, 21.1, 23.0, 29.3, 30.5, 31.5, 33.7, 49.0,
127.5, 128.7, 135.8, 136.9, 139.0, 147.9; MS m/z (%) 255
.
(37), 213 (100). Anal. (C18H25N HCl) C, H, N.
2-Butyl-2,3-didehydro-3-(3-methylphenyl)quinuclidine (32).
Colorless liquid; yield 97%; 1H NMR (300 MHz,
CDCl3) d 0.88 (3H, t, J=6.5 Hz), 1.26–1.36 (2H, m),
1.52–1.71 (6H, m), 2.20 (2H, t, J=8.1 Hz), 2.36 (3H, s),
2.60–269 (2H, m), 2.80–2.86 (1H, m), 2.97–3.06 (2H, m),
7.01–7.06 (3H, m), 7.23 (1H, t, J=7.8 Hz); 13C NMR
(CDCl3) d 13.9, 21.5, 22.9, 29.4, 30.5, 31.5, 33.7, 49.1,
124.7, 127.0, 127.9, 128.3, 137.6, 139.3, 139.9, 148.2; MS
.
m/z (%) 255 (24), 42 (100). Anal. (C18H25N HCl) C, H,
N.
2-Butyl-2,3-didehydro-3-(2-methylphenyl)quinuclidine (33).
Colorless liquid; yield 92%; 1H NMR (300 MHz,
CDCl3) d 0.79 (3H, t, J=7.4 Hz), 1.14–1.27 (2H, m),
1.43–1.49 (2H, m), 1.62–1.74 (4H, m), 1.94–2.01 (2H,
m), 2.25 (3H, s), 2.55–2.60 (1H, m), 2.63–2.73 (2H, m),
2.98–3.12 (2H, m), 7.02–7.06 (1H, m), 7.11–7.26 (3H,
m); 13C NMR (CDCl3) d 13.9, 20.4, 22.7, 28.6, 29.7,
29.9, 31.6, 33.6, 49.1, 49.6, 125.4, 126.6, 128.9, 129.8,
135.4, 139.5, 140.2, 148.2; MS m/z (%) 255 (26), 56
Synaptosomal uptake of [3H]DA. The effect of candidate
compounds in antagonizing dopamine high-affinity
uptake was determined using a method previously
employed.10 For [3H]DA uptake, freshly dissected rat
striata were homogenized with a Teflon–glass pestle in
ice-cold 0.32 M sucrose and centrifuged for 10 min at
1000 g. The supernatant was centrifuged at 17,500 g for
20 min. This P2 synaptosomal pellet was resuspended in
30 volumes of ice-cold modified KRH buffer. An ali-
quot of the synaptosomal suspension was preincubated
with the buffer and drug for 30 min at 37 ꢂC, and uptake
initiated by the addition of [3H]dopamine (5 nM, final
concn). After 5 min, uptake was terminated by adding 5
mL of cold buffer containing glucosamine as a sub-
stitute for NaCl and then finally by rapid vacuum fil-
tration over GF-C glass fiber filters, followed by
washing with two 5 mL volumes of ice-cold, sodium-
free buffer. Radioactivity retained on the filters was
determined by liquid scintillation spectrometry. Specific
uptake is defined as that which is sensitive toinhibition by
30 mM cocaine. It is identical to that calculated by sub-
tracting the mean of identical tubes incubated at 0ꢂC.
.
(100). Anal. (C18H25N HCl) C, H, N.
2-Butyl-2,3-didehydro-3-(4-chlorophenyl)quinuclidine (34).
Colorless liquid; yield 89%; 1H NMR (300 MHz,
CDCl3) d 0.88 (3H, t, J=7.3 Hz), 1.24–1.34 (2H, m),
1.50–1.75 (6H, m), 2.18 (2H, t, J=8.0 Hz), 2.58–2.67
(2H, m), 2.78–2.82 (1H, m), 2.97–3.06 (2H, m), 7.15
(2H, d, J=7.7 Hz), 7.30 (2H, d, J=8.3 Hz); 13C NMR
(CDCl3) d 14.2, 23.2, 29.5, 30.6, 31.7, 33.9, 49.2, 125.7,
127.8, 128.3, 128.4, 129.1, 132.2, 138.5, 149.3; MS m/z (%)
.
275 (18), 233 (100). Anal. (C17H22ClN HCl) C, H, N.
HPLC separation of (ꢁ)-29. The chiral HPLC was per-
formed on a Shimadzu SCL-10A-VP system at a flow
rate of 5 mL/min at room temperature and UV detec-
tion at 254 and 280 nm. The sample for injection was
prepared by dissolving racemic compound (5 mg/mL) in

S. Sakamuri et al. / Bioorg. Med. Chem. 11 (2003) 1123–1136
1135
IC50 values were determined by a computer-assisted,
Table 2. Elemental analysis
iterative fit to a four-parameter sigmoidal equation
(ALLFIT). These values were then converted to Ki
values according to the Cheng–Prusoff equation assum-
ing classical competitive inhibition. Preincubation of the
drug and synaptosomes at 37 ꢂC for 30 min was used to
approximate equilibrium conditions as necessary to
satisfy the requirements of the Cheng–Prusoff equation.
Compd Molecular
No.
Calculated
Found
H
formula
C
H
N
C
N
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
C8H11NO
C11H17NO HCl
C17H23NO HCl
C18H25NO HCl
C18H25NO HCl
70.04 8.08 10.21 70.06 8.19 10.20
61.25 8.41 6.49 61.18 8.52 6.44
69.49 8.23 4.77 69.41 8.17 4.69
70.22 8.51 4.55 70.34 8.42 4.59
70.22 8.51 4.55 70.09 8.61 4.48
70.22 8.51 4.55 69.98 8.46 4.43
.
.
.
.
.
C18H25NO HCl
.
Behavioral pharmacology
C
17H22ClNO HCl 62.20 7.06 4.27 62.11 6.97 4.18
.
C17H21N HCl
C18H23N HCl
C18H23N HCl
74.03 8.04 5.08 73.95 7.94 5.03
74.59 8.35 4.83 74.52 8.26 4.79
74.59 8.35 4.83 74.41 8.34 4.73
74.59 8.35 4.83 74.72 8.48 4.70
65.81 6.82 4.51 65.69 6.94 4.66
72.96 9.36 5.01 73.06 9.27 4.99
73.57 9.60 4.77 73.49 9.60 4.55
73.49 8.71 5.04 73.22 8.62 4.97
74.07 8.98 4.80 74.19 9.10 4.92
74.07 8.98 4.80 73.97 8.89 4.76
74.07 8.98 4.80 74.10 8.90 4.81
65.38 7.42 4.49 65.34 7.38 4.40
Drug discrimination. The drug discrimination study was
conducted using male Sprague–Dawley rats according
tothe procedure described elsewhere. 36 Rats were trained
todiscriminate 10 mg/kg ip cocaine from saline. All drugs
were administered ip in a volume of 1 mL/kg 10 min prior
to the testing. The response rate on both keys and the
percent cocaine lever-appropriate responding were calcu-
lated for each rat. The response rates following a given test
drug injection were presented as the percent of its corre-
sponding vehicle control response rates.
.
.
.
C18H23N HCl
.
C17H20ClN HCl
.
C17H25N HCl
.
C18H27N HCl
.
C17H23N HCl
.
C18H25N HCl
.
C18H25N HCl
.
C18H25N HCl
.
C17H22ClN HCl
Acknowledgements
Locomotor activity
Drugs alone. Locomotor activity of male Swiss Webster
mice was recorded using Truscan activity monitors
(Coulbourn Instruments, Allentown, PA) and a compu-
ter. The activity monitors consist of acrylic chambers,
which are placed inside the sensor ring. The sensor ring is
equipped with light-sensitive detectors and the infrared
light beams. The X–Y coordinates of the body center of
the subject are sampled by scanning the beams and then
the successive locations of coordinates are compared.
The sum of distances between successive coordinates is
measured as the distance traveled, while the total num-
ber of coordinate changes are recorded as the stereo-
typic movements. Following 1 h of habituation to test
arenas, several groups of mice were injected intraper-
itoneally (ip) with different doses of cocaine, compound
34 or its corresponding vehicles, saline and 10%
DMSO, in a volume of 10 mL/kg. Locomotor activity
was recorded in 10-min bins for the next 2 h. The raw data
were converted to 30-min totals. The maximal 30-min
activity occurring within the 2-h session following a given
test drug injection was determined for each dose level and
expressed as the percent of its corresponding vehicle con-
trol response for plotting the dose–response curves.
We are indebted to the National Institute on Drug
Abuse (NIDA), National Institutes of Health and the
Office of Naval Research for their financial support of
this work. The sample of the compounds 7–12 used in
the initial screening was kindly provided by the Chemi-
cal Synthesis Branch, the Developmental Therapeutics
Program, National Cancer Institute, NIH and their help
is greatly appreciated. Receptor binding at dopamine
and serotonin receptors for 34 was provided by the
Medications Development Division, National Institute
on Drug Abuse under NIDA contract #N01DA-7-8072,
‘Receptor Activity Testing for Medication Discovery’,
and their collaboration and help on this project is
greatly appreciated.
References and Notes
1. Carrol, F. I.; Howell, L. L.; Kuhar, M. J. J. Med. Chem.
1999, 42, 2721.
2. Ritz, M. C.; Lamb, R. J.; Goldberg, S. R.; Kuhar, M. J.
Science 1987, 237, 1219.
3. Carroll, F. I.; Lewin, A. H.; Boja, J. W.; Kuhar, M. J.
J. Med. Chem. 1992, 35, 969.
4. Kuhar, M. J.; Ritz, M. C.; Boja, J. W. Trends Neurosci.
1991, 14, 299.
5. Carroll, F. I.; Lewin, A. H.; Biswas, J. Pharm. News 1994,
1, 11.
6. Wise, R. A. Brain Res. 1978, 152, 215.
7. Smith, M. P.; Hoepping, A.; Johnson, K. M.; Trzcinska,
M.; Kozikowski, A. P. Drug Discovery Today 1999, 7, 322.
8. Howell, L. L.; Czoty, P. W.; Kuhar, M. J.; Carroll, F. I.
J. Pharmacol. Exp. Ther. 2000, 292, 521.
9. Glowa, J. R.; Wojnicki, F. H. E.; Matecka, D.; Rice, K. C.;
Rothman, R. B. Exp. Clin. Psychopharmacol. 1995, 3, 232.
10. Wang, S.; Sakamuri, S.; Enyedy, I. J.; Kozikowski, A. P.;
Deschaux, O.; Bandyopadhyay, B. C.; Tella, S. R.; Zaman,
W. A.; Johnson, K. M. J. Med. Chem. 2000, 43, 351.
11. Wang, S.; Sakamuri, S.; Istvan, J.; Enyedy, I. J.; Alan, P.;
Kozikowski, A. P.; Zaman, W. A.; Johnson, K. M. Bioorg.
Med. Chem. 2001, 9, 1753.
Interactions between cocaine and drugs. Following 40
min of habituation to test environment, different doses
of drug (34 or saline) were injected ip in a volume of 1
mL/100 g body weight and immediately placed back in
the activity monitors. Twenty min later, a 20-mg/kg
dose of cocaine was injected ip to all groups of animals
and immediately placed back in the activity monitors.
The data were recorded for a minimum of 2 h. Each
dose of a given drug was studied in a separate group of
mice and each mouse was used only once. The 30-min
period responses were computed from the 2-h data. The
30-min period during which the maximal responses
would occur was used for plotting the dose–response
function. Data were analyzed using standard analysis of
variance (Table 2).

1136
S. Sakamuri et al. / Bioorg. Med. Chem. 11 (2003) 1123–1136
12. Sakamuri, S.; Enyedy, I. J.; Kozikowski, A. P.; Zaman,
W. A.; Johnson, K. M.; Wang, S. Bioorg. Med. Chem. Lett.
2001, 11, 495.
Flippen-Anderson, J. L.; Zhang, M.; Wang, C. Z.; Johnson,
K. M.; Tella, S.; Kozikowski, A. P. J. Med. Chem. 2001, 44,
1615.
13. Enyedy, I. J.; Zaman, W. A.; Sakamuri, S.; Kozikowski,
A. P.; Johnson, K. M.; Wang, S. Bioorg. Med. Chem. Lett.
2001, 11, 1113.
14. Enyedy, I. J.; Wang, J.; Zaman, W. A.; Johnson, K. M.;
Wang, S. Bioorg. Med. Chem. Lett. 2002, 12, 1775.
15. Sakamuri, S.; Enyedy, I. J.; Kozikowski, A. P.; Wang, S.
Tetrahedron Lett. 2000, 41, 9949.
27. Cline, E. J.; Scheffel, U.; Boja, J. W.; Carroll, F. I.; Katz,
J. L.; Kuhar, M. J. J. Pharmacol. Exp. Ther. 1992, 260, 1174.
28. Vaugeois, J. M.; Bonnet, J. J.; Duterte-Boucher, D. Eur.
J. Pharmacol. 1993, 230, 195.
29. Izenwasser, S.; Terry, P.; Heller, B.; Witkin, J. M.; Katz,
J. L. Eur. J. Pharmacol. 1994, 263, 277.
30. QUANTA, a molecular modeling system, is supplied by
Molecular Simulations Inc., 9685 Scranton Road, San Diego,
CA 92121-3752.
31. Brooks, B. R.; Bruccoleri, R. E.; Olafson, B. D.; States,
D. J.; Swaminathan, S.; Jaun, B.; Karplus, M. J. Comput.
Chem. 1983, 4, 187.
32. Nicklaus, M. C.; Wang, S.; Driscoll, J.; Milne, G. W. A.
Bioorg. Med. Chem. 1995, 3, 411.
33. Chem-X is a product of Oxford Molecular Group PLC,
the Medawar Centre, Oxford Science Park, Oxford OX4 4GA,
UK.
34. Milne, G. W. A.; Nicklaus, M. C.; Driscoll, J. S.; Wang,
S.; Zaharevitz, D. W. J. Chem. Inf. Comput. Sci. 1994, 34,
1219.
35. Crystallographic data (excluding structure factors) for the
structures in this paper have been deposited with the Cam-
bridge Crystallographic Data Centre as supplementary pub-
lication no. CCDC 188259. Copies of the data can be
obtained, free of charge, on application to CCDC, 12 Union
Road, Cambridge CB2 1EZ, UK (fax: +44-(0)1223-336033 or
e-mail: deposit@ccdc.cam.ac.uk).
16. Wang, S. et al. Unpublished results.
17. Kitayama, S.; Shimada, S.; Xu, H.; Markham, L.; Dono-
van, D. M.; Uhl, G. R. Proc. Natl. Acad. Sci. U.S.A. 1992, 89,
7782.
18. Lin, Z.; Wang, W.; Uhl, G. R. Mol. Pharmacol. 2000, 58,
1581.
19. Kozikowski, A. P.; Roberti, M.; Xiang, L.; Bergmann,
J. S.; Callahan, P. M.; Cunningham, K. A.; Johnson, K. M.
J. Med. Chem. 1992, 35, 4764.
20. Carroll, F. I.; Gao, Y.; Rahman, M. A.; Abraham, P.;
Parham, K.; Lewin, A. H.; Boja, J. W.; Kuhar, M. J. J. Med.
Chem. 1991, 34, 2719.
21. Hansen, A. R.; Bader, H. J. Heterocycl. Chem. 1966, 3, 109.
22. Cleveland, T. J. Liq. Chromatogr. 1995, 18, 649.
23. Carroll, F. I.; Kotian, P.; Dehghani, A.; Gray, J. L.;
Kuzemko, M. A.; Parham, K. A.; Abraham, P.; Lewin, A. H.;
Boja, J. W.; Kuhar, M. J. J. Med. Chem. 1995, 38, 379.
24. Newman, A. H.; Allen, A. C.; Izenwasser, S.; Katz, J. L.
J. Med. Chem. 1994, 37, 2258.
25. Acri, J. B.; Siedleck, B. K.; Witkin, J. M. J. Pharmacol.
Exp. Ther. 1996, 277, 198.
26. (a) Zhao, L.; Johnson, K. M.; Zhang, M.; Flippen-
Anderson, J.; Kozikowski, A. P. J. Med. Chem. 2000, 43,
3283. (b) Tamiz, A. P.; Bandyopadhyay, B. C.; Zhang, J.;
36. Yasar, S.; Schindler, C. W.; Thorndike, E. B.; Szelenyi, I.;
Goldberg, S. R. J. Pharmacol. Exp. Ther. 1993, 265, 1.
37. Lieske, S. F.; Yang, B.; Eldefrawi, M. E.; MacKerell,
A. D. Jr.; Wright, J. J. Med. Chem. 1998, 41, 864.