Synthesis, modelling, and μ-opioid receptor affinity of N-3(9)-arylpropenyl-N-9(3)-propionyl-3,9-diazabicyclo[3.3.1]nonanes

Article abstract of DOI:10.1016/S0014-827X(00)00036-7

A series of N-3-arylpropenyl-N-9-propionyl-3,9-diazabicyclo[3.3.1]nonanes (1a-g) and of reverted N-3-propionyl-N-9-arylpropenyl isomers (2a-g), as homologues of the previously reported analgesic 3,8-diazabicyclo[3.2.1]octanes (I-II), were synthesized and evaluated for the binding affinity towards opioid receptor subtypes μ, δ and κ. Compounds 1a-g and 2a-g exhibited a strong selective μ-affinity with K_i values in the nanomolar range, which favourably compared with those of I and II. In addition, contrary to the trend observed for DBO-I, II, the μ-affinity of series 2 is markedly higher than that of the isomeric series 1. This aspect was discussed on the basis of the conformational studies performed on DBN which allowed hypotheses on the mode of interaction of these compounds with the μ receptor.

Full text of DOI:10.1016/S0014-827X(00)00036-7

www.elsevier.nl/locate/farmac
Il Farmaco 55 (2000) 553–562
Synthesis, modelling, and m-opioid receptor affinity of
N-3(9)-arylpropenyl-N-9(3)-propionyl-3,9-
diazabicyclo[3.3.1]nonanes
G.A. Pinna a, G. Murineddu a, M.M. Curzu a, S. Villa b, P. Vianello b, P.A. Borea c,
S. Gessi c, L. Toma d, D. Colombo e, G. Cignarella b,*
a Dipartimento Farmaco Chimico Tossicologico, Uni6ersita` di Sassari, Via Muroni 23, 07100 Sassari, Italy
b Istituto di Chimica Farmaceutica e Tossicologica, Uni6ersita` di Milano, Viale Abruzzi 42, 20131 Milan, Italy
c Dipartimento di Medicina Clinica e Sperimentale, Via Fossato di Mortara 19, 44100 Ferrara, Italy
d Dipartimento di Chimica Organica, Uni6ersita` di Pa6ia, 6ia Taramelli 10, 27100 Pa6ia, Italy
e Dipartimento di Chimica e Biochimica Medica, Uni6ersita` di Milano, Via Saldini 50, 20133, Milan, Italy
Received 30 November 1999; accepted 28 February 2000
Dedicated to Professor Pietro Pratesi
Abstract
A series of N-3-arylpropenyl-N-9-propionyl-3,9-diazabicyclo[3.3.1]nonanes (1a–g) and of reverted N-3-propionyl-N-9-aryl-
propenyl isomers (2a–g), as homologues of the previously reported analgesic 3,8-diazabicyclo[3.2.1]octanes (I–II), were synthe-
sized and evaluated for the binding affinity towards opioid receptor subtypes m, d and k. Compounds 1a–g and 2a–g exhibited
a strong selective m-affinity with Ki values in the nanomolar range, which favourably compared with those of I and II. In addition,
contrary to the trend observed for DBO-I, II, the m-affinity of series 2 is markedly higher than that of the isomeric series 1. This
aspect was discussed on the basis of the conformational studies performed on DBN which allowed hypotheses on the mode of
interaction of these compounds with the m receptor. © 2000 Elsevier Science S.A. All rights reserved.
Keywords: Diazabicyclononanes; Modelling; m-Affinity
1. Introduction
doethylenic bridge of DBO-I and -II plays an essential
role in modulating m-affinity by fitting lipophylic pock-
ets of the receptor. The hypothesis was supported by
the significantly lower m-affinity found in the corre-
sponding piperazine and equatorial cis-2,6-dimethyl
piperazine derivatives [3].
On the basis of this assumption, we have now
planned to evaluate the affinity towards m-opioid recep-
tors of 3,9-diazabicyclo[3.3.1]nonanes (DBN), higher
homologues of DBO-I and -II, in which the two carbon
bridge of DBO is replaced by a three carbon bridge [4].
The class of N-8-propionyl-N-3-arylpropenyl-3,8-di-
azabicyclo[3.2.1]octane (DBO-I) as well as of the N-3-
propionyl-N-8-arylpropenyl isomers (DBO-II) are
central analgesics displaying a potency comparable or
greater than morphine [1,2].
Despite the structural dissimilarity with common opi-
oids, their activity is related to an affinity towards
m-opioid receptors in the nanomolar range, similarly to
morphine but with higher m/d, k selectivity [2].
Molecular modelling studies suggest that the en-
Accordingly,
representative
9-propionyl-3-aryl-
propenyl DBN (1a–g) and their 3-propionyl-9-aryl-
propenyl isomers (2a–g) have been synthesized and
tested in vitro towards m-opioid receptors (m, d, k for
1a, 2a) comparing their affinity values with those of the
corresponding DBO recently reported [5,6].
* Corresponding author. Tel.: +39-02-2951 0843; fax: +39-02-
2951 4197.
E-mail address: giorgio.cignarella@unimi.it (G. Cignarella).
0014-827X/00/$ - see front matter © 2000 Elsevier Science S.A. All rights reserved.
PII: S0014-827X(00)00036-7

554
G.A. Pinna et al. / Il Farmaco 55 (2000) 553–562
[3.3.1]nonane (4) still unknown in the literature, which
was obtained in 88% yield by heating 3 at 150°C for 2 h
with consequent N9–N3 acyl migration, following a
procedure described for the DBO analogues [7].
The needed chlorides (14) were prepared from the
appropriate aryl acrylic acid (12b,d) or ester (12c,e–g)
by reduction with diisobutylaluminum hydride in tolu-
ene to the corresponding arylpropenyl alcohol (13)
eventually converted to 14 with thionyl chloride or
concentrated HCl (Scheme 3).
3. Receptor binding
2. Chemistry
3.1. Opioid receptor binding assays
The synthesis of 1a–g and 2a–g required as interme-
diate the known 3-benzyl-3,9-diazabicyclo[3.3.1]nonane
(10) which was obtained by modifying the previously
reported procedure1. Thus, as outlined in Scheme 1,
meso-dimethyl-a,a%-dibromopimelate (6), was con-
densed with benzylamine in refluxing benzene to give in
88% yield N-benzyl-2,6-dicarbomethoxypiperidine (7)
as a mixture of the cis and the trans isomers, which, as
such, was allowed to react with benzylamine in reflux-
ing toluene for 18 h, and, after evaporation of the
solvent, at 160–170°C for a further 4 h. The oily
product was dissolved in ethanol–ether (1/1) and
treated with ethereal HCl to separate in 40% yield
3,9-dibenzyl-3,9-diazabicyclo[3.3.1]nonane-2,4-dione (8)
as the HCl salt. Owing to the comparable yield of 8,
this procedure was preferred to that involving a prelim-
inar separation by flash-chromatography of cis-7 and
trans-7, followed by condensation of cis-7 with benzyl-
amine. Catalytic hydrogenolysis (Pd/C) of 8 · HCl in
ethanol afforded in almost quantitative yield 3-benzyl-
3,9-diazabicyclo[3.3.1]nonane-2,4-dione (9) as the HCl
salt. Reduction of 9 with LiAlH4 in ether gave in 80%
yield the expected 10 as an oil.
Binding affinities to opioid receptors were measured
on mouse brain homogenates in the presence of [3H]-
DAMGO for m, [3H]-DELTORPHINE II for d. [3H]-
U69 593 was used on guinea pig homogenates to
evaluate k binding. Morphine was used as the reference
compound. The inhibition constants of the compounds
tested are reported in Tables 1 and 2.
Acylation of 10 with propionic anhydride to the
3-benzyl-9-propionyl-3,9-diazabicyclo[3.3.1]nonane (11)
followed by debenzylation (H2, Pd/C) led to 9-propi-
onyl-3,9-diazabicyclo[3.3.1]nonane (3) in 82% overall
yield from 10. Finally, condensation of 3 with the
appropriate arylpropenyl chlorides (14) in refluxing ace-
tone or butanone in the presence of K2CO3 gave com-
pounds 1a–g (Scheme 2).
The isomeric series 2a–g was synthesized by a similar
procedure starting from 3-propionyl-3,9-diazabicyclo-
1 Actually, at the very beginning of our research on DBOs, one
example was reported of homologation of the mild analgesic 3-
methyl-8-propionyl DBO to 3-methyl-9-propionyl DBN. The result-
ing decrease of the activity by about 50% (Randall and Selitto test,
rat) discouraged this approach, see [4].
Scheme 1. Reagents and conditions: (i) SOCl2, MeOH, Br2, D; (ii)
PhCH2NH2, C6H6, D; (iii) PhCH2NH2, toluene, D, Et2O/HCl; (iv)
EtOH, Pd/C 10%, H2; (v) H2O, NaCO3 10%; (vi) Et2O, LiAlH4, D.

G.A. Pinna et al. / Il Farmaco 55 (2000) 553–562
555
Scheme 2. Reagents and conditions: (i) CH2Cl2, (CH3CH2CO)2O, D, NaOH 20%; (ii) EtOH, Pd/C 10%, H2, (iii) D; 150°C,
2 h;
(iv)
, K2CO3, acetone or butanone, D.
(Chart 2). Actually, exploration of the conformational
space of compounds 1h and 2h showed a large prefer-
ence for the chair–chair conformations (type A) in both
cases; in fact, all other conformations were more than 4
kcal/mol higher in energy.
4. Modelling
The conformational properties of compounds 1a and
2a, ground terms of the two series of compounds 1 and
2, were investigated with the MM+ force field of the
H
YPER C
HEMTM package [9] considering the protonated
form of each compound. First, attention was focused
on the bicyclic moiety by considering the simplified
models 1h and 2h (Chart 1) in which a methyl group
replaces the phenylpropenyl group, as it was previously
shown [10] on the DBO analogues that the conforma-
tional mobility of this group does not influence the
conformational mobility of this group does not influ-
ence the conformational behaviour of the remaining
part of the molecule.
Chart 1.
In principle, the DBN system could present a confor-
mational flexibility higher with respect to the DBO
system. In fact, the two six-membered rings that consti-
tute the bicycle could assume four combinations of
conformations: chair–chair, chair–boat, boat–chair,
boat–boat exemplified by the four structures A–D
Chart 2.

556
G.A. Pinna et al. / Il Farmaco 55 (2000) 553–562
Scheme 3. Reagents and conditions: (i) DIBALH, Et2O or toluene, 0–5°C; (ii) SOCl2, dry pyridine, CH2Cl2 or concentrated HCl, D.
(Table 1). Having found for both compounds a m-
affinity in the nanomolar range (Ki=29 and 13 nM,
respectively) with a m/d and m/k selectivity from two to
three orders of magnitude, the compounds 1b–g and
2b–g were only tested on m receptors. The Ki values of
DBN series 1 and 2 are shown in Table 3 in comparison
with the Ki values of corresponding DBO-I and DBO-II
recently reported [6].
It is worth noting the m-affinity of DBN-2 favourably
compared with that of the lower homologues DBO-II,
as indicated by the Ki values of 2c,d,f,g which are
significantly lower than those of the corresponding II.
On the contrary, DBN-1 are definitely less potent
than the corresponding DBO-I with the only exception
being 1a (Ki=29 nM) versus Ia (Ki=55 nM).
Interestingly, 2a–g show m-affinity markedly greater
than that of the isomers 9-propionyl DBN (1a–g), the
most active terms (2b–d,f,g) exhibiting Ki values (from
5.8 to 7.6 nM) very close to that of morphine (Ki=2.8
nM).
Fig. 1 reports the 3D plots of the only populated
conformation of 1h (1hA) and of the two populated
conformations of 2h (2hA and 2hA%); these last two
conformations differ only for the orientation of the
methyl group at N9 and are practically isoenergetic.
The modelling study was then extended to compounds
1a and 2a, exploring the conformations deriving from
rotation of the single bonds of the phenylpropenyl group.
As expected, this group is very flexible as several confor-
mations were found in a range of 1 kcal/mol above the
global minima; however, the geometry of the bicyclic
system strictly corresponded to those shown in Fig. 1 for
1h and 2h.
In order to confirm the calculated conformation of 1a
and 2a, they were submitted to high field 1H NMR
spectroscopy; the chemical shifts of all the atoms were
assigned as well as almost all the coupling constants of
the vicinal protons in the diazabicyclo[3.3.1]nonane sys-
tem (Table 3). Application of the Altona equation [11]
to the calculated conformations of 1a and 2a yielded, as
weighed averages, the calculated coupling constants also
reported in Table 3. The agreement between calculated
and experimental values ensures that the calculated con-
formations of 1a and 2a represent their solution confor-
mations thus confirming the chair–chair geometry of the
bicyclic system. As the other compounds in the 1 and in
the 2 series differ from 1a and 2a only for the presence
of substituents on the phenyl ring, it can be safely
affirmed that the conformations calculated for the latter
compounds hold also for all compounds in each series.
These results deserve some comments, taking into
account the previous hypothesis [3] of the existence on
m receptors of two hydrophobic pockets (HP-1 and
Table 1
Binding affinities to m, d and k receptors
Comp.
Binding affinities (Ki (nM)) a
m
d
k
1a
2a
2992.0
1391.5
12 00091152
17509144
\50 000
20009180
5. Discussion
a Ki values were calculated with the LIGAND program [8]. Each
value represents the mean9SEM from three experiments, each per-
formed in triplicate (n=3).
The ground terms 1a and the isomer 2a were evalu-
ated for their affinity towards m, d and k receptors

G.A. Pinna et al. / Il Farmaco 55 (2000) 553–562
557
Table 2
Table 3
Inhibition constants of morphine, I, II and compounds 1 and 2
towards m-opioid receptors
Calculated and experimental 1H NMR coupling constants (Hz) a of
the vicinal hydrogen atoms of the bicyclic system of compounds 1a
and 2a
a
Comp.
[3H]-DAMGO (Ki (nM))
J
1a (calc.)
1a (exp.)
2a (calc.)
2a (exp.)
DBN
DBO
DBN
DBO
(1)
(I)
(2)
(II)
1–2ax
1–2eq
1–8a
3.8
2.0
2.4
4.3
5.8
1.6
12.2
5.5
3.7
2.1
2.5
4.2
5.8
1.6
12.2
5.5
3.5
1.5
4.3
1.7
2.4
4.3
5.1
1.7
12.4
5.5
4.3
1.7
2.3
4.3
5.1
1.7
12.5
5.4
4.0
1.5
a
b
c
d
e
f
g
29.0
55.0
25.0
6.2
139.0
48.0
30.0
55.0
13.0
7.66
8.66
5.83
18.0
6.0
160.0
5.1
2.5
n.d. b
5.0
70.0
48.33
246.66
220.0
1000.0
1750.0
2.8
1–8b
n.d. b
6.0
16.3
47.0
14.0
112.0
87.0
8a–7a
8a–7b
8b–7a
8b–7b
5–4ax
5–4eq
5–6a
5.5
2.0
12.0
6.0
3.5
1.5
n.d. b
13.5
6.5
6.0
4.0
Morphine
1.5
2.5
n.d. b
5.0
a See footnote to Table 1.
5–6b
n.d. b
6.0
6a–7a
6a–7b
6b–7a
6b–7b
5.5
HP-2) able to accommodate the endoethylenic bridge of
DBO-I and DBO-II.
2.0
12.0
6.0
n.d. b
13.5
6.5
Considering that, in DBN the preferred chair–chair
geometry of the bicyclic system involves the
trimethylene bridge oriented towards N-3 (Fig. 1), two
explanations are possible for the lower m-affinity of
compounds 1 with respect to the reverted isomers 2.
As the trimethylene bridge of DBN-1 and of DBN-2
are supposed to accommodate, respectively, into HP-1
and HP-2 pockets, steric reasons could determine an
interaction DBN-1–HP-1 less efficient than that DBN-
2–HP-2. The alternative explanation could be related
to the fact that the protonated nitrogen of DBO or
DBN is supposed to interact with the Asp147 carboxy-
late group of the receptor [12]; the additional CH2 in
compounds 1 pointing towards the N-3 hydrogen atom
could disturb this interaction while in compounds 2 the
additional CH2 cannot have any influence on this inter-
action as it is located on the opposite side of the
molecule with respect to the N-9 hydrogen atom.
a The spectra were recorded in CDCl3 solutions on a Bruker
AM500 spectrometer.
b n.d., not determined.
781 (nujol mulls or films). All NMR spectra were taken
1
on a Varian XL-200 NMR spectrometer with H and
13C being observed at 200 and 50 MHz, respectively.
Chemical shifts for 1H and 13C NMR spectra were
reported in l or ppm downfield from TMS [(CH3)4Si].
Multiplicities are recorded as s (singlet), br s (broad
singlet), d (doublet), t (triplet), q (quartet), dd (doublet
of doublets), dt (doublet of triplets), dq (doublet of
quartets), m (multiplet). Elemental analyses were per-
formed by Laboratorio di Microanalisi, Dipartimento
di Scienze Farmaceutiche, Universita` di Padova, Italy
and are within 90.4% of the calculated values. All
reactions involving air or moisture-sensitive compounds
were performed under an atmosphere of argon ‘S’.
Unless otherwise specified, all materials were ob-
tained from commercial suppliers and used without
purification.
6. Experimental
6.1. Chemistry
Flash chromatography was performed using Merck
silica gel 60 (230–400 mesh ASTM). Thin layer chro-
matography (TLC) was performed with Polygram® SIL
N-HR/HV254 precoated plastic sheets (0.2 mm).
Precursors 14 were prepared by standard procedures
(14b,g) or obtained from commercial suppliers (14a).
Melting points (m.p.) were obtained on an elec-
trothermal IA 9100 digital melting point apparatus or
on a Ko¨fler melting point apparatus and are uncor-
rected. IR spectra were recorded on a Perkin–Elmer
6.2. General arylpropenylation procedure for compounds
1a–g and 2a–g
A mixture of the appropriate propionyl-3,9-diazabi-
cyclo[3.3.1]nonane 3 or 4 (2.30 mmol), the required
cinnamyl halide 14 (2.30 mmol) and K2CO3 (2.30
mmol) in acetone or butanone (13.5 ml) was refluxed
for 4–12 h. The inorganic salts were filtered off, the
Fig. 1. Three-dimensional plots of the most populated conformations
of compounds 1h and 2h.

Table 4
Physicochemical properties of compounds 1 and 2
a
Compound
R
Yield (%)
M.p. (°C)
Formula
IR c (w (cm−1)) 1H NMR (l (ppm))
(analysis b)
1a
1b
1c
1d
1e
H
72
34
64
72
76
oil
oil
oil
oil
oil
C19H26N2O
(C,H,N)
C19H25N3O3
(C,H,N)
C19H25ClN2O
(C,H,N)
C19H24Cl2N2O
(C,H,N)
1635
1.16 (t, 3H); 1.40–1.60 (m, 1H); 1.70–1.95 (m, 4H); 2.20–2.40 (m, 4H); 2.70–3.15 (m, 5H);
3.88 (br s, 1H); 4.70 (br s, 1H); 6.20–6.40 (dt, 1H); 6.50 (d, 1H); 7.20–7.40 (m, 5H)
1.17 (t, 3H); 1.50–1.70 (m, 1H); 1.70–1.92 (m, 4H); 2.20–2.40 (m, 4H); 2.65–3.20 (m, 5H);
3.95 (br s, 1H); 4.73 (br s, 1H); 6.40–6.60 (m, 2H); 7.55 (d, 2H); 8.20 (d, 2H)
1.18 (t, 3H); 1.40–1.60 (m, 1H); 1.70–1.93 (m, 4H); 2.20–2.40 (m, 4H); 2.80–3.10 (m, 5H);
3.88 (br s, 1H); 4.68 (br s, 1H); 6.10–6.30 (dt, 1H); 6.50 (d, 1H); 7.20–7.30 (m, 4H)
1.11 (t, 3H); 1.42–1.63 (m, 1H); 1.70–1.90 (m, 4H); 2.20–2.40 (m, 4H); 2.80–3.10 (m, 5H);
4.05 (br s, 1H); 4.65 (br s, 1H); 6.10–6.30 (dt, 1H); 6.40 (d, 1H); 7.10–7.50 (m, 3H)
1.15 (t, 3H); 1.50–1.70 (m, 1H); 1.75–1.95 (m, 4H); 2.22–2.42 (m, 4H); 2.85–3.25 (m, 5H);
3.89 (br s, 1H); 4.73 (br s, 1H); 6.15–6.24 (dt, 1H); 6.40–6.50 (m, 2H) 7.40 (br s, 2H);
7.80 (s, 1H)
4%-NO2
3%-Cl
1350, 1510,
1620
1640
3%,4%-Cl2
3%-NO2, 4%-Cl
1635
C19H24ClN3O3
(C,H,N)
1335, 1524,
1630
1f
2%-NO2, 5%-Cl
2%-Cl, 5%-NO2
H
25
31
36
22
27
36
60
25
30
130–134 a
208–210 a
oil
C19H24ClN3O3 · H 1340, 1520,
Cl (C,H,N) 1630
1.17 (t, 3H); 1.50–1.70 (m, 1H); 1.70–1.95 (m, 4H); 2.23–2.45 (m, 4H); 2.65–3.20 (m, 5H);
3.90 (br s, 1H); 4.72 (br s, 1H); 6.17–6.24 (dt, 1H); 7.05 (d, 1H); 7.30 (dd, 1H); 7.56 (d,
1H); 7.92 (d, 1H)
1.17 (t, 3H); 1.50–1.70 (m, 1H); 1.70–1.95 (m, 4H); 2.25–2.45 (m, 4H); 2.80–3.20 (m, 5H);
3.95 (br s, 1H); 4.72 (br s, 1H); 6.34–6.48 (dt, 1H); 6.95 (d, 1H); 7.53 (d, 1H); 8.03 (dd,
1H); 8.40 (d, 1H)
1.19 (t, 3H); 1.46–1.66 (m, 2H); 1.72–2.20 (m, 4H); 2.21–2.40 (m, 2H); 2.92 (br s, 2H);
3.18 (dd, 1H); 3.50–3.80 (m, 4H); 4.40 (d, 1H); 6.20–6.30 (dt, 1H);6.60 (d, 1H); 7.20–7.40
(m, 5H)
/
1g
2a
2b
2c
2d
2e
2f
C19H24ClN3O3 · H 1345, 1525,
Cl (C,H,N)
1640
0
C19H26N2O
(C,H,N)
1525, 1635
5
4%-NO2
oil
C19H25N3O3
(C,H,N)
1360, 1515,
1630
1.19 (t, 3H); 1.47–1.70 (m, 2H); 1.72–2.20 (m, 4H); 2.21–2.40 (m, 2H); 3.01 (br s, 2H);
3.50–3.70 (m, 5H); 4.37 (d, 1H); 6.30–6.40 (dt, 1H); 6.60 (d, 1H); 7.50 (d, 2H); 8.20 (d,
2H)
562
3%-Cl
oil
C19H25ClN2O
(C,H,N)
1630
1.17 (t, 3H); 1.40–1.60 (m, 2H); 1.70–2.20 (m, 4H); 2.30–2.50 (m, 2H); 2.98 (br s, 2H);
3.10 (dd, 1H); 3.40–3.60 (m, 4H); 4.40 (d, 1H); 6.20–6.40 (dt, 1H); 6.45 (d, 1H); 7.01–7.40
(m, 4H)
3%,4%-Cl2
oil
C19H24Cl2N2O
(C,H,N)
1635
1.17 (t, 3H); 1.40–1.60 (m, 2H); 1.70–2.10 (m, 4H); 2.20–2.40 (m, 2H); 2.89 (br s, 2H);
3.40–3.60 (m, 5H); 4.20 (d, 1H); 6.20–6.30 (dt, 1H); 6.40 (d, 1H); 7.10–7.20 (m, 1H);
7.30–7.50 (m, 2H)
3%-NO2, 4%-Cl
2%-NO2, 5%-Cl
2%-Cl, 5%-NO2
oil
C19H24ClN3O3
(C,H,N)
1330, 1520,
1630
1.19 (t, 3H); 1.42–1.62 (m, 2H); 1.70–2.20 (m, 4H); 2.20–2.40 (m, 2H); 2.92 (br s, 2H);
3.15 (dd, 1H); 3.40–3.60 (m, 4H); 4.40 (d, 1H); 6.20–6.40 (dt, 1H); 6.52 (d, 1H); 7.40–7.60
(m, 2H); 7.80 (s, 1H)
1.17 (t, 3H); 1.42–1.65 (m, 2H); 1.70–2.20 (m, 4H); 2.37 (q, 2H); 2.93 (br s, 2H); 3.12 (dd,
1H); 3.50–3.75 (m, 4H); 4.40 (d, 1H); 6.15–6.30 (dt, 1H); 7.01 (d, 1H); 7.30 (dd, 1H); 7.56
(d, 1H); 7.92 (d, 1H)
1.17 (t, 3H); 1.48–1.68 (m, 2H); 1.72–2.18 (m, 4H); 2.34 (dq, 2H); 2.93 (br s, 2H); 3.15
(dd, 1H); 3.42–3.78 (m, 4H); 4.40 (d, 1H); 6.30–6.50 (dt, 1H); 7.01 (d, 1H); 7.65 (d, 1H);
8.05 (dd, 1H); 8.42 (d, 1H)
130 (dec) a
245 a
C19H24ClN3O3 · H 1340, 1520,
Cl (C,H,N) 1635
2g
C19H24ClN3O3 · H 1340, 1520,
Cl (C,H,N) 1560, 1635
a As the hydrochloride.
b Compounds gave satisfactory analyses within 90.4% of theoretical calculation unless otherwise stated.
c As the free base.

G.A. Pinna et al. / Il Farmaco 55 (2000) 553–562
559
filtrate evaporated and the oily residue purified by flash
chromatography (eluent: CH2Cl2:(CH3)2CO, 9:1) to
give the desired 1a–g and 2a–g as oils; when the stable
corresponding hydrochloride was obtained: see Table 4
for data.
(41 ml) was refluxed for 18 h, and the solvent evaporated.
The oily residue was heated at 170°C for 4 h removing
methanol by distillation to provide crude 8 which was
dissolved in dry ethanol (40 ml) and dry Et2O (40 ml) and
treated at 0–5°C with ethereal hydrochloric acid to
separate 2,4-dioxo-3,9-dibenzyl-3,9-diazabicyclo[3.3.1]-
nonane (8) hydrochloride as a white solid (3.64 g, 40%),
m.p. 173–175°C. The solid crystallized from ethanol as
white needles, m.p. 175–176°C.
6.3. Dimethyl h,h%-dibromopimelate (6)
A mixture of pimelic acid (5) (50 g, 312 mmol) in
thionyl chloride (55 ml) was heated at 40°C for 18 h.
To the irradiated (300 W lamp) solution, bromine
(117.3 g, 733 mmol) was slowly added (ꢀ8 h) under
stirring at 95°C. After the addition of Br2 was complete,
the brown solution was heated at the same temperature
for a further hour and cooled to room temperature (r.t.).
Methanol (180 ml) was added and the resulting reaction
mixture was poured into ice-water (180 ml) and the whole
solution concentrated and extracted with Et2O three
times. The combined organic extracts were washed with
2% NaHSO3, 5% NaHCO3 and H2O, then dried
(Na2SO4), filtered, and concentrated to a yellow oil
(102.78 g, 95.6%, b.p. 96–98°C/0.2) [4] which was used
in the next step without further purification.
Rf 0.58 (petroleum ether 50–70:ethyl acetate, 8:2); IR
(nujol mull, cm−1) w: 1680 (CꢀO), 2200–2400 (N+H);
1
UV umax (log m): 196.7 (4.21), 242.1 (2.86); H NMR
(CDCl3) lH: 1.20–1.50 (m, 1H), 1.80–1.95 (m, 1H),
2.02–2.20 (m, 2H), 2.80–3.02 (m, 2H), 4.11 (br s, 4H),
5.07 (s, 2H), 7.15–7.55 (m, 10H); Anal. C21H23ClN2O2
(C, H, Cl, N).
6.6. 2,4-Dioxo-3-benzyl-3,9-diazabicyclo[3.3.1]nonane
(9)
A solution of 8 · HCl (8.9 g, 24.1 mmol) in ethanol
(215 ml) was hydrogenated at 60 psi at r.t. for 2 h with
10% Pd–C (0.89 g).
The catalyst was filtered off and the solution was
concentrated to afford 9 · HCl (6.75 g, 100%) as white
powder, m.p. 180–182°C.
1
Rf 0.66 (benzene); IR (film, cm−1) w: 1740 (CꢀO); H
NMR (CDCl3) lH: 1.32–1.85 (m, 2H), 1.98–2.20 (m,
4H), 3.79 (s, 6H), 4.24 (t, 2H, J=70 Hz).
Analytical sample was crystallized from ethanol, m.p.
182–184°C.
6.4. Dimethyl cis- and trans-N-benzyl-2,6-
piperidinedicarboxylate (7)
Rf 0.20 (petroleum ether 50–70:ethyl acetate, 6:4); IR
(nujol mull, cm−1) w: 1690 (CꢀO), 2300–2600 (N+H); 1H
NMR (CDCl3) lH: 1.12–1.42 (m, 1H), 1.70–1.88 (m,
1H), 2.01–2.20 (m, 2H), 2.30–2.55 (m, 2H), 4.57 (br s,
2H), 4.95 (s, 2H), 6.80–8.01 (m, 7H).
A solution of 6 (1.50 g, 4.3 mmol) and benzylamine
(1.47 g, 13 mmol) in benzene (5.5 ml) was refluxed for
24 h. After cooling, the salts were filtered off, the solvent
evaporated and the oily residue was distilled at 114–
150°C/0.01–0.1 to give 7 as a yellowish oil (0.61 g, 64%)
[4]. Flash chromatography with 20% ethyl acetate in
petroleum ether afforded in the order trans-7 (0.21 g,
22.8%), as an oil; Rf 0.58 (petroleum ether 50–70:ethyl
The free base 9 was isolated from the HCl salt with
10% Na2CO3 (16 ml) and extracted with CH2Cl2. Evap-
oration of the solvent led to a viscous oil which on
standing solidified, m.p. 92–94°C. Analytical sample:
m.p. 94–96°C (ethyl ether–petroleum ether) [4].
Rf 0.24 (petroleum ether 50–70:ethyl acetate, 6:4); IR
(nujol mull, cm−1) w: 1670 (CꢀO), 3240, 3300 and 3360
acetate, 8:2); IR (film, cm−1) w: 1740 (CꢀO); H NMR
1
(CDCl3) lH: 1.50–1.62 (m, 2H), 1.75–1.98 (m, 4H), 3.65
(q, 2H, J=13.4 Hz), 3.72 (s, 6H), 3.87 (t, 1H, J=5 Hz),
3.90 (t, 1H, J=5Hz), 7.26–7.41 (m, 5H); 13C NMR
(CDCl3) lC: 19.16, 28.36, 51.34, 56.76, 59.02, 127.10,
128.11, 128.93, 137.98, 174.39; and cis-7 (0.33 g, 35.4%),
as a wax; Rf 0.45 (petroleum ether 50–70:ethyl acetate,
1
(NH); H NMR (CDCl3) lH: 1.25–1.50 (m, 1H), 1.65–
1.82 (m, 1H), 1.82–2.05 (m, 4H), 2.50 (br s, 1H, exch with
D2O), 3.79 (br s, 2H), 4.97 (s, 2H), 7.23–7.51 (m, 5H).
6.7. 3-Benzyl-3,9-diazabicyclo[3.3.1]nonane (10)
8:2); IR (nujol mull, cm−1) w: 1740 (CꢀO); H NMR
1
(CDCl3) lH: 1.20–1.50 (m, 2H), 1.75–1.95 (m, 4H),
3.15–3.30 (m, 2H), 3.60 (s, 6H), 3.85 (s, 2H), 7.20–7.40
(m, 5H); 13C NMR (CDCl3) lC: 20.27, 28.47, 51.19,
58.64, 62.01, 126.90, 127.71, 129.15, 136.81, 173.25.
To a stirred mixture of lithium aluminium hydride (7.9
g, 210 mmol) in dry Et2O (269 ml) at 0°C a solution of
9 (7.04 g, 29 mmol) in dry benzene (101 ml) was added
dropwise. The reaction mixture was refluxed for 6 h, then
cooled at 5°C, decomposed with water (27 ml) and kept
at r.t. for 1 h. The salts were filtered off and the filtrate
was dried (Na2SO4) and evaporated to leave a dark oil.
The oil was purified by distillation at 170–178°C/0.8 [4]
to afford the desired compound 10 (5.48 g, 88%) as a
brown oil.
6.5. 2,4-Dioxo-3,9-dibenzyl-3,9-diazabicyclo-
[3.3.1]nonane (8) hydrochloride
A solution of 7 (as a cis+trans mixture, 19.45 g, 67
mmol) and benzylamine (7.15 g, 767 mmol) in toluene

560
G.A. Pinna et al. / Il Farmaco 55 (2000) 553–562
Rf 0.27 (chloroform:methanol, 9:1); IR (film, cm−1
)
DEPT (135°) and HETCOR experiments. Anal.
C10H18N2O (C, H, N).
w: 3340 (NH); 1H NMR (CDCl3) lH: 1.56–1.95 (m,
6H), 2.25–2.43 (m, 2H), 2.75–2.95 (m, 3H, one which
exch with D2O), 2.97–3.08 (br s, 2H), 3.37 (s, 2H),
7.20–7.40 (m, 5H).
6.11. General procedure for arylpropenyl alcohols (13)
The required 3-arylacrylic acid (12b,d) (10 mmol) or
ester (12c,e,f,g) in dry Et2O (30 ml for 12b) or dry
toluene (50 ml for 12c–g) was treated under argon with
1.5 M DIBALH (17.3 ml, 26 mmol) in toluene at
−5°C. The mixture was stirred for 1 h, the temperature
reaching 20°C. The reaction was cooled at 0–5°C,
quenched by dropwise addition of saturated solution of
potassium sodium tartrate (24.3 ml) and stirred
overnight at r.t. To the mixture Et2O (60 ml) was
added, the organic layers separated, washed (H2O),
dried (Na2SO4), the solvent was evaporated and the
crude 13 was purified by flash chromatography or by
crystallization.
6.8. 3-Benzyl-9-propionyl-3,9-diazabicyclo[3.3.1]nonane
(11)
To a chilled solution of 10 (3.65 g, 16.9 mmol) in
dichloromethane (10.3 ml), propionic anhydride (4.5 g,
35 mmol) was added dropwise. When addition was
complete, the mixture was refluxed for 1 h. After cool-
ing at −5°C, 20% NaOH (15 ml) was added and the
mixture was stirred overnight at r.t. and then extracted
with CH2Cl2. The organic layers were dried (Na2SO4),
filtered and evaporated to give 11 as a crude oil which
was purified by distillation at 130°C/0.2 (4.0 g, 87%) [4].
Rf 0.58 (petroleum ether 50–70:ethyl acetate, 6:4); IR
(film, cm−1) w: 1660 (CꢀO); 1H NMR (CDCl3) lH:
1.10–1.30 (m, 4H), 1.50–1.90 (m, 5H), 2.20–2.50 (m,
4H), 2.80–3.00 (m, 2H), 3.38 (ABq, 2H), 3.89 (br s,
1H), 4.69 (br s, 1H), 7.20–7.40 (m, 5H).
6.11.1. 13b
Yield: 34%; m.p. 129–130°C (ether–petroleum ether)
1
(128–129°C [1]); H NMR (CDCl3) lH: 1.91 (br s, 1H),
4.30–4.40 (d, 2H), 6.20–6.40 (m, 1H), 6.50–6.70 (d,
1H), 7.50–7.60 (d, 2H), 8.20–8.30 (d, 2H).
6.9. 9-Propionyl-3,9-diazabicyclo[3.3.1]nonane (3)
A solution of 11 (3.1 g, 11.4 mmol) in ethanol (28 ml)
was hydrogenated at 60 psi at 60°C for 7 h in the
presence of 10% Pd–C (0.6 g). The catalyst was filtered
off, the solution was evaporated and the oily residue
was purified by flash chromatography (CHCl3–MeOH,
9:1) to give 3 as a yellow oil (1.96 g, 94.5%), b.p.
110–115°C/0.2, which on standing, solidified and was
crystallized from ligroine (m.p. 54–56°C) [4].
6.11.2. 13c
Yield: 97%; b.p. 118–120°C/0.2 (125–130°C/0.4 [1]);
1H NMR (CDCl3) lH: 2.30–2.40 (br s, 1H), 4.30–4.40
(d, 2H), 6.20–6.40 (m, 1H), 6.50–6.70 (d, 1H), 7.10–
7.30 (m, 3H), 7.30–7.40 (s, 1H).
6.11.3. 13d
1
Rf 0.34 (chloroform:methanol, 9:1); IR (film, cm−1
)
Yield: 64%; b.p. 134–136°C/0.7; H NMR (CDCl3)
w: 1630 (CꢀO), 3390 (NH); 1H NMR (CDCl3) lH:
1.00–1.20 (m, 4H), 1.60–1.75 (m, 1H), 1.80–2.00 (m,
4H), 2.20–2.65 (m, 3H), 3.05–3.20 (m, 4H), 3.87 (br s,
lH: 1.60 (br s, 1H), 4.30–4.40 (t, 2H), 6.30–6.40 (m,
1H), 6.50–6.70 (d, 1H), 7.20–7.30 (m, 1H), 7.30–7.50
(d, 1H), 7.50 (d, 1H).
1H), 4.19 (br s, 1H, exch with D2O), 4.64 (br s, 1H). 13
C
NMR (CDCl3) lC: 43.78 and 49.16 (CH×2), 170.80
(CꢀO) assigned with DEPT and HETCOR experiments.
6.11.3.1. 13e. Yield: 68%; m.p. 95–97°C (EtOH–H2O);
1H NMR (CDCl3) lH: 1.69 (br s, 1H exch with D2O),
4.38 (dd, 2H), 6.45 (dt, 1H), 6.63 (d, 1H), 7.49 (br s,
2H), 7.86 (s, 1H). Anal. C9H8ClNO3 (C, H, Cl, N).
6.10. 3-Propionyl-3,9-diazabicyclo[3.3.1]nonane (4)
The compound 3 (0.83 g, 4.56 mmol) was heated in
an open round-bottom flask at 150°C for 2 h. The
crude product was chromatographed (silica gel) eluting
with CHCl3–CH3OH (8:2) to give in the order 3 (0.1 g,
12%) and 4 (0.71 g, 88%) as an oil, b.p. 125–130°C/0.4.
6.11.4. 13f
1
Yield: 96%; m.p. 64–66°C (EtOH–H2O); H NMR
(CDCl3) lH: 1.85 (br s, 1H exch with D2O), 4.44 (d,
2H), 6.53 (dt, 1H), 7.03 (d, 1H), 7.53 (d, 1H), 8.00 (dd,
1H), 8.40 (d, 1H).
1
IR (film, cm−1) w: 1630 (CꢀO), 3390 (NH); H NMR
(CDCl3) lH: 1.16 (t, 3H), 1.50–1.70 (m, 2H), 1.80–2.20
(m, 4H), 2.35 (q, 2H), 3.15 (dd, 1H), 3.33 (br s, 2H),
3.65 (dd, 1H), 3.88 (d, 1H), 4.62 (d, 1H), 4.79 (br s, 1H
exch with D2O). 13C NMR (CDCl3) lC: 9.05 (CH3),
18.24, 26.64, 29.48, 29.49, 45.08 and 49.22 (CH2×6),
46.53 and 46.61 (CH×2), 172.58 (CꢀO) assigned by
6.11.5. 13g
1
Yield: 78%; m.p. 85–86°C (EtOH–H2O); H NMR
(CDCl3) lH: 1.66 (br s, 1H exch with D2O), 4.40 (d,
2H), 6.38 (dt, 1H), 7.13 (d, 1H), 7.38 (dd, 1H), 7.58 (d,
1H), 7.93 (d, 1H). Anal. C9H8ClNO3 (C, H, Cl, N).

G.A. Pinna et al. / Il Farmaco 55 (2000) 553–562
561
6.12. General procedures for arylpropenyl chlorides (14)
binding experiments. An aliquot of homogenate was
removed for protein assay [14].
6.12.1. A
A solution of the appropriate 13b,e–g (10 mmol) and
dry pyridine (1 ml) in dry dichloromethane (37 ml) was
added within 1 h at r.t. to a solution of thionyl chloride
(2.7 ml, 37 mmol) in dry dichloromethane (13 ml).
After a further 0.5 h stirring, the mixture was poured
into an aqueous solution of NaHCO3 (6.8 g, 81 mmol)
and the layers were separated and the aqueous phase
was extracted with CH2Cl2. The combined organic lay-
ers were washed with brine, dried (Na2SO4) and evapo-
rated to give 14b,e–g which were purified by flash
chromatography.
6.13.2. Binding studies
Binding assays were carried out as described by
Gillan and Kosterlitz [15] with small modifications.
Displacement experiments were performed in 250 ml
of 50 nM Tris–HCl, pH 7.4 containing 1 nM [3H]-
DAMGO and 1 nM [3H]-DELTORPHINE II to label
m and d receptors in rat brain membranes (18 mg
tissue); 1 nM [3H]-U69 593 was used to label k recep-
tors in guinea pig brain membranes (15 mg of tissue).
To determinate IC50 values (where IC50 is the inhibitor
concentration displacing 50% of the labelled ligand),
the test compounds were added in triplicate to the
binding-assay samples at a minimum of six different
concentrations. After 60 min incubation, bound and
free radioactivity were separated by filtering the assay
mixture through GF/B glass fibre filters (Whatman)
using a MicroMate 196 Cell Harvester (Packard Instru-
ment Company). The filter-bound radioactivity was
counted on Top Count (efficiency 57%) with Micro-
Scint-20 (30 ml in 96-well plates). Non-specific binding
was defined as the binding in the presence of 10 mM
bremazocine: this was always B10% of the total
binding.
1
6.12.1.1. 14b. Yield: 78%; m.p. 58.5–59.5°C; H NMR
(CDCl3) lH: 5.3–5.6 (m, 3H), 6.0–6.2 (m, 1H), 7.5–7.6
(d, 2H).
6.12.1.2. 14e. M.p. 59–61°C [6].
6.12.1.3. 14f. Yield: 96%; m.p. 38–40°C; 1H NMR
(CDCl3) lH: 4.26 (d, 2H), 6.31 (dt, 1H), 7.17 (d, 1H),
7.42 (dd, 1H), 7.60 (d, 1H), 8.00 (d, 1H). Anal.
C9H7Cl2NO2 (C, H, Cl, N).
1
6.12.1.4. 14g. Yield: 98%; m.p. 159–161°C; H NMR
Ki values were calculated from IC50 values using the
Cheng–Prusopff equation [16].
(CDCl3) lH: 4.30 (d, 2H), 6.50 (dt, 1H), 8.10 (dd, 1H),
8.41 (d, 1H). Anal. C9H7Cl2NO2 (C, H, Cl, N).
6.12.2. B
References
A mixture of the appropriate 13c,d (10 mmol) and
concentrated HCl (15 ml) was heated at 78–80°C for 3
h with stirring. The mixture was cooled and extracted
with Et2O. The organic layer was washed several times
with water, dried (Na2SO4) and then evaporated to
provide the crude 14c,d which were purified by bulb to
bulb distillation.
[1] G. Cignarella, E. Occelli, E. Testa, Bicyclic homologs of piper-
azine. VII. Synthesis and analgesic activity of 3-aralkenyl-8-pro-
pionyl-3,8-diaza-bicyclo[3.2.1]octanes, J. Med. Chem. 8 (1965)
326–331.
[2] G. Cignarella, D. Barlocco, M.E. Tranquillini, A. Volterra, N.
Brunello,
G.
Racagni,
Interaction
of
3,8-diazabicy-
clo[3.2.1]octanes with m and d-opioid receptors, Pharmacol. Res.
Commun. 20 (1988) 383–394.
6.12.2.1. 14c. Yield: 77%; b.p. 90°C/0.1; 1H NMR
(CDCl3) dH: 4.2 (d, 2H), 6.2–6.4 (m, 1H), 6.5–6.7 (d,
1H), 7.2–7.4 (m, 4H).
[3] D. Barlocco, G. Cignarella, G. Greco, E. Novellino, Computer
aided structure-affinity relationships in a set of piperazine and
3,8- diazabicyclo[3.2.1]octane derivatives binding to the m opioid
receptor, J. Comput.-Aided Mol. Design 7 (1993) 557–571.
[4] G. Cignarella, G. Maffii, E. Testa, Ricerche sui 3,9-diazabici-
clo[3.3.1]nonani. Stereochimica e attivita` analgesica nei derivati
6.12.2.2. 14d. B.p. 72–74°C/0.01 [13].
6.13. Biology
del
3,9-diazabiciclo[3.3.1]nonanoe
nel
3,8-diazabici-
clo[3.2.1]ottano, Gazz. Chim. Ital. 93 (1963) 226–237.
[5] P. Fadda, D. Barlocco, S. Tronci, G. Cignarella, W. Fratta,
Antinociceptive action of DBO 17 and DBO 11 in mice: two
3,8-diazabicyclo[3.2.1]octane derivatives with selective m opioid
receptor affinity, Naunyn-Schmiedeberg’s Arch. Pharmacol. 356
(1997) 596–602.
[6] D. Barlocco, G. Cignarella, P. Vianello, S. Villa, G.A. Pinna, P.
Fadda, W. Fratta, Synthesis and m-opioid receptor affinity of a
new series of nitro substituted 3,8-diazabicyclo[3.2.1]octane
derivatives, Il Farmaco 53 (1998) 557–562.
6.13.1. Membrane preparation
Male Wistar rats and guinea pigs were decapitated
and the whole brain (minus brainstem, striatum and
cerebellum) was dissected on ice. The tissue was dis-
rupted in a Polytron (setting 5) in 20 vols of 50 nM
Tris–HCl, pH 7.4. The homogenate was centrifuged at
34 000×g for 10 min and the pellet was resuspended in
the same buffer. After 30 min incubation at 37°C the
membranes were centrifuged and pellets were used in
[7] G. Cignarella, E. Testa, R. Pasqualucci, Thermal and chemical
intramolecular N“N acyl-migration in 8-acyl-3,8-diazabici-
clo[3.3.1]octanes, Tetrahedron 19 (1963) 143–148.

562
G.A. Pinna et al. / Il Farmaco 55 (2000) 553–562
[8] P.J. Munson, D. Rodbard, Ligand: a versatile computerized
structurally related to 3,8-diazabicyclo [3.2.1] octanes, J. Med.
Chem., accepted for publication.
[13] H. Ohtaka, M. Mijake, T. Kanazawa, K. Ito, G. Tsukamoto,
Benzylpiperazine derivatives. I. Syntheses and biological activi-
ties of 1-(2,3,4-trimethoxyben-zyl)piperazine derivatives, Chem.
Pharm. Bull. 35 (1987) 2774–2781.
[14] M.M. Bradford, A rapid and sensitive method for the quantita-
tion of microgram quantities of protein utilizing the principle of
protein dye binding, Anal. Biochem. 72 (1976) 248–254.
[15] M.G. Gillan, H.W. Kosterlitz, Spectrum of the m, d and k
binding sites in homogenates of rat brain, Br. J. Pharmacol. 77
(1982) 461–469.
approach for characterization of ligand-binding systems, Anal.
Biochem. 107 (1980) 220–239.
[9] HYPERCHEM™, HyperCube, Inc., Canada.
[10] L. Toma, G. Cignarella, D. Barlocco, F. Ronchetti, Molecular
mechanics and 1H NMR conformational study of 3,8-diazabicy-
clo[3.2.1]octanes and related cis-2,6-dimethylpiperazines active
on opioid receptors, Tetrahedron 48 (1992) 159–166.
[11] C.A.G. Haasnoot, F.A.A.M. de Leeuw, C. Altona, The relation-
ship between proton–proton NMR coupling constants and sub-
stituent electronegativities—I. An empirical generalization of the
Karplus equation, Tetrahedron 36 (1980) 2783–2792.
[12] P. Vianello, A. Albinati, G.A. Pinna, A. Lavecchia, L. Marinelli,
P.A. Borea, S. Gessi, P. Fadda, S. Tronci, G. Cignarella, Synthe-
sis and opioid receptor affinity of 9,10-diazabicy-
clo[4.2.1.12,8]decanes and 2,7-diazabi-cyclo [4.4.0.03,8]decanes
[16] Y.C. Cheng, W.H. Prusoff, Relationship between the inhibition
constant (Ki) and the concentration of inhibitor which causes
50% inhibition (I50%) of an enzymatic reaction, Biochem. Phar-
macol. 22 (1973) 3099–3108.
.
.