Highly selective chiral N-substituted 3α-[bis(4'- fluorophenyl)methoxy]tropane analogues for the dopamine transporter: Synthesis and comparative molecular field analysis

Article abstract of DOI:10.1021/jm990265s

In a continuing effort to further characterize the role of the dopamine transporter in the pharmacological effects of cocaine, a series of chiral and achiral N-substituted analogues of 3α-[bis(4'-fluorophenyl)methoxy]tropane (5) has been prepared as potential selective dopamine transporter ligands. These novel compounds displaced [₃H]WIN 35,428 binding from the dopamine transporter in rat caudate putamen with K(i) values ranging from 13.9 to 477 nM. Previously, it was reported that 5 demonstrated a significantly higher affinity for the dopamine transporter than the parent drug, 3α- (diphenylmethoxy)tropane (3; benztropine). However, 5 remained nonselective over muscarinic m₁ receptors (dopamine transporter, K(i) = 11.8 nM; m₁, K(i) = 11.6 nM) which could potentially confound the interpretation of behavioral data, for this compound and other members of this series. Thus, significant effort has been directed toward developing analogues that retain high affinity at the dopamine transporter but have decreased affinity at muscarinic sites. Recently, it was discovered that by replacing the N-methyl group of 5 with the phenyl-n-butyl substituent (6) retention of high binding affinity at the dopamine transporter (K(i) = 8.51 nM) while decreasing affinity at muscarinic receptors (K(i) = 576 nM) was achieved, resulting in 68-fold selectivity. In the present series, a further improvement in the selectivity for the dopamine transporter was accomplished, with the chiral analogue (S)N-(2-amino-3-methyl-n-butyl)-3α-[bis(4'- fluorophenyl)methoxy]tropane (10b) showing a 136-fold selectivity for the dopamine transporter versus muscarinic m₁ receptors (K(i) = 29.5 nM versus K(i) = 4020 nM, respectively). In addition, a comparative molecular field analysis (CoMFA) model was derived to correlate the binding affinities of all the N-substituted 3α-[bis(4'-fluorophenyl)methoxy]tropane analogues that we have prepared with their 3D-structural features. The best model (q² = 0.746) was used to accurately predict binding affinities of compounds in the training set and in a test set. The CoMFA coefficient contour plot for this model, which provides a visual representation of the chemical environment of the binding domain of the dopamine transporter, can now be used to design and/or predict the binding affinities of novel drugs within this class of dopamine uptake inhibitors.

Full text of DOI:10.1021/jm990265s

J . Med. Chem. 2000, 43, 1085-1093
1085
High ly Selective Ch ir a l N-Su bstitu ted 3r-[Bis(4′-flu or op h en yl)m eth oxy]tr op a n e
An a logu es for th e Dop a m in e Tr a n sp or ter : Syn th esis a n d Com p a r a tive
Molecu la r F ield An a lysis
Michael J . Robarge,# Gregory E. Agoston,†,# Sari Izenwasser,‡, Theresa Kopajtic, Clifford George,§
J onathan L. Katz, and Amy Hauck Newman*,#
Medicinal Chemistry and Psychobiology Sections, National Institute on Drug Abuse-Intramural Research Program,
National Institutes of Health, 5500 Nathan Shock Drive, Baltimore, Maryland 21224, and Laboratory for the Structure of
Matter, Naval Research Laboratory, Washington, D.C. 20375-5000
Received May 28, 1999
In a continuing effort to further characterize the role of the dopamine transporter in the
pharmacological effects of cocaine, a series of chiral and achiral N-substituted analogues of
3R-[bis(4′-fluorophenyl)methoxy]tropane (5) has been prepared as potential selective dopamine
transporter ligands. These novel compounds displaced [3H]WIN 35,428 binding from the
dopamine transporter in rat caudate putamen with Ki values ranging from 13.9 to 477 nM.
Previously, it was reported that 5 demonstrated a significantly higher affinity for the dopamine
transporter than the parent drug, 3R-(diphenylmethoxy)tropane (3; benztropine). However, 5
remained nonselective over muscarinic m1 receptors (dopamine transporter, Ki ) 11.8 nM; m1,
Ki ) 11.6 nM) which could potentially confound the interpretation of behavioral data, for this
compound and other members of this series. Thus, significant effort has been directed toward
developing analogues that retain high affinity at the dopamine transporter but have decreased
affinity at muscarinic sites. Recently, it was discovered that by replacing the N-methyl group
of 5 with the phenyl-n-butyl substituent (6) retention of high binding affinity at the dopamine
transporter (Ki ) 8.51 nM) while decreasing affinity at muscarinic receptors (Ki ) 576 nM)
was achieved, resulting in 68-fold selectivity. In the present series, a further improvement in
the selectivity for the dopamine transporter was accomplished, with the chiral analogue (S)-
N-(2-amino-3-methyl-n-butyl)-3R-[bis(4′-fluorophenyl)methoxy]tropane (10b) showing a 136-
fold selectivity for the dopamine transporter versus muscarinic m1 receptors (Ki ) 29.5 nM
versus Ki ) 4020 nM, respectively). In addition, a comparative molecular field analysis (CoMFA)
model was derived to correlate the binding affinities of all the N-substituted 3R-[bis(4′-
fluorophenyl)methoxy]tropane analogues that we have prepared with their 3D-structural
features. The best model (q2 ) 0.746) was used to accurately predict binding affinities of
compounds in the training set and in a test set. The CoMFA coefficient contour plot for this
model, which provides a visual representation of the chemical environment of the binding
domain of the dopamine transporter, can now be used to design and/or predict the binding
affinities of novel drugs within this class of dopamine uptake inhibitors.
In tr od u ction
further illicit drug taking. To this end, the preponder-
ance of research suggests that the primary mechanism
by which cocaine exerts its reinforcing and psychomotor
stimulant effects appears to be through the blockade of
the dopamine transporter and subsequent accumulation
of dopamine in the synapse.1,2 As a result, a thorough
understanding of the structure and function of the
dopamine transporter and of the cocaine binding site
has been deemed critical for development of drugs that
may be useful in the treatment of cocaine abuse. To
date, the dopamine transporter has been cloned and
expressed, but its 3D-structure is unknown.3-6 In efforts
to develop a pharmacotherapeutic in absence of topo-
logical information about the dopamine transporter,
considerable research has been conducted toward the
elucidation of the cocaine recognition site or sites
through the use of novel, high-affinity ligands.7,8 Struc-
ture-activity relationship (SAR) studies of cocaine and
3â-phenyltropane-2â-carboxylic acid methyl ester (WIN
35,065-2; 2) analogues have proven useful in the iden-
The socioeconomic costs and serious health risks
associated with illicit drug use, and in particular cocaine
(1) abuse, have accentuated the need for prevention and
treatment. One approach to the development of a
pharmaceutical treatment for cocaine abuse has been
to focus on identifying the mechanistic correlates to
cocaine’s pharmacological effects. The prevailing concept
that underlies this pursuit has been that as we better
understand how cocaine produces its reinforcing effects
that may lead to its abuse and addiction, we can
potentially intervene on a molecular level, to prevent
* To whom correspondence should be addressed. Tel: 410-550-1455.
Fax: 410-550-1648. E-mail: anewman@intra.nida.nih.gov.
#
Medicinal Chemistry Section, NIH.
Psychobiology Section, NIH.
§ Naval Research Laboratory.
† Current address: Entremed, Inc., 9640 Medical Center Dr.,
Rockville, MD 20850.
‡ Current address: University of Miami School of Medicine, Dept.
of Neurology, 1501 NW 9th Ave., Rm. #4061, Miami, FL 33136.
10.1021/jm990265s This article not subject to U.S. Copyright. Published 2000 by the American Chemical Society
Published on Web 02/25/2000

1086 J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 6
Robarge et al.
tification of structural requirements for potency and
selectivity at the dopamine transporter within this class
of drugs.7
gence from cocaine and related analogues at the dopa-
mine transporter.27 Recently a photoaffinity ligand, [125I]-
N-[n-butyl-4-(4′′′-azido-3′′′-iodophenyl)]-4′,4′′-difluoro-
3R-(diphenylmethoxy)tropane, has been shown to label
a binding site on the dopamine transporter that is
distinct from that which is labeled by the cocaine-based
photoaffinity label, RTI 82 ([125I]3â-(p-chlorophenyl)-
tropane-2â-carboxylic acid 4′-azido-3′-iodophenylethyl
ester).17,28 Taken together, these studies suggest that
this class of dopamine uptake inhibitors may be access-
ing a binding site distinct from that of cocaine. We have
hypothesized that the discrepant behavioral profile of
this class of dopamine uptake inhibitors may, in part,
be due to interactions at a distinctive binding domain
on the dopamine transporter.
The generation of 3D-quantitative structure-activity
relationship (3D-QSAR) models using comparative mo-
lecular field analysis (CoMFA)9 has been explored in the
3â-phenyltropane-2â-carboxylic acid methyl ester class
of dopamine uptake inhibitors.10-14 The molecular mod-
els generated have provided predictive models of the
steric and electrostatic environment of the binding site,
which in turn can provide guidance in optimization of
further 2- and 3-substituted tropane analogues. These
reports demonstrate the usefulness of molecular model-
ing in the determination of binding site topology and
generation of high-affinity drugs. While other structur-
ally distinct dopamine uptake inhibitors may access
binding sites that are unique,15-17 molecular models of
these compounds at the dopamine transporter have been
limited to one report.18 Thus, the need for further
investigation using this technique is warranted and will
undoubtedly provide additional information on the
structure and function of the dopamine transporter.
In our initial studies on the 3R-(diphenylmethoxy)-
tropanes we found that while certain aryl ring substitu-
tions improved binding affinity at the dopamine trans-
porter, high selectivity for the dopamine transporter
over muscarinic m1 receptors was not achieved.22-24 As
a result, it was possible that the antimuscarinic activity
of this class of compounds might be masking stimulation
of locomotor activity or cocaine-like subjective effects.
Although, behavioral studies using classical muscarinic
antagonists, such as atropine or scopolamine, did not
support this notion, development of compounds without
the muscarinic component was desirable.26
In an attempt to eliminate the muscarinic binding
component for this class of dopamine uptake inhibitors,
a series of N-substituted 3R-[bis(4′-fluorophenyl)methoxy]-
tropane analogues were designed.29 In this study, a
separation of binding affinities for the dopamine trans-
porter versus muscarinic m1 receptors was achieved,
with the most potent and selective compound being
N-(4′′-phenyl-n-butyl)-3R-[bis(4′-fluorophenyl)methoxy]-
tropane (dopamine transporter, Ki ) 8.51 nM; m1, Ki )
576 nM).29
One class of dopamine transporter inhibitors that our
group has been investigating is based on 3R-(diphenyl-
methoxy)-1RH,5RH-tropane (benztropine; 3).19-21 Benz-
tropine binds equipotently to cocaine at the dopamine
transporter but demonstrates lower affinity than co-
caine for the other monoamine transporters.20,21 The
benztropines are intriguing since they contain a tropane
ring as seen in cocaine and a diphenylmethoxy moiety
found in the arylpiperazine class of dopamine uptake
inhibitors (e.g. GBR 12909; 4).21 We have previously
reported on several series of phenyl ring-substituted
analogues of benztropine,22-24 of which the most potent
compound was 3R-[bis(4′-fluorophenyl)methoxy]tropane
(5). Compound 5 demonstrated significantly higher
Since the 3R-[bis(4′-fluorophenyl)methoxy]tropanes do
not have a chiral center, enantioselectivity in this series
could not be determined. Profound enantioselectivity
has previously been demonstrated in the 2-carbomethoxy-
substituted 3R-[bis(4′-fluorophenyl)methoxy]tropane se-
ries of dopamine transporter ligands.30 Thus, our rea-
soning for synthesizing chiral N-substituted analogues
was that higher affinity and selectivity may be achieved
in this structurally similar group of compounds by
incorporating a chiral center in the N-substituent.
Furthermore, the synthesis of the present series of
chiral compounds, using commercially available chiral
amino acids, provides the advantage of far simpler and
higher yield chemistry than previously reported chiral
analogues.30 To this end, chiral N-substituted analogues
were designed to determine whether enantioselective
binding could be achieved and thus to potentially
improve selectivity of these compounds at the dopamine
transporter. Herein, the synthesis and evaluation for
binding at the dopamine transporter and muscarinic m1
receptors of several novel chiral N-substituted analogues
of 3R-[bis(4′-fluorophenyl)methoxy]tropane are described.
In addition, we also report the derivation of a highly
predictive CoMFA model based on the expanded series
of N-substituted 3R-[bis(4′-fluorophenyl)methoxy]tro-
pane analogues at the dopamine transporter.
affinity for the dopamine transporter than benztropine
(3) but retained high affinity for muscarinic receptors
(dopamine transporter, Ki ) 11.8 nM; m1, Ki ) 11.6 nM).
Interestingly, several of these 3R-(diphenylmethoxy)-
tropane analogues bound with high affinity to the
dopamine transporter and potently blocked dopamine
uptake but did not produce stimulation of locomotor
activity or cocaine-like discriminative stimulus ef-
fects.25,26 Structure-activity relationships derived from
this series of compounds revealed a significant diver-

N-Substituted 3R-[Bis(4′-fluorophenyl)methoxy]tropanes
J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 6 1087
Sch em e 1a
Ta ble 1. N-Substituted 3R-[Bis(4′-fluorophenyl)methoxy]-
tropanes Binding to the Dopamine Transporter (DAT) and to
Muscarinic (m1) Receptors
Ki, nM ((SEM)a
e
compd
N-substituent
DAT
m1
5
6
7
8a
8b
8c
methyl
4′′-phenyl-n-butyl
hydrogen
3′′-butene
naphthylmethyl
3′′′,4′′′-dichlorobenzyl
11.8 ( 1.3b 11.6 ( 0.9
8.51 ( 1.2c 575.5 ( 11
11.2 ( 1.2c
31.9 ( 2.8
367 ( 37
392 ( 35
203 ( 17
278 ( 19
1790 ( 49
3880 ( 370
10a (R)-2′′-amino-3′′-methyl-n-butyl 56.4 ( 9.6 2180 ( 86
10b (S)-2′′-amino-3′′-methyl-n-butyl 29.5 ( 3.5 4020 ( 590
10c 2′′-aminoethyl
10d (R)-2′′-phenyl-n-propyl
10e (S)-2′′-phenyl-n-propyl
10f
10g (S)-2′′-phenyl-n-butyl
10h 4′′-(4′′′-nitrophenyl)-n-butyl
11
12a n-butyl
12b allyl
12c cyclopropylmethyl
12d 3′′-phenyl-n-propyl
12e indol-3′′-ylethyl
13.9 ( 1.7 1250 ( 140
477 ( 53
301 ( 27
400 ( 52
228 ( 21
20.2 ( 2.2c
392 ( 15
1900 ( 48
1440 ( 180
2360 ( 300
299 ( 20
(R)-2′′-phenyl-n-butyl
4′′-(4′′′-aminophenyl)-n-butyl
29.7 ( 3.6d 134 ( 16
24.6 ( 1.9c
29.9 ( 2.9c
32.4 ( 2.9c
41.9 ( 4.6c
399 ( 28
177 ( 21
257 ( 29
312 ( 11
44.6 ( 4.9c 3280 ( 220
12f
2′′-{[(4′′′-nitrophenyl)phenyl]-
methoxy}ethyl
57.0 ( 9.7c
60.7 ( 7.3c
460 ( 51
686 ( 86
12g 3′′-(4′′′-fluorophenyl)-n-propyl
12h benzyl
12i
12j
82.2 ( 12c 1030 ( 150
cinnamyl
4′′′-fluorobenzyl
86.4 ( 10c
95.6 ( 9.6c 1540 ( 120
401 ( 22
a
Each Ki value represents data from at least three independent
b
experiments, each performed in triplicate. Data from ref 23.
c Data from ref 29. Data from ref 28. e Muscarinic data analyzed
d
with the computer program PRISM and differs from affinities
previously reported by Agoston et al.29 in which the program
LIGAND was used to analyze the data.
hydrate (HOBt). The corresponding amides 9a -g were
reduced with AlH3 or LiAlH4 (LAH) to give compounds
10a -g.
Molecu la r Mod elin g: CoMF A
The aim of the present study was to extend the SARs
of the N-substituted 3R-[bis(4′-fluorophenyl)methoxy]-
tropane analogues and to derive valid CoMFA models
for this series at the dopamine transporter and musca-
rinic m1 receptor. Table 1 lists the Ki values for the
compounds that were used in the CoMFA study.
Molecular modeling studies were performed using the
Sybyl (version 6.5)33 and MOPAC (version 6.0)34 soft-
ware packages installed on a Silicon Graphics Octane
workstation running IRIX 6.5. CoMFA was used for the
development of the 3D-QSAR model. The dopamine
transporter inhibitors considered in this work were
divided into two data sets: the ‘training set’ (Table 1
minus 10f, 12d , and 12g), on which the CoMFA analysis
was actually performed, and the ‘test set’ (10f, 12d , and
12g) used only for validating the predictive ability of
the 3D-QSAR model. Due to the relatively small number
of chiral analogues in the training set, only one chiral
analogue (10f) was chosen to be in the test set. Com-
pounds 12d and 12g were chosen to be in the test set
to see how well the model could predict activity for small
changes in ligand structure.
a
(a) RBr, K2CO3, DMF; (b) RCOOH, DCC, HOBt, Et3N, DMF;
(c) RCOCl, CHCl3, H2O, NaHCO3; (d) 4-(aminomethyl)piperidine,
CHCl3; (e) LiAlH4, THF; (f) AlH3, THF.
Ch em istr y
The compounds that were used in the present study
are listed in Table 1. The synthesis of the novel
analogues 8a -c and 10a -g is depicted in Scheme 1,
and Table 2 lists their chemical yields and physical
properties.
Achiral analogues 8a -c were prepared directly by
alkylation of 731 using the appropriate alkyl halide in
DMF with K2CO3.29 Amino acid analogues 9a -c were
prepared by modification of the method by Carpino et
al.32 The appropriate N-(9-fluorenylmethoxycarbonyl)
amino acid was converted to its acid chloride by reflux-
ing in thionyl chloride/dichloromethane. The resulting
amino acid chlorides were coupled with 7 in chloroform
and aqueous sodium bicarbonate. Deprotection with
4-(aminomethyl)piperidine gave the free amines 9a -c.
Chiral analogues 9d -g were prepared using an ami-
dation method, previously described,29 with 1,3-dicyclo-
hexylcarbodiimide (DCC) and 1-hydroxybenzotriazole
The entire set of N-substituted 3R-[bis(4′-fluorophen-
yl)methoxy]tropane analogues was drawn using the
SKETCH option in Sybyl 6.5 starting from the X-ray
coordinates of 6. Multiple conformations were drawn for
N-substituents that could adopt more than one orienta-

1088 J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 6
Ta ble 2. Physical Properties of Compounds 8a -c and 10a -ga
Robarge et al.
compd
recryst solv
mp, °C
formulab
OR: [R]23 (c, MeOH)
yield,d
%
D
8a
8b
8c
10a
10b
10c
10d
10e
10f
10g
ether/IPA
ether/MeOH
MeOH
ether/MeOH
ether/MeOH
ether/MeOH
amorphous solid
amorphous solid
EtOAc
181
C
C
C
C
C
C
C
C
C
C
24H27F2NO‚HCl
NAc
NA
NA
8.15 (1.08)
-9.61 (1.04)
NA
17.18 (1.03)
-17.29 (0.96)
16.05 (1.19)
-15.91 (1.30)
52
20
50
25
40
34
73
66
41
49
>220
>220
182-185
180-183
>230
31H30F2NO‚HBr
27H25Cl2F2NO‚HBr
25H32F2N2O‚2HBr
25H32F2N2O‚2HBr
22H26F2N2O‚2HBr
31H31F2NO‚fumarate‚H2O
31H31F2NO‚fumarate‚H2O
30H33F2NO‚HCl‚1/4H2O
30H33F2NO‚HCl‚1/4H2O
119-121
119-121
EtOAc
a
b
General procedures for the synthesis of 8a -c and 10a -g are given in the Experimental Section. All compounds were analyzed for
C,H,N and the results agreed to (0.4% with theoretical values. c NA, not applicable. from compound 7.
d
tion. This resulted in a number of compounds having
more than one possible conformer. The conformer which
gave the best q2 (cross-validated r2) value (highest) after
energy minimization, charge assignment, and alignment
was chosen for the final model. Geometries were opti-
mized and partial charges assigned using the AM1
model Hamiltonian35 as implemented in the MOPAC
program using PRECISE convergence criteria. It is
recognized that 6 may not adopt the X-ray coordinates
when bound to the dopamine transporter; however, the
use of a reasonable low-energy conformation as a
template is a practical starting point for statistical
comparisons of flexible structures within the Sybyl
CoMFA model.36 All compounds in the CoMFA study
were aligned via root-mean-square fitting of the non-
hydrogen atoms of the tropane ring (Figure 1).
A table was constructed of all the aligned compounds.
F igu r e 1. Alignment of the non-hydrogen atoms in the
tropane ring of all 25 compounds within the CoMFA study.
Logarithms of 1/inhibition of [3H]WIN 35,428 binding
(Ki) was added as the biological observable to obtain a
more evenly distributed dependent variable. The com-
puted log P (cLogP) values were also determined and
added to the table.
value of the model. During this procedure, each com-
pound in the ‘training set’ was systematically removed
and its activity predicted by a model derived from the
remaining compounds. Column filtering was set to 2
kcal/mol and the number of components to 10. The
number of components in the cross-validated run, which
corresponded to the highest q2 value that did not
increase by at least 5% from the addition of further
components, was chosen as optimal. The final PLS
model was derived using the optimal number of com-
ponents with no validation.
The CoMFA columns, which are descriptors of the
shape of the noncovalent fields surrounding the tested
molecules, were added. These analyses were performed
using steric (Lennard-J ones) and electrostatic (Cou-
lombic) fields either in equal combination or individu-
ally. Energy cutoff values of 30 kcal/mol were selected
for both steric and electrostatic fields. The CoMFA
region was created automatically and extended 4 Å
beyond every aligned molecule in the x-, y-, and z-
directions. The steric and electrostatic energy fields
were individually calculated by a sp3 carbon probe atom
with a charge of +1 located at the intersection of grid
points spaced at 2 Å within the CoMFA region.
To obtain a 3D-QSAR, partial least squares (PLS) was
used to correlate selected columns within the table. The
leave-one-out procedure was initially selected for cross-
validation of the PLS statistical method to determine
the optimal number of components and the predictive
Resu lts a n d Discu ssion
All compounds were tested for their displacement of
[3H]WIN 35,428 from the dopamine transporter in rat
caudate putamen. Muscarinic m1 binding was evaluated
by displacement of [3H]pirenzipine from rat brain ho-
mogenate. The results are shown in Table 1.
The structure-activity relationships for compounds
5-7, 10h , and 12a-j have been previously discussed.28,29
To summarize, a variety of N-substitutions providing

N-Substituted 3R-[Bis(4′-fluorophenyl)methoxy]tropanes
J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 6 1089
Ta ble 3. Summary of CoMFA Results
tertiary amines, including alkyl (e.g. methyl 5, n-butyl
12a , allyl 12b, and cyclopropylmethyl 12c) and arylalkyl
(phenylalkyl 6, 10h , 11, 12d ,g,h ,j, indol-3-ylethyl 12e,
and cinnamyl 12i), were well-tolerated at the dopamine
transporter. In general, for the phenylalkyl substitu-
ents, it was determined that increasing the length of
the alkyl tether between the tropane nitrogen and the
phenyl ring increased binding affinity. None of the
previously reported compounds bound with high affinity
to the other monoamine transporters. However, by
substituting other moieties for the methyl group of 5,
muscarinic affinity was significantly decreased while
generally retaining the other binding characteristics.
The most potent and selective compound in this series
was the N-(4-phenyl-n-butyl) analogue 6 (dopamine
transporter, Ki ) 8.51 nM; m1, Ki ) 576 nM).
optimal no. of components
4
q2
0.746
0.297
0.983
0.077
247.59
0
standard error of prediction
r2
standard error of estimate
F values
probability of R2 ) 0 (n1 ) 4, n2 ) 17)
contributions
cLogP
steric
0.236
0.764
A priority in the present study was to determine the
effect of chiral N-substitution of 3R-[bis(4′-fluorophenyl)-
methoxy]tropane analogues on binding to the dopamine
transporter and muscarinic m1 receptors. We found that
the valine-derived N-substituent, 10a and 10b, was
well-tolerated at the dopamine transporter. The com-
bination of structural features of 12a and 12h into chiral
analogues 10d -g resulted in loss of binding affinity at
the dopamine transporter. From this limited set of chiral
N-substituted analogues, it appears that the S-config-
uration is preferred slightly over the R-configuration.
For example, analogues 10b, 10e, and 10g are ap-
proximately 2-fold higher in potency at the dopamine
transporter than each of their R-configuration stereo-
isomers. Moreover, since the opposite enantioselectivity
is apparent at muscarinic m1 receptors, a further
separation in the binding affinities for the dopamine
transporter versus muscarinic m1 receptors was achieved.
Selectivity for the dopamine transporter was increased
from 68-fold in our previously most selective compound
6 to 136-fold in compound 10b (dopamine transporter,
Ki ) 29.5 nM; m1, Ki ) 4020 nM) while still retaining
good potency at the dopamine transporter.
F igu r e 2. Plot of predicted versus measured values from non-
cross-validated CoMFA model.
porter affinity in the cocaine and WIN series of tropane-
based compounds. When the nitrogen in the tropane
ring of a series of 2-carbomethoxy-3-(diarylmethoxy)-
tropane (difluoropine) analogues was replaced with
oxygen, a significant decrease in binding affinities
resulted.39 These findings support our results and
further substantiate a separation between SARs in
these two tropane-based classes of dopamine transporter
ligands.
The butenyl N-substituent 8a supported the previ-
ously reported SARs with a Ki ) 31.9 nM. However,
aromatic substituted compounds 8b and 8c showed a
substantial decrease in affinity (Ki ) 367 and 392 nM,
respectively). Amine-containing N-substituents 10a -c
bound with moderate to high affinity to the dopamine
transporter. Comparison of these analogues with their
non-amine-containing counterparts or with other N-
substituted analogues of similar size and shape indi-
cates that steric factors may be playing a greater role
than electronic factors in binding of these compounds
to the dopamine transporter. These findings were
substantiated further using the 3D-QSAR method of
analysis.
Analogues from our previous report29 that had an
N-substituent which rendered the nitrogen of the tro-
pane ring nonbasic were not included in the CoMFA
study because of their pernicious effects on the model.
These ligands did not bind appreciably to the dopamine
transporter (Ki > 2000 nM),29 and this result was
deemed likely due to the lack of a basic nitrogen,
although this finding is in contrast to recent reports by
Kozikowski,37 Madras,38 and Meltzer39 which convey
that a neutral tropane nitrogen or non-nitrogen bridge
does not have a detrimental effect on dopamine trans-
Several predictive models were derived using the
training set mentioned earlier. The most predictive
model resulted when only steric fields were considered
in the CoMFA. Addition of the cLogP values to the PLS
analysis resulted in an increased q2 value in every
model. The summary of the CoMFA results is shown in
Table 3. The optimal number of components was deter-
mined to be 4 by the selection criteria described above.
The q2 obtained was 0.746 with a standard error of
prediction of 0.297. The PLS run with no cross-valida-
tion yielded an R2 of 0.983 with a standard error of
estimate of 0.077. The relative contribution for sterics
is 0.764 versus 0.236 for cLogP. From this model, the
activity of each compound was predicted. These were
compared to the actual values (experimentally deter-
mined) in Figure 2. Compounds 10f, 12d , and 12g (test
set) were used to evaluate the predictive power of this
model. Table 4 shows the actual, predicted, and residual
values for the training set and the test set. This model
showed exceptional predictive ability with a residual
standard deviation in the test set of 0.05 log unit across
a range of 1.75 log units. Chiral analogue 10f was very
well-predicted with a residual value of 0.01. The ability
of the model to distinguish between small structural
alterations of the N-substituent is illustrated by the

1090 J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 6
Ta ble 4. CoMFA Actual, Predicted, and Residual Activitiesa
Robarge et al.
compd
actual
predicted
Training Set
residual
5
6
7
8a
8b
8c
7.93
8.07
7.95
7.50
6.44
6.41
7.25
7.53
7.85
6.32
6.52
6.64
7.69
7.53
7.61
7.52
7.49
7.35
7.24
7.09
7.06
7.02
7.82
8.04
7.98
7.44
6.49
6.40
7.25
7.59
7.81
6.35
6.50
6.68
7.63
7.68
7.65
7.61
7.46
7.26
7.22
7.05
7.12
6.94
0.11
0.03
-0.03
0.06
-0.05
0.01
0.00
-0.06
0.04
-0.03
0.02
-0.04
0.06
-0.15
-0.04
-0.09
0.03
0.09
0.02
10a
10b
10c
10d
10e
10g
10h
11
12a
12b
12c
12e
12f
12h
12i
12j
std dev
0.04
-0.06
0.08
0.06
Test Set
10f
12f
12g
6.40
7.38
7.22
6.39
7.28
7.17
0.01
0.10
0.05
std dev
a
Activities are expressed as log[1/Ki(M)].
F igu r e 3. CoMFA steric STDEV*COEFF contour plots from
the analysis model derived with no cross-validation. Compound
6 is shown inside the field. Green contours (level 75%)
surround regions where steric bulk favors binding. Yellow
contours (level 25%) surround regions were steric bulk disfa-
vors binding.
phenylpropyl analogue 12d and the p-fluorophenylpro-
pyl analogue 12g with residual values of 0.10 and 0.05,
respectively.
A graphical representation of the CoMFA model can
be seen in Figure 3. The CoMFA contours corresponding
to steric fields are shown together with compound 6 for
visualization purposes. The contour map suggests that
moving steric bulk away from the basic nitrogen of the
tropane (green areas) ring is advantageous for dopamine
transporter binding. Interestingly, steric bulk is disfa-
vored (yellow areas) surrounding the 2-position of the
N-substituent. This is evident from the chiral analogues
10d -g which have a phenyl group protruding from the
2-position.
Attempts to derive a valid 3D-QSAR model for mus-
carinic m1 binding were unsuccessful. There was no
apparent correlation of structure with activity in this
series of compounds. While the reasons for this are
unclear at this time, it may be that the alignment of
our test set is incorrect for the muscarinic m1 receptor
binding site. Further examination will be required to
determine the pharmacophore at the muscarinic m1
receptor, in this class of compounds, before proper
alignment may be accomplished. Also, our attempts at
eliminating affinity at the muscarinic m1 receptor may
have biased our set of compounds and resulted in a
nonpredictive model. Nonetheless, we have achieved our
goal of further improving selectivity for binding at the
dopamine transporter versus muscarinic receptors and
have determined that certain chiral N-substitutions in
this series of compounds provide separation.
receptor binding. The binding selectivity for the dopa-
mine transporter over the muscarinic m1 receptor was
improved with the chiral analogue 10b showing 136-
fold selectivity.
A CoMFA model has been derived which accounts for
the 3D-QSAR at the dopamine transporter. Good cor-
relations were achieved between steric fields with cLogP
and binding affinities. Previous CoMFA studies on
tropane-based ligands also suggest that the steric
component is the predominant factor in the binding
affinities of these analogues with electostatics playing
a smaller role.10-14 A comparison of these 3D-QSAR
models with our CoMFA models further substantiates
the hypothesis that the benztropine class of compounds
may be interacting with a different binding domain at
the dopamine transporter as compared to cocaine and
its analogues.17 Furthermore, the information obtained
from the present model, coupled with our previous
CoMFA model18 on phenyl ring-substituted benz-
tropines, will provide leads toward mapping the topology
of the dopamine transporter. Moreover, these novel
molecular probes, and the structural information they
provide, will undoubtedly contribute direction toward
the development of a pharmacotherapeutic for cocaine
abuse.
Exp er im en ta l Section
Con clu sion
All melting points were determined on a Thomas-Hoover
melting point apparatus and are uncorrected. The 1H NMR
spectra were recorded on a Bruker (Billerica, Mass) AC-300
instrument. Samples were dissolved in an appropriate deu-
Several new chiral N-substituted 3R-[bis(4′-fluoro-
phenyl)methoxy]tropane ligands have been synthesized
and tested for dopamine transporter and muscarinic m1

N-Substituted 3R-[Bis(4′-fluorophenyl)methoxy]tropanes
J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 6 1091
terated solvent (CDCl3 or CD3OD). Proton chemical shifts are
reported as parts per million (δ) relative to tetramethylsilane
(Me4Si, 0.00 ppm) which was used as an internal standard.
Infrared spectra were recorded as a neat film on NaCl plates
with a Perkin-Elmer 1600 series FTIR. Optical rotations were
obtained at the sodium D line on a J asco DIP-370 digital
polarimeter (100-mm cell). Microanalyses were performed by
Atlantic Microlab, Inc. (Norcross, GA) and agree within (0.4%
of calculated values. TLC solvent used was CHCl3/MeOH/NH4-
OH, 90:10:1, unless otherwise indicated. All chemicals and
reagents were purchased from Aldrich Chemical Co. or Lan-
caster Synthesis, Inc. unless otherwise indicated and used
without further purification.
was then allowed to warm to room temperature and stirred
for an additional 48 h. After completion of the reaction
(assessed by TLC), 15 mL of H2O was added. The reaction
mixture was basified by adding a few drops of concentrated
NH4OH to pH 9. The organic products were extracted with
ether (3 × 30 mL) and the combined ether fractions were
washed with H2O (2 × 25 mL), dried (Na2SO4), filtered, and
evaporated to an orange oil. Alane (AlH3) was made by careful
addition of 1.0 g 98% sulfuric acid in 2 mL anhydrous THF to
0.76 g (20 mmol) of LiAlH4 in 40 mL anhydrous THF at 0 °C
under an atmosphere of argon. After 15 min of stirring at room
temperature, a solution of the intermediate amide in 10 mL
of anhydrous THF was added dropwise to the reaction mixture
under argon. After 2 h of stirring at room temperature, the
reaction mixture was hydrolyzed by slow addition of a 1:1
mixture of THF and H2O at 0 °C. The gelatinous product was
dissolved in 60 mL of ether and 4 mL of 15% NaOH. After 20
min of stirring at room temperature, the mixture was filtered
over a pad of sodium sulfate. The organic filtrate was evapo-
rated to give the crude product as an oil which was purified
by flash chromatography using 3:1 hexane:ethyl acetate (1%
TEA). The oil was dissolved in a minimal volume of methanol
and acidified to pH 2 with a saturated solution of methanolic
HCl. Alternatively, the fumarate salt was produced by dis-
solving free base in acetone followed by addition of an
equimolar amount of fumaric acid. Evaporation of the solvent
and recrystallization gave pure product as the HCl or fumarate
salt (41-73% yield).
Syn th esis of N-Su bstitu ted 3r-[Bis(4′-flu or op h en yl)-
m eth oxy]tr op a n es 8a -c. Gen er a l Meth od . Compound 729
(727 mg, 2.5 mmol) was converted to its free base form by
extracting with CHCl3 (3 × 10 mL) from 20% NH4OH (20 mL),
drying, evaporating and then dissolving it in 20 mL dry DMF.
Anhydrous K2CO3 (376 mg, 2.72 mmol) and the appropriate
alkyl bromide (2.7 mmol) were added, and the reaction mixture
was allowed to stir at room temperature for 5 h. Inorganics
were removed by suction filtration, the filter pad was washed
with ether and the filtrate was poured into a separatory funnel.
Extraction from H2O (10 mL) with ether (3 × 10 mL) was
followed by washing the combined organic portions with H2O
(1 × 10 mL) and drying (Na2SO4). Evaporation of the volatiles
gave product as a clear oil. The crude free base was dissolved
in a minimal volume of methanol and acidified to pH 2 with a
saturated solution of methanolic HBr or HCl. Evaporation of
the solvent and recrystallization gave pure product as the HBr
or HCl salt (20-52% yield).
Syn th esis of N-Su bstitu ted 3r-[Bis(4′-flu or op h en yl)-
m eth oxy]tr op a n es 10a -c. Gen er a l Meth od . The appropri-
ate ((9-fluorenylmethyl)oxy)carbonyl (Fmoc) amino acid chlo-
ride was prepared according to Carpino et al.32 Compound 729
(350 mg, 1.1 mmol) was converted to its free base form by
extracting with CHCl3 (3 × 10 mL) from 20% NH4OH (20 mL),
drying, evaporating and then dissolving it in 10 mL amylene
stabilized chloroform and 10 mL of 10% sodium bicarbonate.
To this biphasic mixture was added the appropriate Fmoc-
protected amino acid chloride (1 mmol) dissolved in 10 mL of
amylene-stabilized chloroform. After 10 min of vigorous stir-
ring, the layers were separated and 0.5 mL of N-methylpip-
erazine was added, followed by immediate washing with 5%
HCl. To the organic layer was added 5 mL of 4-(aminomethyl)-
piperidine and after 40 min the organic phase was washed
with brine (2 × 20 mL), phosphate buffer (0.5 M, pH ) 5.5)
(4 × 20 mL), and brine (1 × 20 mL), dried (Na2SO4), filtered,
and solvent removed in vacuo. The crude amide was purified
by flash chromatography using 95:5:0.5 chloroform:methanol:
ammonium hydroxide (CMA). The amide was dissolved in 10
mL of anhydrous THF and added to a 30 mL suspension of
LAH (143 mg, 3.76 mmol) in THF at 0 °C. The reaction was
stirred at 0 °C for 5 min and then allowed to warm to room
temperature for 2 h. The reaction was quenched by slow
addition of a 1:1 mixture of THF and H2O at 0 °C. The
gelatinous product was dissolved in 60 mL of ether and 4 mL
of 15% NaOH. After 20 min of stirring at room temperature,
the mixture was filtered over a pad of sodium sulfate. The
organic filtrate was evaporated to give the crude product as
an oil which was purified by flash chromatography using 95:
5:0.5 CMA. The oil was dissolved in a minimal volume of
methanol and acidified to pH 2 with a saturated solution of
methanolic HBr. Evaporation of the solvent and recrystalli-
zation gave pure product as the HBr salt (25-40% yield).
Rep r esen ta tive sp ectr a l d a ta of 8c: 1H NMR (300 MHz,
CDCl3) δ 1.79-2.03 (m, 6H), 2.09-2.14 (m, 2H), 3.08 (br s,
2H), 3.56 (app t, J ) 3.45, 2.3 Hz, 1H), 5.36 (s, 1H), 6.99 (app
t, J ) 8.61, 8.69 Hz, 4H), 7.17-7.37 (m, 6H), 7.47 (d, J ) 1.7
Hz, 1H); 13C NMR (75 MHz, CDCl3) δ 26.1, 36.4, 55.6, 58.2,
69.5, 79.2, 115.0, 115.2, 127.6, 128.2, 128.3, 129.9, 130.1, 132.0,
138.5, 140.6, 160.2, 163.5; IR (neat film NaCl plate) 1501
(ROR), 1602 (aromatic), 2800-3000 (aliphatic stretch), 3000-
3200 (aromatic stretch) cm-1
.
Rep r esen ta tive sp ectr a l d a ta of 10a : [R]25D +8.15, CH3-
1
OH, c ) 1.08; H NMR (300 MHz, CDCl3) δ 0.89 (d, J ) 6.8
Hz, 6H), 1.4-1.8 (m, 8H), 2.05-2.08 (m, 2H), 2.4-2.6 (m, 2H),
3.1 (d, J ) 1.8 Hz, 2H), 3.5 (s, 1H), 5.36 (s, 1H), 6.96-7.02 (m,
4H), 7.24-7.29 (m, 4H); 13C NMR (75 MHz, CDCl3) δ 17.99,
19.18, 25.4, 26.8, 31.7, 36.2, 36.5, 54.8, 57.1, 57.7, 60.4, 69.4,
79.2, 114.9, 115.2, 128.2, 128.3, 137.9, 160.3, 163.5.
Rep r esen ta tive sp ectr a l d a ta of 10f: [R]25D +16.05, CH3-
1
OH, c ) 1.19; H NMR (300 MHz, CDCl3) δ 0.75 (t, J ) 7.35
Hz, 3H), 1.43-1.56 (m, 1H), 1.68-2.05 (m, 8H), 2.54 (br s, 3H),
3.08 (br s, 1H), 3.49 (s, 1H), 5.32 (s, 1H), 6.94-7.00 (app t,
J ) 8.35, 8.65 Hz, 4H), 7.16-7.30 (m, 9H); 13C NMR (75 MHz,
CDCl3) δ 11.6, 24.9, 25.8, 27.8, 34.6, 58.4, 59.0, 68.2, 79.6,
115.1, 115.3, 126.6, 127.7, 128.1, 128.2, 128.5, 138.0, 143.0,
160.3, 163.6.
Sin gle-Cr ysta l X-r a y An a lysis of 6. A clear rectangular
0.40 × 0.20 × 0.04 mm crystal recrystallized from acetone,
C
30H34OF2N+. Br, FW ) 542.49, was selected for data collec-
tion. Data were collected on a computer-controlled diffractom-
eter with an incident beam graphite monochromator (Siemens
P4 with Cu KR radiation, λ ) 1.54178 Å, T ) 295 K). A least-
squares refinement using 30 centered reflections within 14°
< 2θ < 56° gave the triclinic P1 cell, a ) 6.539(1), b ) 7.170-
(1), c ) 28.181(4) Å, with V ) 1311.5(3) Å3, Z ) 2, and dcalc
)
1.374 gm/cm3. A total of 4826 reflections were measured in
the θ/2θ mode to 2θmax ) 115°, of which there were 4673
independent reflections. Corrections were applied for Lorentz
and polarization effects. A face indexed numerical absorption
correction was applied, µ ) 2.439 mm1, and maximum and
minimum transmissions were 0.90 and 0.40, respectively. The
structure was solved by direct methods with the aid of the
program SHELXTL97.40 The full-matrix least-squares refine-
Syn th esis of N-Su bstitu ted 3r-[Bis(4′-flu or op h en yl)-
m eth oxy]tr op a n es 10d -g. Gen er a l Meth od . Compound 729
(663 mg, 2.0 mmol) and triethylamine (1.2 mL, 8.6 mmol) were
added to a mixture of the appropriate chiral acid (2.2 mmol),
dicyclohexylcarbodiimide (DCC; 453 mg, 2.2 mmol), and 1-hy-
droxybenzotriazole hydrate (HOBt; 297 mg, 2.2 mmol) in 50
mL dry DMF. The reaction mixture was allowed to stir for 1
h at 0 °C under an atmosphere of argon. The reaction mixture
2
ment on Fo varied 443 parameters including the coordinates
and anisotropic thermal parameters for all non-hydrogen
atoms. H atoms were included using a riding model [coordinate
shifts of C applied to attached H atoms, CH distances set to

1092 J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 6
Robarge et al.
0.96 Å, H angles idealized, Uiso(H) were set to 1.2Ueq(C) or, if
methyl, 1.5Ueq(C)]. The final R values for the 4072 observed
reflections with Fo > 4σFo were R1 ) 0.0728, and wR2 ) 0.22
for all data.41 The final difference Fourier excursions were 0.64
and -0.52 eÅ3.
The molecule was refined as a P1h but cannot have a center
due to molecular approaches of the fluorophenyl groups across
the inversion center. The disordered fluorophenyl groups are
at equal occupancy and can occupy the cell with a Z of 2 with
different parts of the disorder across the false center with
normal approaches. A refinement as a P1 is also disordered
for both molecules and is of poor geometry. The false center is
likely due to racemic twinning and stacking faults.
Assays were conducted in binding buffer (10 mM Tris-HCl,
5 mM MgCl2). The total volume in each tube was 0.5 mL and
the final concentration of membrane after all additions was
approximately 200-300 mg of protein/sample. [3H]Pirenzepine
(specific activity 73.9 Ci/mmol; from New England Nuclear,
Boston, MA; final concentration 3 nM) was added and the
incubation was continued for 1 h at 37 °C. The incubation was
terminated by the addition of 5 mL of ice-cold buffer (10 mM
Tris-HCl, pH 7.4) and rapid filtration through Whatman GF/B
glass fiber filter paper (presoaked in 0.5% polyethylenimine
in water to reduce nonspecific binding) using a Brandel cell
harvester (Gaithersburg, MD). The filters were washed with
two additional 5-mL washes and transferred to scintillation
vials. Absolute ethanol (0.5 mL) and Beckman Ready Value
scintillation cocktail (2.75 mL) were added to the vials that
were counted the next day at an efficiency of about 36%. Under
these assay conditions, an average experiment yielded ap-
proximately 15 000 dpm total binding per sample and ap-
proximately 900 dpm nonspecific binding, defined as binding
in the presence of 10 µM QNB (quinuclidinyl benzilate). Each
compound was tested with concentrations ranging from 0.01
nM to 100 µM for competition against binding of [3H]piren-
zepine, in at least three independent experiments, each
performed in triplicate. Displacement data were analyzed by
use of the nonlinear curve-fitting computer program PRISM.43
P h a r m a cology. 1. Dop a m in e Tr a n sp or t er Bin d in g
Assa y. Male Sprague-Dawley rats (200-250 g; Taconic,
Germantown, NY) were decapitated and their brains removed
to an ice-cooled dish for dissection of the caudate putamen.
The tissue was homogenized in 30 volumes ice-cold modified
Krebs-HEPES buffer (15 mM HEPES, 127 mM NaCl, 5 mM
KCl, 1.2 mM MgSO4, 2.5 mM CaCl2, 1.3 mM NaH2PO4, 10 mM
D-glucose, pH adjusted to 7.4) using a Brinkman polytron and
centrifuged at 20000g for 10 min at 4 °C. The resulting pellet
was then washed two more times by resuspension in ice-cold
buffer and centrifugation at 20000g for 10 min at 4 °C. Fresh
homogenates were used in all experiments. Binding assays
were conducted in modified Krebs-HEPES buffer on ice. The
total volume in each tube was 0.5 mL and the final concentra-
tion of membrane after all additions was 0.5% (w/v) corre-
sponding to 200-300 mg of protein/sample. Triplicate samples
of membrane suspension were preincubated for 5 min in the
presence or absence of the compound being tested. [3H]WIN
35,428 (2-â-carbomethoxy-3-â-(4-fluorophenyl)tropane 1,5-
naphthalenedisulfonate; specific activity 82.4 Ci/mmol; from
New England Nuclear, Boston, MA; final concentration 1.5
nM) was added and the incubation was continued for 1 h on
ice. The incubation was terminated by the addition of 3 mL of
ice-cold buffer and rapid filtration through Whatman GF/B
glass fiber filter paper (presoaked in 0.1% BSA in water to
reduce nonspecific binding) using a Brandel cell harvester
(Gaithersburg, MD). The filters were washed with three
additional 3-mL washes and transferred to scintillation vials.
Absolute ethanol (0.5 mL) and Beckman Ready Value scintil-
lation cocktail (2.75 mL) were added to the vials which were
counted the next day at an efficiency of about 36%. Under
these assay conditions, an average experiment yielded ap-
proximately 6000 dpm total binding per sample and ap-
proximately 250 dpm nonspecific binding, defined as binding
in the presence of 100 µM cocaine. Each compound was tested
with concentrations ranging from 0.01 nM to 100 µM for
competition against binding of [3H]WIN 35,428, in three
independent experiments, each performed in triplicate.
Ack n ow led gm en t. M.J .R. and G.E.A. were sup-
ported by a National Institutes of Health Intramural
Research Training Award fellowship. Single-crystal
X-ray analysis of 6 was funded in part by the Office of
Naval Research and NIDA Contract DA09045. We
thank Phyllis Terry and Richard Loeloff for expert
technical assistance. Animals used in this study were
maintained in facilities fully accredited by the American
Association for the Accreditation of Laboratory Animal
Care (AAALAC), and all experimentation was conducted
according to the guidelines for the Institutional Care
and Use Committee of the Intramural Research Pro-
gram, National Institute on Drug Abuse, NIH, and the
Guide for Care and Use of Laboratory Animals of the
Institute of Laboratory Animal Resources, National
Research Council, Department of Health, Education and
Welfare, Publication (NIH) 85-23, revised 1985.
Refer en ces
(1) Kuhar, M. J .; Ritz, M. C.; Boja, J . W. The Dopamine Hypothesis
of the Reinforcing Properties of Cocaine. Trends Neurosci. 1991,
14, 299-301.
(2) Ritz, M. C.; Lamb, R. J .; Goldberg, S. R.; Kuhar, M. J . Cocaine
Receptors on Dopamine Transporters are Related to Self-
Adiministration of Cocaine. Science 1987, 237, 1219-1223.
(3) Shimada, S.; Kitayama, S.; Lin, C.-L.; Patel, A.; Nanthakumar,
E.; Gregor, P.; Kuhar, M.; Uhl, G. Cloning and Expression of a
Cocaine-Sensitive Bovine Dopamine Transporter Complemen-
tary DNA. Science 1991, 254, 576-578.
In both saturation and competition experiments, two com-
ponents of [3H]WIN 35,428 binding were apparent. Analysis
of the data utilizing the LIGAND program revealed a high-
affinity component with a KD of 7 ( 5 nM and a Bmax of 445 (
338 fmol/mg protein and a low-affinity component with a KD
of 126 ( 115 nM and a Bmax of 1995 ( 559 fmol/mg protein.
Saturation and displacement data were analyzed by the use
of the nonlinear least-squares curve-fitting computer program
LIGAND.42 Data from replicate experiments were modeled
together to produce a set of parameter estimates and the
associated standard errors of these estimates. In each case,
the model reported fit significantly better than all others
according to the F test at p < 0.05. The Ki values reported are
the dissociation constants derived for the unlabeled ligands.
2. Mu sca r in ic m 1 Bin d in g Assa y. Whole frozen rat brains
excluding cerebellum (Taconic, Germantown, NY) were thawed
in ice-cold buffer (10 mM Tris-HCl, 320 mM sucrose, pH 7.4)
and homogenized with a Brinkman polytron in a volume of
10 mL/g of tissue. The homogenate was centrifuged at 1000g
for 10 min at 4 °C. The resulting supernatant was then
centrifuged at 10000g for 20 min at 4 °C. The resulting pellet
was resuspended in a volume of 5 mL/g in 10 mM Tris buffer
(pH 7.4).
(4) Kilty, J . E.; Lorang, D. B.; Amara, S. G. Cloning and Expression
of a Cocaine-Sensitive Rat Dopamine Transporter. Science 1991,
254, 578-579.
(5) Usdin, T. B.; Mezey, E.; Chen, C.; Brownstein, M. J .; Hoffman,
B. J . Cloning of the Cocaine-Sensitive Bovine Dopamine Trans-
porter. Proc. Natl. Acad. Sci. U.S.A. 1991, 88, 11168-11171.
(6) Giros, B.; Mestikawy, S. E.; Godinot, N.; Zheng, K.; Han, H.;
Yang-Feng, T.; Caron, M. G. Cloning, Pharmacological Charac-
terization, and Chromosome Assignment of the Human Dopa-
mine Transporter. Mol. Pharmacol. 1992, 42, 383-390.
(7) Carroll, F. I.; Lewin, A. H.; Kuhar, M. J . Dopamine Transporter
Uptake Blockers: Structure-Activity Relationships. In Neu-
rotransmitter Transporters: Structure, Function, and Regula-
tion; Reith, M. E. A., Ed.; Humana Press Inc.: Totowa, NJ , 1996;
pp 263-295.
(8) Newman, A. H. Novel Dopamine Transporter Ligands: The
State of the Art. Med. Chem. Res. 1998, 8, 1-11.
(9) Crammer, R. D.; Patterson, D. E.; Bunce, J . D. Comparative
Molecular Field Analysis (CoMFA). 1. Effect of Shape on Binding
of Steroids to Carrier Proteins. J . Am. Chem. Soc. 1988, 110,
5959-5967.

N-Substituted 3R-[Bis(4′-fluorophenyl)methoxy]tropanes
J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 6 1093
(10) Carroll, F. I.; Gao, Y.; Rahaman, A.; Abraham, P.; Parham, K.;
Lewin, A. H.; Boja, J . W.; Kuhar, M. J . Synthesis, Ligand
Binding, QSAR, and CoMFA Study of 3â-(p-Substituted phenyl)-
tropane-2â-carboxylic Acid Methyl Esters. J . Med. Chem. 1991,
34, 2719-2725.
(11) Carroll, F. I.; Mascarella, S. W.; Kuzemko, M. A.; Gao, Y.;
Abraham, P.; Lewin, A. H.; Boja, J . W.; Kuhar, M. J . Synthesis,
Ligand Binding, and QSAR (CoMFA and Classical) Study of 3â-
(3′-Substituted phenyl)-, 3â-(4′-Substituted phenyl)-, and 3â-
(3′,4′-Disubstituted phenyl)tropane-2â-carboxylic Acid Methyl
Esters. J . Med. Chem. 1994, 37, 2865-2873.
(12) Kotian, P.; Mascarella, S. W.; Abraham, P.; Lewin, A. H.; Boja,
J . W.; Kuhar, M. J .; Carroll, F. I. Synthesis, Ligand Binding,
and Quantitative Structure-Activity Relationship Study of 3â-
(4′-Substituted phenyl)-2â-heterocyclic Tropanes: Evidence for
an Electrostatic Interaction at the 2â-position. J . Med. Chem.
1996, 39, 2753-2763.
(13) Yang, B.; Wright, J .; Eldefrawi, M. E.; Pou, S.; MacKerell, A. D.
Conformation, Aqueous Solvation, and pKa Contributions to the
Binding Activity of Cocaine, WIN 35,065-2, and the WIN Vinyl
Analog. J . Am. Chem. Soc. 1994, 116, 8722-8732.
(14) Lieske, S. F.; Yang, B.; Eldefrawi, M. E.; MacKerell, A. D.;
Wright, J . (-)-3â-Substituted Ecgonine Methyl Esters as Inhibi-
tors for Cocaine Binding and Dopamine Uptake. J . Med. Chem.
1998, 41, 864-876.
(15) Vaughan, R. A. Photoaffinity-labeled Ligand Binding Domains
on Dopamine Transporters Indentified by Peptide Mapping. Mol.
Pharmacol. 1995, 47, 956-964.
(16) Vaughan, R. A.; Kuhar, M. J . Dopamine Transporter Ligand
Binding Domains. J . Biol. Chem. 1996, 271, 21672-21680.
(17) Vaughan, R. A.; Agoston, G. E.; Lever, J . R.; Newman, A. H.
Differential Binding Sites of Tropane-Based Photoaffinity Ligands
on the Dopamine Transporter. J . Neurosci. 1999, 19, 630-636.
(18) Newman, A. H.; Robarge, M. J .; Izenwasser, S.; Kline, R. H. A
Comparative Molecular Field Analysis (CoMFA) Study of Novel
Phenyl Ring-Substituted 3R-(Diphenylmethoxy) tropane Ana-
logues at the Dopamine Transporter. J . Med. Chem. 1999, 42,
3502-3507.
(19) Coyle, J . T.; Snyder, S. H. Antiparkinsonian Drugs: Inhibition
of Dopamine Uptake in the Corpus Striatum as a Possible
Mechanism of Action. Science 1969, 166, 899-901.
(20) van der Zee, P.; Hespe, W. A. Comparison of the Inhibitory
Effects of Aromatic Substituted Benzhydryl Esters on the
Uptake of Catecholamines and Serotonin into Synaptosomal
Preparations of the Rat Brain. Neuropharmacology 1978, 17,
483-490.
(21) van der Zee, P.; Koger, H. S.; Gootjes, J .; Hespe, W. A. Aryl 1,4-
Dialk(en)ylpiperazines as Selective and Very Potent Inhibitors
of Dopamine Uptake. Eur. J . Med. Chem. 1980, 15, 363-370.
(22) Newman, A. H.; Allen, A. C.; Izenwasser, S.; Katz, J . L. Novel
3R-(Diphenylmethoxy)tropane Analogues: Potent Dopamine
Uptake Inhibitors without Cocaine-like Behavioral Profiles. J .
Med. Chem. 1994, 37, 2258-2261.
(23) Newman, A. H.; Kline, R. H.; Allen, A. C.; Izenwasser, S.; George,
C.; Katz, J . L. Novel 4′-Substituted and 4′,4′′-Disubstituted 3R-
(Diphenylmethoxy)tropane Analogues as Potent and Selective
Dopamine Uptake Inhibitors. J . Med. Chem. 1995, 38, 3933-
3940.
(24) Kline, R. H.; Izenwasser, S.; Katz, J . L.; Newman, A. H.
3′-Chloro-3R-(diphenylmethoxy)tropane but not 4′-Chloro-3R-
(diphenylmethoxy)tropane Produces a Cocaine like Behavioral
Profile. J . Med. Chem. 1997, 40, 851-857.
(25) Katz, J . L.; Izenwasser, S.; Newman, A. H. Relations Between
Heterogeneity of Dopamine Transporter Binding and Function
and the Behavioral Pharmacology of Cocaine. Pharmacol. Bio-
chem. Behav. 1997, 57, 505-512.
(26) Katz, J . L.; Izenwasser, S.; Kline, R. H.; Allen, A. C.; Newman,
A. H. Novel 3R-Diphenylmethoxytropane Analogues: Selective
Dopamine Uptake Inhibitors with Behavioral Effects Distinct
from those of Cocaine. J . Pharmacol. Exp. Ther. 1999, 288, 302-
315.
(27) Newman, A. H.; Agoston, G. E. Novel Benztropine [3R-(Diphen-
ylmethoxy)tropane Analogues: Selective Ligands for the Dopa-
mine Transporter. Curr. Med. Chem. 1998, 5, 301-315.
(28) Agoston, G. E.; Vaughan, R.; Lever, J . R.; Izenwasser, S.; Terry,
P. D.; Newman, A. H. A Novel Photoaffinity Label for the
Dopamine Transporter Based on N-Substituted 3R-[bis(4′-fluo-
rophenyl)methoxy]tropane. Bioorg. Med. Chem. Lett. 1997, 7,
3027-3032.
(29) Agoston, G. E.; Wu, J . H.; Izenwassser, S.; George, C.; Katz, J .;
Kline, R. H.; Newman, A. H. Novel N-Substituted 3R-[Bis(4′-
fluorophenyl)methoxy]tropane Analogues: Selective Ligands for
the Dopamine Transporter. J . Med. Chem. 1997, 40, 4329-4339.
(30) Meltzer, P. C.; Liang, A. Y.; Madras, B. K. The Discovery of an
Unusually Selective and Novel Cocaine Analog: Difluoropine.
Synthesis and Inhibition of Binding at Cocaine Recognition Sites.
J . Med. Chem. 1994, 37, 2001-2010.
(31) Olofson, R. A.; Martz, J . T.; Senet, J .-P.; Piteau, M.; Malfroot,
T. A. A New Reagent for the Selective, High-Yield N-Dealkyla-
tion of Tertiary Amines: Improved Synthesis of Naltrexone and
Nalbuphine. J . Org. Chem. 1984, 49, 2081-2082.
(32) Carpino, L. A.; Cohen, B. J .; Stephens, K. E., J r.; Sadat-Aalace,
S. Y.; Tien, J .-H.; Langridge, D. C. ((9-Fluorenylmethyl)oxy)-
carbonyl (Fmoc) Amino Acid Chlorides. Synthesis, Characteriza-
tion, and Application to the Rapid Synthesis of Short Peptide
Segments. J . Org. Chem. 1986, 51, 3732-3734.
(33) Sybyl (Version 6.5), Tripos Inc., 1699 South Hanley Rd., St.
Louis, MO 63144; 1998.
(34) MOPAC (Version 6.5) [QCPE No. 455], Quantum Chemistry
Exchange Program, Indiana University, Creative Arts Bldg. 181,
Bloomington, IN 47405; 1990.
(35) Dewar, M. J . S. E.; Zoebisch, G.; Healy, E. F. AM1: A New
General Purpose Quantum Mechanical Molecular Model. J . Am.
Chem. Soc. 1985, 107, 3902-3909.
(36) Crammer, R. D., III; DePreist, S. A.; Patterson, D. E.; Hecht, P.
The Developing Practice of Comparative Molecular Field Analy-
sis. In 3D QSAR in Drug Design-Theory, Methods and Applica-
tions; Kubinyi, H., Ed.; ESCOM: Leiden, 1993; pp 443-485.
(37) Kozikowski, A. P.; Saiah, M. K. E.; Bergmann, J . S.; J ohnson,
K. M. Structure-Activity Relationship Studies of N-Sulfonyl
Analogues of Cocaine: Role of Ionic Interaction of Cocaine
Binding. J . Med. Chem. 1994, 37, 3440-3442.
(38) Madras, B. K.; Pristupa, Z. B.; Niznik, H. B.; Liang, A. Y.;
Blundell, P.; Gonzalez, M. D.; Meltzer, P. C. Nitrogen-Based
Drugs are not Essential for Blockade of Monoamine Transport-
ers. Synapse 1996, 24, 340-348.
(39) Meltzer, P. C.; Liang, A. Y.; Blundell, P.; Gonzalez, M. D.; Chen,
Z.; George, C.; Madras, B. K. 2-Carbomethoxy-3-aryl-8-oxabicyclo-
[3.2.1]octanes: Potent Non-Nitrogen Inhibitors of Monoamine
Transporters. J . Med. Chem. 1997, 40, 2661-2673.
(40) Sheldrick, G. M. SHELXTL97: Program for the Refinement of
Crystal Structures; University of Go¨ttingen: Go¨ttingen, Ger-
many, 1997.
(41) (a) R1 ) ∑||Fo| - |Fc||/∑|Fo|. (b) wR2 ) [∑[w(Fo - Fc2)2]/∑-
2
[w(Fo2)2]]1/2
.
(42) Snedecor, G. W.; Cochran, W. G. In Statistical Methods, 6th ed.;
Iowa State University Press: Ames, IA, 1967; pp 135-171.
(43) PRISM (Version 2.01), Graph Pad Software, 10855 Sorrento
Valley Rd. #203, San Diego, CA 92121; 1996.
J M990265S