3D QSAR study, synthesis, and in vitro evaluation of (+)-5-FBVM as potential PET radioligand for the vesicular acetylcholine transporter (VAChT)

Article abstract of DOI:10.1016/j.bmc.2010.08.028

Located in presynaptic cholinergic nerve terminals, the vesicular acetylcholine transporter (VAChT) represents a potential target for quantitative visualization of early degeneration of cholinergic neurons in Alzheimer's disease using PET. Benzovesamicol

Full text of DOI:10.1016/j.bmc.2010.08.028

Bioorganic & Medicinal Chemistry 18 (2010) 7659–7667
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
3D QSAR study, synthesis, and in vitro evaluation of (+)-5-FBVM as potential
PET radioligand for the vesicular acetylcholine transporter (VAChT)
b
a
b
Mitja Kovac a, Sylvie Mavel a, , Winnie Deuther-Conrad , Nathalie Méheux , Jana Glöckner ,
*
Barbara Wenzel b, Marko Anderluh c, Peter Brust b, Denis Guilloteau a, Patrick Emond a
a Université François-Rabelais de Tours, INSERM U930, CHRU, Hôpital Bretonneau, Service de Médecine Nucléaire, 37000 Tours, France
b Forschungszentrum Dresden-Rossendorf, Research Site Leipzig, Institute of Radiopharmacy, Permoserstr. 15, 04318 Leipzig, Germany
c
ˇ
University of Ljubljana, Faculty of Pharmacy, Department of Pharmaceutical Chemistry, Aškerceva 7, 1000 Ljubljana, Slovenia
a r t i c l e i n f o
a b s t r a c t
Article history:
Located in presynaptic cholinergic nerve terminals, the vesicular acetylcholine transporter (VAChT) rep-
resents a potential target for quantitative visualization of early degeneration of cholinergic neurons in
Alzheimer’s disease using PET. Benzovesamicol derivatives are proposed as radioligands for this purpose.
We report QSAR studies of vesamicol and benzovesamicol derivatives taking into account the stereose-
lectivity of the VAChT binding site. Use of different data sets and different models in this study revealed
that both enantiomers of 5-fluoro-3-(4-phenyl-piperidin-1-yl)-1,2,3,4-tetrahydro-naphthalen-2-ol
(5-FBVM) are promising candidates, with predicted VAChT affinities between 6.1 and 0.05 nM. The syn-
thesis of enantiopure (R,R)- and (S,S)-5-FBVM and their corresponding triazene precursors for future
radiofluorination is reported. Both enantiomers exhibited high in vitro affinity for VAChT [(+)-5-FBVM:
Ki = 6.95 nM and (À)-5-FBVM: Ki = 3.68 nM] and were selective for r2 receptors (ꢀ70-fold), only (+)-5-
Received 18 May 2010
Revised 26 May 2010
Accepted 12 August 2010
Available online 17 August 2010
Keywords:
Benzovesamicol derivative
VAChT
Triazene
Fluoro-dediazoniation
3D QSAR
FBVM is selective for
r
1 receptors (ꢀfivefold). These initial results suggest that (+)-(S,S)-5-FBVM warrants
further investigation as a potential radioligand for in vivo PET imaging of cholinergic nerve terminals.
Ó 2010 Elsevier Ltd. All rights reserved.
1. Introduction
ing by PET have focused on compounds such as benzovesamicols,3–6
trozamicols,7,8 or morpholino vesamicols,9,10 all based on the struc-
The degeneration of cholinergic neurons in the brain is one of the
most significant neuropathological features in Alzheimer’s disease
(AD) synapse disorder has been shown to precede the neurofibrillary
and neuritic aspects of AD during the course of the disease,1 and the
loss of synaptic terminals correlates better with cognitive decline
than extracellular plaque load or loss of neurons.2 Located in presyn-
aptic cholinergic nerve terminals, the vesicular acetylcholine trans-
porter (VAChT) is postulated to be a valuable target for in vivo
quantitative visualization of early neurodegenerative processes in
AD by using molecular imaging techniques such as SPECT (Single
Photon Emission Computed Tomography) and PET (Positron Emis-
sion Tomography), PET being regarded as superior in terms of detec-
tion efficiency, spatial resolution, and quantification. However,
VAChT-specific tracer compounds labeled with the short-lived PET
radionuclides 18F and 11C have not been developed for clinical use
to date. Approaches to the development of tracers for VAChT imag-
ture of (À)-vesamicol, a drug that binds to a side allosteric to the act-
eylcholine transport side with low-nanomolar affinity.11–14
However, (À)-vesamicol possesses moderate affinity to
ceptors and nanomolar affinity to
-receptors.15 In combination
with similar expression patterns of VAChT16 and 1-receptors17 in
a-adreno-
r
r
the brain the latter cause marginal signal-to-background ratios in
in vivo imaging by PET, one reason why so many VAChT radioligands
have failed in pre-clinical evaluation to date. On the other hand, the
successful application of the benzovesamicol-related SPECT ligand
(À)-5-[123I]IBVM (5-iodo-3-(4-phenyl-piperidin-1-yl)-1,2,3,4-tet-
rahydro-naphthalen-2-ol, Kd = 0.30 nM)18–21 has encouraged fur-
ther research into the design of VAChT-specific PET tracers.
Structure-affinity studies assessing the potential of iodobenzove-
samicol derivatives for visualization of cholinergic terminals have
revealed that there is considerable bulk tolerance at different
structural positions of the benzovesamicol molecule.18,22
Therefore we have recently synthesized and evaluated new
benzovesamicol derivatives21 and aza-analogs of trozamicol deriv-
atives both in vitro and in vivo in animals models.23–25 Several 18F-
labeled benzovesamicol derivatives have now been synthesized
* Corresponding author. Address: Faculté de Pharmacie, Laboratoire de Biophy-
sique Médicale et Pharmaceutique, 31 avenue Monge, 37200 Tours, France. Tel.: +33
2 47 36 72 40; fax: +33 2 47 36 72 24.
including
[ [ [ [
18F]NEFA, 18F]FAA,6 18F]FEOBV,5 and 18F]FPOBV3
(Table 1). However, except for [18F]FEOBV, which is currently
E-mail address: sylvie.mavel@univ-tours.fr (S. Mavel).
0968-0896/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2010.08.028

7660
M. Kovac et al. / Bioorg. Med. Chem. 18 (2010) 7659–7667
Table 1
Vesamicol analogs I and benzovesamicols II used for the QSAR studies
HO
HO
9
8
A
N
C
11
N
13
R1
R3
R
D
R2
I
II
5
R
I
Refa
21
R1
H
R2
R3
H
II
Refa
21
H
(À)-vesamicol
(+)-vesamicol
(À)-oMV
(+)-oMV
5-I
(À)-5-IBVM
(+)-5-IBVM
(À)-NEFA
o-CH3
p-CH3
o-I
30
H
H
H
H
H
H
5-N(Et)COCH2F
5-O(CH2)2F
5-(CH2)3F
H
H
H
H
H
50
(À)-pMV
(+)-pMV
(À)-oIV
30
(À)-FEOBV
(+)-FEOBV
(À)-FPOBV
(+)-FPOBV
(À)-AOIBV
(+)-AOIBV
(À)-MOIBV
(+)-MOIBV
(À)-MAIBV
(+)-MAIBV
(À)-6-IBVM
(À)-7-IBVM
(À)-8-IBVM
(À)-pI-BVM
(À)-mI-BVM
(À)-ABV
52
21
21
21
21
(+)-oIV
52
52
m-I
(À)-mIV
5-OCH2CHCHI
5-I and 8-OCH3
5-I
(+)-mIV
p-I
(À)-pIV
CH2NH2
H
H
H
p-I
m-I
H
6-I
7-I
8-I
H
H
5-NH2
H
H
H
H
H
H
22
22
22
22
53
50
50
(+)-ABV
(À)-BVM
H
H
H
a
The data of binding affinities used for QSAR study were taken from these references.
under investigation in PET studies,4 these PET tracers have not
been suitable for in vivo applications, possibly due to the metabolic
susceptibility of the F-carrying substituents, for example, amide in
NEFA and alcoholate in FEOBV. In previous paper, FPOBV instability
in vivo,3 possibly arising from the weak strength of the C(sp3)–F
bond (0.00079 kcal/mol, DiscoveryStudio2.5Ò, Accelrys Inc.) caus-
ing de-fluorination is reported. To increase the metabolic stability
and to reinforce the C–F bond strength by a C(sp2)–F bond
(0.0628 kcal/mol, DiscoveryStudio2.5Ò, Accelrys Inc.), we have
developed the novel benzovesamicol analog 5-FBVM as a fluoro
analog of 5-IBVM (Table 1).
synthesis and provided the basis for the synthesis of enantiopure
(+)-(S,S) and (À)-(R,R)-5-FBVM as well as (S,S) and (R,R)-5-TBV.
2. Results and discussion
2.1. QSAR study
The QSAR study is based on 32 derivatives which belong to the
classes of vesamicols (I) and benzovesamicols (II) (Table 1). Six
derivatives were synthesized and evaluated regarding VAChT affin-
ity and specificity by in-house in vitro assays, the Ki values of the
remaining compounds were taken from literature. Since binding
to the VAChT is known to be highly enantioselective (generally,
the in vitro affinity for the VAChT of (À)-enantiomers is about 10
times greater) we made three sets from the same overlay: Set 1
for all 32 derivatives, Set 2 for the (À)-enantiomers (20 deriva-
tives), and Set 3 for the (+)-enantiomers (12 derivatives). All the
structures were minimized under a CHARMm forcefield with a root
mean squared (RMS) difference of energy gradient reached
0.1 kcal/mol Å (Discovery StudioÒ 2.5, Accelrys Inc., San Diego,
CA). According to crystallographic findings on vesamicol deriva-
tives10,27 and ABV (data not shown, structure see Table 1), the
piperidine ring is in a chair conformation, and is almost perpendic-
ular to the cyclohexanol ring, with the hydroxyl in trans position as
an equatorial conformer. The carbons C-9, C-11, and C-13 of piper-
idine (see Table 1) and C-OH were chosen for the overlay. Ki values
were taken from the literature, and pKi = Àlog Ki was used for the
Genetic Function Approximation (GFA) algorithm.28 The QSAR pro-
gram presents 179 physicochemical descriptors as different elec-
tronic, spatial, shadow, shape, and thermodynamic indices. After
several filters (analyzing the correlation matrix, eliminating the
highly correlated descriptors, and eliminating descriptors with
too wide a range of training data (>700)), only 26 2D or 3D
In 2008, Szymoszek et al.26 have published a first Comparative
Molecular Field Analysis (CoMFA) study using a partial least
squares (PLS) algorithm for a set of 37 vesamicol derivatives, cov-
ering three different structural types, to predict the binding affinity
of vesamicol-type ligands from their respective molecular struc-
ture. To expand these efforts, we have performed a further 3D
QSAR study, which is based on 32 vesamicol and benzovesamicol
derivatives (Table 1) in order to predict the binding affinity for
the new compound 5-FBVM. Furthermore, this study considered
for the first time the stereoselectivity of the binding of vesamicol
derivatives to the VAChT protein.
Moreover, we also wished to improve the accessibility of the
radiolabeling procedure by radiofluorination of non-activated aro-
matic cycles (i.e., without an electron withdrawing group in ortho
or para position to the leaving group). Therefore, in this report we
describe not only the synthesis of 5-FBVM via fluoro-dediazonia-
tion based on the secondary amine 5-ABV, but also the synthesis
of a suitable triazene precursor (5-TBV) for future 18F-labeling. To
demonstrate the suitability of triazene as leaving group, the non-
radioactive fluorination of 5-TBV resulting in 5-FBVM was
accomplished. Because of the stereoselective binding of vesamicol
derivatives, 5-ABV was enantioseparated via Mosher ester

M. Kovac et al. / Bioorg. Med. Chem. 18 (2010) 7659–7667
7661
descriptors were retained as dependent on pKi. The definitions of
2.1.3. Set 3
N = 12: (+)-enantiomers.
mean descriptors are given in Table 2. In the GFA model, linear
or quadratic terms of descriptors were allowed in the selection.
For the statistical parameters, the cross validated r2 is not the most
important, and the Friedman’s lack-of-fit (LOF) score, which evalu-
ates the QSAR model by considering the number of descriptors as
well as the quality of fitness, is chosen: the lower the LOF, the less
likely it is that the GFA model will fit the data. The significant
regression is given by F, and the higher the value, the better the
model. The standard errors of regression coefficients are given in
parentheses. GFA models were tested with 5-FBVM, with the excel-
lent predictive sub-nanomolar Ki.
Linear model form: pKi(+)pred= 27.60( 0.02) + 3.42( 0.003) Â
CHI_2 À 1.73( 0.001) Â CHI_V_3_PÀ 0.059( 0) Â E_ADJ_mag + 19.66
( 0.05) Â JX À 30.19( 0.05) Â JY À 0.654( 0.0003) Â Dipole_Z À 9.64
( 0.01) Â Jurs_RPSA À 3.73( 0.007) Â Shadow_XYfrac + 3.67( 0.002)
 Bond Energy À 0.067( 0)  Dihedral Energy.
r
2 = 1.0000 r2 (adj) = r2 (pred) = RMS residual
1.0000 1.0000
error = 0.0004352 L.O.F. = 1.844eÀ006
Friedman
Predicted affinity: pKi(+)-5-FBVMpred = 8.22.
2.1.1. Set 1
The major contributing factors are JY > JX > Jurs_RPSA > Sha-
dow_XYfrac ꢁ Bond Energy ꢁ CHI_2 > CHI_V_3_P > Dipole_Z > Di-
hedral Energy ꢁ E_ADJ_mag.
N = 32: (+) and (À)-enantiomers.
Linear model form: pKi(+)/(À)pred = 79.61( 10.49) À 39.86
( 5.82) Â JY + 0.99( 0.22) Â Dipole_Z + 0.0185( 0.004) Â Jurs_PN-
SA_2 À 4.61( 1.45) Â Shadow_nu + 3.28( 1.07) Â Bond Energy À
0.894( 0.348) Â Van der Waals Energy.
The best predictive values for 5-FBVM were obtained with the
linear model rather than the quadratic model. Furthermore, consid-
ering the (À)-enantioselectivity of VAChT binding site, we focused
on (+)/(À) and (À) linear models. For set 1 ((+)/(À) model), the main
descriptor was JY (Balaban index which characterizes the shape of a
molecule taking into account the relative covalent radius of the
atoms of the model) with a mean value for the 32 compounds of
1.50. As the coefficient in the equation was negative, the smaller
the value, the better the affinity for the VAChT (JY(À)-5-FBVM = 1.460
r
2 = 0.7425 r2 (adj) = 0.6807 r2 (pred) = 0.6101 RMS residual Friedman
error = 0.7419 L.O.F. = 1.155
Predicted affinity: pKi(+)-5-FBVMpred = 8.91 pKi(À)-5-FBVMpred
=
9.34.
The major contributing factors are JY > Shadow_nu > Bond En-
ergy > Van der Waals Energy > Dipole_Z > Jurs_PNSA_2.
Quadratic model form: pKi(+)/(À)pred = 71.22( 9.53) À 32.92
( 4.57) Â JY + 1.03( 0.21) Â Dipole_Z + 0.0160( 0.003) Â Jurs_PNSA_2
À 4.10( 1.28)  Shadow_nu À 0.73( 0.23)  Bond Energy  Van
der Waals Energy.
compared to JY
= 1.447 and JY(À)-vesamicol = 1.579). For set 2 ((À)
(À)-IBVM
model), the main descriptor was Shadow_XZfrac (a steric descrip-
tor) with a mean value for the 20 compounds of 0.606. The negative
coefficient of shadow in the XZ plane indicated that a decrease in
the area of the molecular shadow in the XZ plane (Shadow_XZfrac
(À)-5-FBVM = 0.577 compared to Shadow_XZfrac(À)-IBVM = 0.581 and
Shadow_XZfrac(À)-vesamicol = 0.611) was favorable for affinity for
the VAChT. The equations showed multiple occurrence of Jurs
descriptors,29 suggesting the importance of charge distribution
and surface areas, in particular for the Jurs_RPSA descriptor (total
polar surface area divided by the total molecular solvent-accessible
surface area) with a mean value for the 20 compounds of 0.0789.
Since Jurs_RPSA depends on polar surface, fluorine derivatives
had the highest values and the coefficient in the equation was
negative. For 5-FBVM the descriptor was the smallest of the
fluorine derivatives (Jurs_RPSA(À)-5-FBVM = 0.128 compared to
Jurs_RPSA(À)-NEFA = 0.152 and Jurs_RPSA(À)-FEOBV = 0.130).
r
2 = 0.7469
r
2 (adj) = 0.6982
r
2 (pred) = 0.6454 RMS residual Friedman
error = 0.7213 L.O.F. = 1.028
Predicted affinity: pKi(+)-5-FBVMpred = 9.21 pKi(À)-5-FBVMpred
=
9.75.
2.1.2. Set 2
N = 20: (À)-enantiomers.
Linear
model
form:
pK
i(À)pred = 14.06( 3.67) + 0.0759
( 0.0120) Â Molecular_SASA À 0.0766( 0.011) Â E_ADJ_mag + 0.9205
( 0.0834) Â Jurs_PPSA_3 À 38.88( 7.73) Â Jurs_RPSA + 2.89( 1.13) Â
RadofGyration À 5.20( 0.76) Â Shadow_nu À 66.45( 10.20)Â
Shadow_XZfrac + 27.27( 4.31) Â Shadow_YZfrac À 0.335( 0.051) Â
Dihedral Energy.
The 3D QSAR model (Discovery StudioÒ 2.5, Accelrys), defines
the critical regions (steric or electrostatic) affecting binding affin-
ity. A contour plot of the electrostatic field region favorable
(in blue) or unfavorable (red) for the VAChT affinity is shown in
Figure 1 with the superposition of both stereoisomers. As reported
by the 3D QSAR modeling shown that substitutions near the 5-po-
sition should increase the affinity for VAChT (as it was previously
suggested by Szymoszek et al.26) as well as an electropositive sub-
stituent (near the 8-position) as shown in Figure 2. A good ligand
should have strong Van der Waals attraction in the green area
and a polar group in the blue electrostatic potential area. Figure
3 shows the differences obtained for the 3D QSAR model with po-
sitive coefficients (in green) on a Van der Waals grid between the
(À)-enantiomers model and the (+)-enantiomers model. Moreover
we observed an overlap of a negative yellow area near the 5-posi-
tion for (+)-enantiomer FBVM in the favorable positive green area
of (À)-enantiomers model, this could explain partially the stereo-
specificity of VAChT binding site. With this good 3D model
(N = 32) pKi(À)-5-FBVM = 9.72 and pKi(+)-5-FBVM = 8.24 with r = 0.922
and r2 = 0.851 could be predicted, corresponding to the experimen-
tal finding reported in Table 3 (pKi = 8.43 and 8.16, respectively).
A third QSAR study was performed based on Bayesian modeling
(Discovery StudioÒ 2.5, Accelrys) which distinguishes ‘active’
r
2 = 0.9591 r2 (adj) = 0.9223 r2 (pred) = 0.8523 RMS residual Friedman
error = 0.3716 L.O.F. = 0.6274
Predicted affinity: pKi(À)-5-FBVMpred = 8.64.
The major contributing factors are Shadow_XZfrac > Jurs_
RPSA > Shadow_YZfrac > Shadow_nu > RadofGyration ꢁ Jurs_PP-
SA_3 > Dihedral Energy > Molecular_SASA ꢁ E_ADJ_mag.
Quadratic
model
form:
pKi(À)pred = 26.21( 3.37) À 5.32
( 0.85)JY Â Shadow_nu À 0.0031( 0.0006)Dipole_Z Â Jurs_PNSA_2
À 0.0053( 0.001) Jurs_DPSA_3 Â Dihedral Energy + 0.145( 0.02)
Jurs_PPSA_3 Â Shadow_nu.
r
2 = 0.8346 r2 (adj) = 0.7904 r2 (pred) = 0.7359 RMS residual Friedman
error = 0.6105 L.O.F. = 0.7662
Predicted affinity: pKi(À)-5-FBVMpred = 10.26.

7662
M. Kovac et al. / Bioorg. Med. Chem. 18 (2010) 7659–7667
Table 2
Description of the molecular properties used as descriptors in QSAR studies
Property
Description
Molecular_SASA
CHI-2
2D surface area
2D topological
descriptors
2D topological
descriptors
2D topological
descriptors
2D topological
descriptors
2D topological
descriptors
Spatial
Spatial
Spatial
Spatial
3D electronic
descriptors
Total solvent accessible surface area
Unmodified molecular connectivity indices. This type of emphasizes different aspects of atom connectivity within a
molecule—the amount of branching, ring, structures present and flexibility
CHI-V_3_P
E_ADJ_mag
JX
JY
Highly discriminating descriptor. This Balaban indices characterize the shape of a molecule taking into account
electronegativity of the atoms of the model
Highly discriminating descriptor. This Balaban indices characterize the shape of a molecule taking into account relative
covalent radius of the atoms of the model
Projection of molecular surface on a plan (XZ)
Projection of molecular surface on a plan (XY)
Projection of molecular surface on a plan (YZ)
Ratio of largest to smallest dimension of projection of molecular surface on a plan
Dipole moment
Shadow XZ frac
Shadow XY frac
Shadow YZ frac
Shadow_nu
Dipole Z
Jurs_PNSA_2
Jurs_PPSA_3
Jurs_RPSA
Total charge weighted negative surface area
Atomic charge weighted positive surface area
Relative polar surface area: total polar surface area divided by the total molecular solvent-accessible surface area
RadofGyration
3D molecular
properties
Molecular
properties
Molecular
properties
Molecular
properties
Molecular
properties
Molecular
properties
Angle energy
Bond energy
Conformational
Conformational
Electrostatic
energy
Dihedral energy
Van der Waals
energy
Conformational
Figure 1. Isosurface of the 3D QSAR model coefficients on Electrostatic Potential
grids with positive (in blue) and negative (in red) coefficients for the aligned
molecular structures of 32 (+)/(À)-enantiomers I and II in solid representation (A, C,
and D visualize cycles as precise in Table 1).
Figure 2. Isosurface of the 3D QSAR model positive coefficients on Van der Waals
grid (in green) and Electrostatic Potential grid (in blue) for (+)/(À)-enantiomer
model in solid representation with (À)-5-FEOBV.
ligands from baseline ligands. The Bayesian statistics assign the
probability for each individual descriptor of a molecule to be a
member of an ‘active’ class. From the data set of 32 known ligands
(Table 1), 81% were defined as ‘active’ or ‘inactive’, with a cutoff va-
lue at a pKi = 7.69 (Ki <25 nM). From this effective model we tested
(À) and (+)-5-FBVM to predict their activity. The results did not
distinguish these two enantiomers, they were both predicted as
‘active’ with a probability of 82%. From the listing scores obtained
in this Bayesian model, the closest compound of FBVM was
(À)oMV, which presented a Ki = 6.7 nM,30 that is, was close to the
experimental findings (Ki(À)-5-FBVM = 3.7 nM and Ki(+)-5-FBVM
=
6.9 nM, Table 3).

M. Kovac et al. / Bioorg. Med. Chem. 18 (2010) 7659–7667
7663
Table 3
Affinities and selectivity of FEOBV, (rac)-5-IBVM, (rac)-5-ABV, and 5-FBVM for rVAChT, r1 and r2 (Ki values are given in means SD)
rVAChT
Ki values, in nM
Ratio of Ki values
r1
r2
r
1/VAChT
r
2/VAChT
(À)-FEOBV
19.6 1.1a
209 94a
269 37a
266 113a
652 60
13.2 5.84
38.1 4.85
3.57 0.86
n.d.
n.d.
n.d
n.d
10.7
4.7
17
n.c.
n.c.
n.c.
n.c.
21
(+)-FEOBV
56.9 4.8a
(rac)-5-IBVM
(rac)-5-ABV
(rac)-5-FBVM
(+)-(S,S)-5-FBVM
(À)-(R,R)-5-FBVM
15.8 5.42
40.6 1.69
10.9 2.78
6.95 0.62 (pKi = 8.16)
3.68 0.48 (pKi = 8.43)
16
229 60a
526b
1.21
5.48
0.97
76
68
252 13a
Ki values were determined by competition of derivatives against bound (À)-[3H]vesamicol under equilibrium at 22 °C, unless stated.
n.d.: not determinated.
n.c.: not calculated.
All experiments were performed in triplicate (n P 3; an = 2; bn = 1).
The Balz-Schiemann reaction (1927) is a well known process of
deamine fluorination of aniline via the SN1 mechanism by the ac-
tion of sodium nitrite (NaNO2), followed by thermal decomposition
with fluoroboric acid (HBF4). The mechanism of halogeno-dediazo-
tization may take place via an ionic pathway (heterolytic decom-
position)32,33 or via a radical pathway.34 An alkyl nitrite such as
n-butylnitrite or t-butylnitrile, that is, soluble in organic solvents
could be a good diazotizating agent according to the literature.35,36
Moreover they are very weak electrophiles (even weaker than the
nitrosonium ion). The solvent, which is one of the key parameters
for a successful reaction, should be a chlorinated organic solvent
because it is aprotic and presents non-nucleophilic and non-oxi-
dizing properties. Ionic liquid solvents are known to improve the
yield of the Balz-Schiemann reaction.37,38 Because of the high re-
dox potential of fluoride anion (e.g., F2/FÀ in CH3CN E = 2.4 V),
the formation of a fluoroaryl bond proceeds generally via a hetero-
lytic mechanism.39 Furthermore, the redox potential could be in-
creased by solvation.39 The source of the fluorine atom is
variable: silicium fluoride (SiF4) is fairly toxic; alkali metal fluo-
rides (CsF or KF) are hygroscopic with poor solubility; and quater-
nary ammonium fluorides are very hygroscopic but are soluble in
aprotic solvents. Other salts such as boron trifluoride-diethyl
etherate (BF3ÁEt2O) are excellent fluorinating agents35 and are sol-
uble in chlorinated solvents in contrast to nitrosonium tetrafluoro-
borate (NOBF4) or nitrosonium hexafluorophosphate (NOPF6). The
decomposition temperature of the diazonium group depends on
further substituents on the aromatic ring.40
Figure 3. Isosurface of the 3D QSAR model positive coefficients on Van der Waals
grids in green and negative coefficients on Van der Waals grid in yellow: in solid
representation for (+)-enantiomers model, in quad mesh representation for (À)-
enantiomers model with (+) and (À)-5-FBVM.
We tested several fluorinating agents and reaction conditions
such as (i) 1.5 equiv of NOBF4, CH2Cl2, 1 h reflux; (ii) 1.5 equiv of
NaBF4, 1.5 equiv t-BuONO, CH2Cl2, 1 h, rt; (iii) 1.5 equiv of
(t-Bu)4NF, 1.5 equiv t-BuONO, CH2Cl2, 1 h, reflux; and (iv) 1.5 equiv
of BF3ÁEt2O, 1.2 equiv t-BuONO, C6H4Cl2, 1 h, 60 °C. Scheme 1
presents the best results of one pot fluoro-dediazoniation of the pri-
mary aromatic amine 5-ABV to obtained 5-FBVM in moderate yield
(around 25% yield). The chlorinated organic solvents dichloroben-
zene (method A)35 or CH2Cl2 (method B) were suitable for such
From these three QSAR studies, 5-FBVM is predicted to be a li-
gand with nanomolar VAChT binding affinity, without significant
differences between the two enantiomers.
2.2. Chemistry
Synthesis of the 5-ABV enantiomers which were the basis for all
fluorinations has already been described.5,27,31
Method A
;Yield 16%
HO
HO
1.5 eq. BF3.Et2O, 1.2 eq. t-BuONO,C6H4Cl2
105°C, 1h
N
N
Method B;Yield 25%
H2N
F
1) 1.5 eq. BF3.Et2O, 1.5 eq. t-BuONO,
CH2Cl2 , RT, overnight
5-FBVM
5-ABV
2) reflux, 1.5h
Scheme 1. Synthesis of 5-FBVM from 5-ABV by one pot fluoro-dediazotization.

7664
M. Kovac et al. / Bioorg. Med. Chem. 18 (2010) 7659–7667
one pot fluoro-dediazoniation. Increasing of the amount of nitrite or
fluorine agent (2 equiv) did not increase the yield (12% yield).
Since this method is not useable for radiolabeling, we developed
a further procedure to introduce a fluorine atom, in particular for
future 18F-radiofluorination. Aryl-dialkyltriazenes (as a protected
form of aryldiazonium ions) have been extensively studied for dif-
ferent purposes (for review, see Ref. 41). We synthesized triazene
5-TBV precursors using two methods: the typical method via a dia-
zonium salt quenched with diethylamine to provide 5-TBV in
quantitative yield (method A, Scheme 2), and method B started
by the synthesis of a diazonium salt obtained by using t-butylni-
trite in CH2Cl2 then quenched with diethylamine to provide 88%
yield of 5-TBV. Triazenes can be decomposed by acidic-thermal-
decomposition through diazonium ion to provide fluorinated42,43
or iodinated44 derivatives as described for certain radiolabelings.
Two pathways are possible: photoinduced decomposition of 1-
aryl-3,3-dialkyltriazenes or thermal decomposition via ionic
paths.39 Both processes start with cleavage of the N(2)–N(3)
bond.34 Most studies agree on strong acid-catalyzed decomposi-
tion of triazenes, involving fast and reversible protonation of
N(3), followed by the ‘slow’ heterolytic cleavage of the N(2)–N(3)
bond to yield the corresponding diazonium ion and amine.39,45–47
Protonation of N(3) is a crucial step to decompose the triazeno
moiety (competing with N(1) protonation): the partial atom
charges of N(1) and N(3) for 5-TBV were À0.138 and À0.0912,
respectively (Discovery StudioÒ 2.5, CFF forcefield, Accelrys Inc),
corresponding to the dipolar charge distribution of the triazene
functional group.
and non-nucleophilic properties. However, we observed the forma-
tion of the corresponding aryl-p-toluenesulfonic ester analog. In
contrast, the reaction with triflic acid in CH2Cl2 was more success-
ful, and we obtained 5-FBVM with 25% yield (Scheme 3). A chlori-
nated organic solvent, especially CH2Cl2 (compared to C6H4Cl2),
was a good solvent, possibly due to the potential stabilization of
the diazonium-boron trifluoride complex intermediate. We also
tested several fluorinating agents, including tetra-n-butylammo-
nium fluoride (TBAF), which is soluble in chlorinated solvents.
Reaction of TBAF, CH2Cl2, and CF3SO3H under reflux was not suc-
cessful (data not shown). The use of CsF in carbon tetrachloride
with triflic acid, as previously described for dimethyltriazene,49
failed in our case (data not shown), possibly due to the insolubility
of CsF and triflic acid in CCl4. Using KF in C2H4Cl2 at reflux was also
not successful.
The enantiomeric purity and the optical rotation of (S,S)- and
(R,R)-5-FBVM were checked by chiral HPLC by using an amylose
based column in RP mode (91% CH3CN/20 mM NH4OAc aq) and a
chiral detector. Under these conditions, (S,S)-5-FBVM was a (+)-
enantiomer and (R,R)-5-FBVM a (À)-enantiomer.
2.3. In vitro evaluation
Vesamicol-derived ligands are generally insufficiently selective
towards VAChT due to a non-negligible affinity to
r receptors.
Since these receptors are distributed in cholinergic brain areas,
specific imaging of cholinergic deficiency is almost impossible with
unselective compounds.
Because the triazene moiety is decomposed through a diazo-
nium ion, the theoretical model for acid-catalyzed thermal decom-
position to yield the desired arylfluoride (fluoro-de-triazenation) is
the same as for fluoro-dediazoniation.48 This acid should not pres-
ent a nucleophilic conjugated base (AÀ) to prevent competition
with fluoride anion for aryl cation. Additionally, an acid with a
low redox potential (e.g., trifluoroacetic acid) is desirable, to pre-
vent reduction of the aryldiazonium ion via the radical pathway
and the formation of radicals.
Binding affinities of 5-FBVM and reference compounds to
VAChT and
r receptor sites were determined in vitro with Ki val-
ues presented in Table 3. 5-FBVM presents very good affinity for
VAChT and, surprisingly, the (+)-(S,S) enantiomer showed almost
the same affinity as the (À)-(R,R) enantiomer (Ki(S,S)FBVM = 6.95 nM
and Ki(R,R)FBVM = 3.68 nM), the affinity being better than for ABV
(Ki = 6.95 nM), that is, known to be a good ligand.50 Furthermore,
(+)-(S,S)-5-FBVM is selective for VAChT towards the r1 receptor
(Ki(S,S) = 38.1 nM and Ki(R,R) = 3.75 nM), and both isomers are selec-
For the synthesis of ( )-5-FBVM from the aryl-dialkyltriazene
( )-5-TBV we first tested TBFA in C6H4Cl2 or in CH2Cl2 at reflux
and p-toluenesulfonic acid (PTSA) with a high redox potential
tive for VAChT towards r2 receptors (Ki >200 nM). The affinity of 5-
FBVM is in the same range as the affinity of 5-IBVM which is cur-
rently regarded as the standard radiotracer.
Method A
;Yield 95%
1) 1.5 eq. NaNO2, HCl aq, 0°C, 0.5 h
2) NaHCO3 aq, 1.5 eq. Et2NH, 20°C, 1h
HO
HO
N
N
Method B; Yield 88%
H2N
Et2N-N=N
1) 1.5 eq. BF3.Et2O, CH2Cl2 ,
1.5 eq. t-BuONO, 0°C, 2 h
5-TBV
5-ABV
2) NaHCO3 aq, 1.5 eq. Et2NH, 20°C, 1h
Scheme 2. Synthesis of (S,S) and (R,R)-5-TBV.
HO
N
HO
N
1) 3 eq. CF3SO3, CH2Cl2
2) 1.5 eq. BF3.Et2O, 40°C, 1h
Et2N-N=N
F
25%
(rac)-5-FBVM
(rac)-5-TBV
Scheme 3. Synthesis of (rac)-5-FBVM from triazene precursor (rac)-5-TBV.

M. Kovac et al. / Bioorg. Med. Chem. 18 (2010) 7659–7667
7665
4.2.15-Fluoro-3-(4-phenyl-piperidin-1-yl)-1,2,3,4-tetrahydro-
3. Conclusion
naphthalen-2-ol: 5-FBVM from 5-ABV
4.2.1.1 Method A.
We obtained a good linear GFA model and a 3D QSAR model
which confirmed the spatial impact on affinity for VAChT via steric
descriptors and the Van der Waals coefficient. Each study predicted
a good affinity of both 5-FBVM enantiomers.
To a cold (0 °C) solution of (rac)-5-ABV (0.161 mg, 0.5 mmol) in
anhydrous 1,2-dichlorobenzene (4 mL), boron trifluoride diethyl
etherate (BF3ÁO(CH2CH3)2) (0.064 mL, 0.75 mmol) was added. The
stirred reaction mixture was warmed to 105 °C and n-butylnitrite
(0.07 mL, 0.6 mmol) was added. The mixture was warmed for
1 h. After cooling to room temperature, the solution was quenched
with water and extracted with EtOAc. The water phase was made
basic by Na2CO3 and was extracted once again with EtOAc. The
combined organic extracts were dried over MgSO4 and concen-
trated under reduced pressure. The crude product was purified
by gradient flash chromatography (Al2O3, n-hexane/EtOAc 4/1 to
n-hexane/EtOAc 1/1). (rac)-5-FBVM was obtained as a white pow-
der in 16% yield.
All one pot acid-catalyzed thermal fluoro-dediazoniation reac-
tions confirmed that parameters such as good solubility, non-
nucleophilicity, and high redox potential of the reagents used, such
as triflic acid, are the most important conditions for successful flu-
oro-detriazenation via the formation of diazonium ion. Using bor-
on trifluoride etherate we succeeded in introducing fluorine on an
aromatic nucleus. Dichloromethane proved to be the most oppor-
tune solvent for one pot fluoro-dediazoniation. The in vitro evalu-
ations confirmed the QSAR model where both enantiomers
exhibited high affinity for VAChT [(+)-5-FBVM: Ki = 6.95 nM and
(À)-5-FBVM: Ki = 3.68 nM]. The stereoisomers were selective to-
wards r2 receptors (ꢀ70-fold), however, only (+)-5-FBVM is also
selective for r1 receptors (ꢀfivefold). Further experiments are
needed to improve the characterization of the pharmacological
and pharmacokinetics profiles of this compound in order to deter-
mine its potential use as an F-18-labeled imaging agent for studies
involving the cholinergic system.
4.2.1.2. Method B.
(0.161 mg, 0.5 mmol) in anhydrous dichloromethane (4 mL) boron
trifluoride diethyl etherate (0.064 mL,
(BF3ÁO(CH2CH3)2)
To a cold (0 °C) solution of (rac)-5-ABV
0.75 mmol) and n-butylnitrite (0.095 mL, 0.75 mmol) were added.
The mixture was stirred at room temperature overnight and after-
ward heated to 60 °C for drying. After cooling, water was added to
this reaction residue. The water phase was made basic by NaHCO3
and was extracted with EtOAc for two times. The combined organic
extracts were dried over MgSO4 and concentrated under reduced
pressure. The crude product was purified by gradient flash chroma-
tography (SiO2, n-hexane/EtOAc 4/1 to n-hexane/EtOAc 1/2). (rac)-
5-FBVM was obtained as a white powder in 25% yield.
4. Experimental section
4.1. Molecular modeling
Computational results were obtained using software programs
from Accelrys Software Inc. The molecules were built and mini-
mized in molecular package (Discovery StudioÒ 2.5.5, Accelrys,
San Diego, CA) by CHARMm with CFF partial charge estimation
method. The GFA model in QSAR protocol was used with a popula-
tion size of 100 and 5000 maximum generations. All the parame-
ters have been left to the system defaults. Two model forms have
been used: linear or full quadratic. For the 3D QSAR model, the grid
spacing was 1 Å.
(+)-5-FBVM and (À)-5-FBVM were obtained by the same proce-
dure as white powders in 27% and 25% yield, respectively, starting
from the corresponding (+)-5-ABV and (À)-5-ABV.
1H NMR (CDCl3): d 1.75–1.96 (m, 4H, 4H-10), 2.53–3.08 (m, 9H,
1H-11, 1H-1, 2H-4, 1H-3, 4H-9), 3.25 (dd, J = 5.4 Hz, J = 16 Hz, 1H-
1), 3.85–3.99 (m, 1H-2), 4.34 (br s, OH), 6.86–6.97 (m, 1HAr), 7.10–
7.17 (m, 2HAr), 7.26–7.41 (m, 5HAr).
13C NMR (CDCl3): d 19.0 (C-4), 33.7, 34.2 (2C-10), 37.6 (C-1),
42.8 (C-11), 44.8, 53.4 (2C-9), 65.1 (C-2), 65.9 (C-3), 112.1 (2JC–F
=
22 Hz, C-6), 122.5 (2JC–F = 18 Hz, C-4a), 124.5 (4JC–F = 3 Hz, C-8),
126.2 (CHAr), 126.7 (2CHAr), 127.1 (3JC–F = 8 Hz, C-7), 128.4 (2CHAr),
136.5 (3JC–F = 8 Hz, C-8a), 145.9 (CAr), 160.9 (1JC–F = 243 Hz, C-5).
MS: m/z = 325 (14), 228 (74), 174 (35), 161 (18), 155 (16), 146
(100), 133 (32), 56 (31).
4.2. Chemistry
NMR spectra were recorded on a Bruker DPX Avance 200 spec-
trometer (200 MHz for 1H, 50.3 MHz for 13C). CDCl3 was used as
solvent; chemical shifts are expressed in ppm relative to TMS as
an internal standard. Mass spectra were obtained on a CG–MS
Hewlett Packard 5989A spectrometer (electronic impact at
70 eV). The thin-layer chromatographic (TLC) analyzes were per-
formed using Merck 60-F254 silica gel plates. Flash chromatography
was used for routine purification of reaction products using silica
gel (230–400 mesh). Visualization was accomplished under UV or
in an iodine chamber. All chemicals and solvents were of commer-
cial quality and were purified following standard procedures. Ele-
mental analyzes of new compounds were within 0.5% of
theoretical values.
Anal. Calcd for rac-C21H24FNO: C, 77.51; H, 7.43. Found: C,
77.90; H, 3.91.
4.2.25-(3,3-Diethyltriaz-1-enyl)-3-(4-phenylpiperidin-1-yl)-
1,2,3,4-tetrahydronaphthalen-2-ol: (rac)-5-ABV-diethyltriazene
(5-TBV)
4.2.2.1 Method A.
(rac)-5-ABV (256 mg, 0.8 mmol) was dissolved in 0.1 mL of 12 N
HCl and the flask was immersed in an ice-bath (0–5 °C). A 1.5 equiv
of NaNO2 (82 mg) was added to the solution and the reaction mix-
ture was stirred for 30 min. After neutralization with saturated
aqueous solution of NaHCO3, 1.5 equiv of diethyl amine (0.12 mL)
was added to react for 1 h. The mixture was extracted three times
with CH2Cl2. Combined organic extracts were dried over MgSO4
and concentrated under reduced pressure. The triazene (rac)-5-
TBV was obtained as a white powder and was pure enough to be
used without any purification (95% yield).
4-Phenylpiperidine, used in the synthesis of 5- and 8-amino-
benzovesamicol (ABV),27 was obtained by alkaline fusion from 4-
cyano-4-phenylpiperidine.31 Enantiomeric resolution of 5-ABV
was done by chromatography separation of diastereomeric N,O-
bis-(À)-
a-methoxy-a-trifluoromethylphenylacetyl (MTPA) deriva-
tives followed by hydrolysis of Mosher esters.5 The optical purity
of both (À)-ABV and (+)-ABV was checked on Chiracel OD column
(+)-5-TBV and (À)-5-TBV were obtained by the same procedure
as white powders in 92% or 96% yield, respectively, starting from
the corresponding (+)-5-ABV and (À)-5-ABV.
(4.6 Â 250 nm, 10
lm particle, Daicel Chemical Industries Ltd, Ill-
kirch France) with n-hexane/isopropanol (80/20) as eluent at a flow
rate of 1.5 mL/min ((+)-ABV; tR = 11 min, (À)-ABV); tR = 13.5 min).
(À)-5-IBVM18,27 and (À)-FEOBV5 were synthesized as previ-
ously described.
4.2.2.2. Method B.
(rac)-5-ABV (256 mg, 0.8 mmol) was dis-
solved in CH2Cl2 (2 mL) and the flask was immersed in an ice-bath

7666
M. Kovac et al. / Bioorg. Med. Chem. 18 (2010) 7659–7667
(0–5 °C). A 1.5 equiv of boron trifluoride diethyl etherate (0.15 mL)
and 1.5 equiv of t-butylnitrite (0.14 mL) were added to the stirred
solution. After 2 h, the reaction mixture was neutralized with sat-
urated aqueous solution of NaHCO3 and 1.5 equiv of diethyl amine
(0.12 mL) was added and stirred for 1 h. The mixture was
quenched with water and extracted three times with CH2Cl2. Com-
bined organic extracts were dried over MgSO4 and concentrated
under reduced pressure. The triazene (rac)-5-TBV was pure enough
to be used without any purification (88% yield).
7.4 at room temperature. Incubation was terminated after 60 min
by filtration (GF-B filter, pre-incubated in 0.3% PEI at room temper-
ature for 90 min; Brandel Cell harvester). Non-specific binding was
determined in the presence of 10 mM ( )-vesamicol.
r
1 affinity was determined by radioligand displacement studies
on homogenates of rat cortical membranes by using (À)-[3H]pen-
tazocine (Perkin Elmer; specific activity: 1070 GBq/mmol). Assays
were incubated in 50 mM TRIS–HCl, pH 7.4 at room temperature.
Incubation was terminated after 120 min by filtration (GF-B filter,
pre-incubated in 0.3% PEI at room temperature for 90 min; Brandel
Cell harvester). Non-specific binding was determined in the pres-
1H NMR (CDCl3): d 1.34 (t, J = 7 Hz, 6H, 2CH3), 1.76–2.01 (m, 4H,
2H-10, 2H-12), 2.48–3.08 (m, 8H, 1H-11, H-1, 2H-4, 2H-9, 2H-13),
3.37 (dd, J = 5.6 Hz, J = 16 Hz, 1H-1), 3.45 (dd, J = 3.4 Hz, J = 15.5 Hz,
1H-1), 3.83 (q, J = 7 Hz, 4H, 2CH2), 3.91–3.97 (m, 1H-2), 6.95 (d,
J = 6.5 Hz, 1HAr), 7.15–7.37 (m, 7HAr).
ence of 10
lM haloperidol.
r
2 affinity was determined by radioligand displacement studies
on homogenates of rat liver membranes by using (À)-[3H]DTG
13C NMR (CDCl3): d 12.6 (br s, 2CH3); 21.5 (C-4), 33.8, 34.3 (2C-
10), 38.1 (C-1), 42.8 (C-11), 45.0, 53.5 (2C-9), 65.4 (C-2), 66.7 (C-3),
114.0 (CHAr), 125.9, 126.1, 126.3 (3CHAr), 126.7 (2CHAr), 128.4
(2CHAr), 129.4 (CAr), 134.6 (CAr), 146.1 (CAr), 148.9 (CAr).
MS: m/z = 307 (84), 174 (100), 129 (26), 117 (44), 115 (41), 91
(49), 70 (33).
(Perkin Elmer; specific activity: 1147 GBq/mmol) in the presence
of 1 lM dextrallorphan (Roche) to block r
1 binding of [3H]DTG. As-
says were incubated in 50 mM TRIS–HCl, pH 7.4 at room tempera-
ture. Incubation was terminated after 120 min by filtration (GF-B
filter, pre-incubated in 0.3% PEI at room temperature for 90 min;
Brandel Cell harvester). Non-specific binding was determined in
the presence of 10 lM haloperidol.
4.2.3. (rac)-5-Fluoro-3-(4-phenyl-piperidin-1-yl)-1,2,3,4-
tetrahydro-naphthalen-2-ol: (rac)-5-FBVM from 5-triazene 5-
TBV
All assays were performed in triplicates at least three times. The
IC50-values were estimated by computational non-linear regres-
sion analysis. Ki-Values were calculated according to Cheng and
Prusoff.51
To a solution of (rac)-5-TBV (0.162 g, 0.4 mmol) in anhydrous
CH2Cl2 (5 mL) trifluoromethanesulfonic acid monohydrate
(CF3SO3HÁH2O) (0.18 mg, 1.2 mmol) dissolved in CH3CN (0.5 mL)
was added. BF3ÁEt2O (boron trifluoride diethyl etherate, 0.075 mL,
0.6 mmol) diluted in CH2Cl2 (0.2 mL) was then added to the stirred
reaction mixture. The reaction mixture was heated to 60 °C for dry-
ing over 1 h. After cooling, water was added to this reaction resi-
due. The water phase was made basic by NaHCO3 and was
extracted with CH2Cl2. The combined organic extracts were dried
over MgSO4 and concentrated under reduced pressure. The crude
product was purified by gradient flash chromatography (SiO2, n-
hexane/EtOAc 4/1 to n-hexane/ EtOAc 1/2). (rac)-5-FBVM was ob-
tained as a white powder in 25% yield.
Acknowledgment
This work was supported by INSERM. This study was funded in
part by FEDER: ImAD project. We thank the ‘Département d’analy-
ses Chimiques et S.R.M. biologique et médicale’ (Tours, France) for
chemical analyzes.
References and notes
1. Bauer, J.; Hull, M.; Berger, M. Z. Gerontol. Geriatr. 1995, 28, 155.
2. Terry, R. D.; Masliah, E.; Salmon, D. P.; Butters, N.; DeTeresa, R.; Hill, R.; Hansen,
L. A.; Katzman, R. Ann. Neurol. 1991, 30, 572.
3. Giboureau, N.; Emond, P.; Fulton, R. R.; Henderson, D. J.; Chalon, S.; Garreau, L.;
Roselt, P.; Eberl, S.; Mavel, S.; Bodard, S.; Fulham, M. J.; Guilloteau, D.; Kassiou,
M. Synapse 2007, 61, 962.
4. Kilbourn, M. R.; Hockley, B.; Lee, L.; Sherman, P.; Quesada, C.; Frey, K. A.;
Koeppe, R. A. Nucl. Med. Biol. 2009, 36, 489.
5. Mulholland, G. K.; Jung, Y. W.; Wieland, D. M.; Kilbourn, M. R.; Kuhl, D. E. J. J.
Labelled Compd. Radiopharm. 1993, 33, 583.
6. Rogers, G. A.; Stone-Elander, S.; Ingvar, M.; Eriksson, L.; Parsons, S. M.; Widén,
L. Nucl. Med. Biol. 1994, 21, 219.
7. Efange, S. M.; Langason, R. B.; Khare, A. B.; Low, W. C. Life Sci. 1996, 58,
1367.
8. Efange, S. M.; Khare, A.; Parsons, S. M.; Bau, R.; Metzenthin, T. J. Med. Chem.
1993, 36, 985.
9. Sorger, D.; Scheunemann, M.; Vercouillie, J.; Grossmann, U.; Fischer, S.; Hiller,
A.; Wenzel, B.; Roghani, A.; Schliebs, R.; Steinbach, J.; Brust, P.; Sabri, O. Nucl.
Med. Biol. 2009, 36, 17.
10. Sorger, D.; Scheunemann, M.; Großmann, U.; Fischer, S.; Vercouille, J.; Hiller,
A.; Wenzel, B.; Roghani, A.; Schliebs, R.; Brust, P.; Sabri, O.; Steinbach, J. Nucl.
Med. Biol. 2008, 35, 185.
4.3. Determination of optical rotation of (R,R)-5-FBVM and
(S,S)-5-FBVM
The analytical separation of the 5-FBVM enantiomers by chiral
HPLC was performed on
a Reprosil Chiral-AM-RP column
(250 Â 4.6 mm), which is based on amylose-tris-(3,5-dimethyl-
phenyl)-carbamate as chiral selector (Dr. Maisch-GmbH, Ger-
many). The optical rotation was determined by using a chiral
detector (OR 2090 model from JASCO, Germany). In general, the
OR detector operated under following conditions: range: 0.05, re-
sponse: SLOW, gain: 10. The polarity of signal amplitudes obtained
by the chiral detector was checked with (À)-vesamicol as reference
compound. By using 91% CH3CN/20 mM NH4OAc aq and a flow rate
of 1 mL/min the enantiomers were separated. With tR = 15.5 min
the (À)-(R,R)-5-FBVM eluted in front of the (+)-(S,S)-5-FBVM with
tR = 23.5 min.
11. Parsons, S. M.; Bahr, B. A.; Rogers, G. A.; Clarkson, E. D.; Noremberg, K.; Hicks, B.
W. Prog. Brain Res. 1993, 98, 175.
12. Bahr, B. A.; Clarkson, E. D.; Rogers, G. A.; Noremberg, K.; Parsons, S. M.
Biochemistry 1992, 31, 5752.
13. Altar, C. A.; Marien, M. R. Synapse 1988, 2, 486.
4.4. Receptor binding studies
14. Bahr, B. A.; Parsons, S. M. Proc. Natl. Acad. Sci. U.S.A. 1986, 83, 2267.
15. Efange, S. M.; Mach, R. H.; Smith, C. R.; Khare, A. B.; Foulon, C.; Akella, S. K.;
Childers, S. R.; Parsons, S. M. Biochem. Pharmacol. 1995, 49, 791.
16. Ichikawa, T.; Ajiki, K.; Matsuura, J.; Misawa, H. J. Chem. Neuroanat. 1997, 13,
23.
For radioligand displacement studies, the test compounds were
solved in DMSO at 10 mM stock solutions. Serial dilutions in the
range of 0.01 nM to 1
incubation buffer.
lM were obtained by further dilution in
17. Phan, V. L.; Miyamoto, Y.; Nabeshima, T.; Maurice, T. J. Neurosci. Res. 2005, 79,
561.
18. Jung, Y. W.; Van Dort, M. E.; Gildersleeve, D. L.; Wieland, D. M. J. Med. Chem.
1990, 33, 2065.
19. Kuhl, D. E.; Minoshima, S.; Fessler, J. A.; Frey, K. A.; Foster, N. L.; Ficaro, E. P.;
Wieland, D. M.; Koeppe, R. A. Ann. Neurol. 1996, 40, 399.
20. Sorger, D.; Schliebs, R.; Kampfer, I.; Rossner, S.; Heinicke, J.; Dannenberg, C.;
Georgi, P. Nucl. Med. Biol. 2000, 27, 23.
VAChT affinity was determined by radioligand displacement
studies on homogenates of PC12 cells stably transfected with
rVAChT (Ali Roghani, Texas Tech University, Lubbock, TX, USA)
by using (À)-[3H]vesamicol (Perkin Elmer; specific activity:
1296 GBq/mmol). Assays were incubated in 50 mM TRIS–HCl, pH

M. Kovac et al. / Bioorg. Med. Chem. 18 (2010) 7659–7667
7667
21. Zea-Ponce, Y.; Mavel, S.; Assaad, T.; Kruse, S. E.; Parsons, S. M.; Emond, P.; Chalon,
34. Galli, C. Chem. Rev. 1988, 88, 765.
S.; Giboureau, N.; Kassiou, M.; Guilloteau, D. Bioorg. Med. Chem. 2005, 13, 745.
22. Jung, Y. W.; Frey, K. A.; Mulholland, G. K.; del Rosario, R.; Sherman, P. S.; Raffel,
D. M.; Van Dort, M. E.; Kuhl, D. E.; Gildersleeve, D. L.; Wieland, D. M. J. Med.
Chem. 1996, 39, 3331.
35. Garel, L.; Saint-Jalmes, L. Tetrahedron Lett. 2006, 47, 5705.
36. Doyle, M. P.; Bryker, W. J. J. Org. Chem. 1979, 44, 1572.
37. Laali, K. K.; Gettwert, V. J. J. Fluorine Chem. 2001, 107, 31.
38. Hubbard, A.; Okazaki, T.; Laali, K. K. J. Org. Chem. 2008, 73, 316.
23. Assaad, T.; Mavel, S.; Emond, P.; Chalon, S.; Drossard, M. L.; Vergote, J.; Bodard,
S.; Allouchi, H.; Besnard, J.-C.; Guilloteau, D. J. Labelled Compd. Radiopharm.
2007, 50, 139.
39. Satyamurthy, N.; Barrio, J. R.; Schmidt, D. G.; Kammerer, C.; Bida, G. T.; Phelps,
M. E. J. Org. Chem. 1990, 55, 4560.
40. Yoneda, N.; Fukuhara, T. Tetrahedron 1996, 52, 23.
24. Assaad, T.; Mavel, S.; Parsons, S. M.; Kruse, S.; Galineau, L.; Allouchi, H.;
Kassiou, M.; Chalon, S.; Guilloteau, D.; Emond, P. Bioorg. Med. Chem. Lett. 2006,
16, 2654.
41. Kimball, D. B.; Haley, M. M. Angew. Chem., Int. Ed. 2002, 41, 3338.
42. Berridge, M. S.; Sayre, L. M.; Arora, P. K.; Terris, A. H.; Riachi, N. J.; Harik, S. I. J.
Med. Chem. 1993, 36, 1284.
25. Emond, P.; Mavel, S.; Zea-Ponce, Y.; Kassiou, M.; Garreau, L.; Bodard, S.;
Drossard, M. L.; Chalon, S.; Guilloteau, D. Nucl. Med. Biol. 2007, 34, 967.
26. Szymoszek, A.; Wenzel, B.; Scheunemann, M.; Steinbach, J.; Schüürmann, G. J.
Med. Chem. 2008, 51, 2128.
43. Pages, T.; Langlois, B. R. J. Fluorine Chem. 2001, 107, 321.
44. Harmata, M.; Hong, X.; Zheng, P. Tetrahedron Lett. 2006, 47, 7343.
45. Schmidt, B. F.; Snyder, E. J.; Carroll, R. M.; Farnsworth, D. W.; Michejda, C. J.;
Smith, R. H. J. Org. Chem. 1997, 62, 8660.
27. Rogers, G. A.; Parsons, S. M.; Anderson, D. C.; Nilsson, L. M.; Bahr, B. A.;
Kornreich, W. D.; Kaufman, R.; Jacobs, R. S.; Kirtman, B. J. Med. Chem. 1989, 32,
1217.
28. Rogers, D.; Hopfinger, A. J. J. Chem. Inf. Comput. Sci. 1994, 34, 854.
29. Stanton, D. T.; Jurs, P. C. Anal. Chem. 1990, 62, 2323.
30. Shiba, K.; Ogawa, K.; Ishiwata, K.; Yajima, K.; Mori, H. Bioorg. Med. Chem. 2006,
14, 2620.
31. Berkoff, C. E.; Rivard, D. E.; Kirkpatrick, D.; Ives, J. L. Synth. Commun. 1980, 10,
939.
32. Hashida, Y.; Landells, R. G. M.; Lewis, G. E.; Szele, I.; Zollinger, H. J. Am. Chem.
Soc. 1978, 100, 2816.
46. Smith, R. H.; Wladkowski, B. D.; Mehl, A. F.; Cleveland, M. J.; Rudrow, E. A.;
Chmurny, G. N.; Michejda, C. J. J. Org. Chem. 1989, 54, 1036.
47. Tewson, T. J.; Welch, M. J. J. Chem. Soc., Chem. Commun. 1979, 1149.
48. Barrio, J. R.; Satyamurthy, N.; Ku, H.; Phelps, M. E. J. Chem. Soc., Chem. Commun.
1983, 443.
49. Pages, T.; Langlois, B. R.; Bars, D. L.; Landais, P. J. Fluorine Chem. 2001, 107, 329.
50. Rogers, G.; Kornreich, W.; Hand, K.; Parsons, S. Mol. Pharmacol. 1993, 44, 633.
51. Cheng, Y.; Prusoff, W. H. Biochem. Pharmacol. 1973, 22, 3099.
52. Shiba, K.; Yano, T.; Sato, W.; Mori, H.; Tonami, N. Life Sci. 2002, 71, 1591.
53. Efange, S. M. N.; von Hohenberg, K.; Khare, A. B.; Tu, Z.; Mach, R. H.; Parsons, S.
M. Nucl. Med. Biol. 2000, 27, 749.
33. Szele, I.; Zollinger, H. J. Am. Chem. Soc. 1978, 100, 2811.