Synthesis and pharmacological studies of new hybrid derivatives of fentanyl active at the μ-opioid receptor and I2-imidazoline binding sites

Article abstract of DOI:10.1016/j.bmc.2006.06.007

Two series of fentanyl-derived hybrid molecules bearing potent I₂-imidazoline binding site (IBS) ligands (i.e., guanidine and BU224 moieties) linked with an aliphatic (m = 2, 3, 4, 6, 7, 8, 9 and 12 methylene units) or aromatic spacer were prepared. Their affinities for the μ-opioid receptors and for the I₂-IBS were determined through competition binding studies on human postmortem brain membranes. Whereas the BU224 hybrid molecules bound to the μ-opioid receptor and the I₂-IBS in the micromolar to low micromolar range, the alkaneguanidine series exhibited remarkable affinities in the nanomolar range for both receptors. [³⁵S]GTPγS functional assays were performed on human postmortem brain membranes with selected ligands from each series (4f and 8g) showing the highest dual affinity for the μ-opioid receptor and I₂-IBS affinities. Both compounds displayed agonist properties: at the μ-opioid receptor for the alkaneguanidine derivative 4f (spacer: six methylene units) and at a G-protein coupled receptor (GPCR) which remains to be determined for 8g. The lack of analgesic properties of 4f in vivo (i.e., hot plate and writhing tests in mice), discordant with the good in vitro binding data (K_i μ = 1.04 ± 0.28 nM, K_i I₂ = 409 ± 238 nM), may possibly be due to the low intrinsic efficacy of the compound. Alternatively, a low access to the central nervous system for this kind of hybrid molecules cannot be ruled out. Two new compounds reported here (9f and 13), which were not dual acting, are worth mentioning for their outstanding binding affinities; 9f bound to the μ-opioid receptor with a picomolar affinity (K_i = 0.0098 ± 0.0033 nM), whereas 13 presented an I₂-IBS affinity (K_i = 18 ± 11 nM) similar to the reference compound BU224.

Full text of DOI:10.1016/j.bmc.2006.06.007

Bioorganic & Medicinal Chemistry 14 (2006) 6570–6580
Synthesis and pharmacological studies of new hybrid derivatives
of fentanyl active at the l-opioid receptor
and I2–imidazoline binding sites
Christophe Dardonville,a,* Cristina Fernandez-Fernandez,a Sarah-Louise Gibbons,a
Gary J. Ryan,a Nadine Jagerovic,a Ane M. Gabilondo,b
J. Javier Meanab and Luis F. Calladob
aInstituto de Quı´mica Me´dica, CSIC, Juan de la Cierva 3, E-28006 Madrid, Spain
bDepartamento de Farmacologı´a, Universidad del Paı´s Vasco/Euskal Herriko Unibertsitatea, Leioa, E-48940 Bizkaia, Spain
Received 29 March 2006; revised 11 May 2006; accepted 6 June 2006
Available online 23 June 2006
Abstract—Two series of fentanyl-derived hybrid molecules bearing potent I2–imidazoline binding site (IBS) ligands (i.e., guanidine
and BU224 moieties) linked with an aliphatic (m = 2, 3, 4, 6, 7, 8, 9 and 12 methylene units) or aromatic spacer were prepared. Their
affinities for the l-opioid receptors and for the I2–IBS were determined through competition binding studies on human postmortem
brain membranes. Whereas the BU224 hybrid molecules bound to the l-opioid receptor and the I2–IBS in the micromolar to low
micromolar range, the alkaneguanidine series exhibited remarkable affinities in the nanomolar range for both receptors. [35S]GTPcS
functional assays were performed on human postmortem brain membranes with selected ligands from each series (4f and 8g) show-
ing the highest dual affinity for the l-opioid receptor and I2–IBS affinities. Both compounds displayed agonist properties: at the
l-opioid receptor for the alkaneguanidine derivative 4f (spacer: six methylene units) and at a G-protein coupled receptor (GPCR)
which remains to be determined for 8g. The lack of analgesic properties of 4f in vivo (i.e., hot plate and writhing tests in mice),
discordant with the good in vitro binding data (Ki l = 1.04 0.28 nM, Ki I2 = 409 238 nM), may possibly be due to the low intrin-
sic efficacy of the compound. Alternatively, a low access to the central nervous system for this kind of hybrid molecules cannot be
ruled out. Two new compounds reported here (9f and 13), which were not dual acting, are worth mentioning for their outstanding
binding affinities; 9f bound to the l-opioid receptor with a picomolar affinity (Ki = 0.0098 0.0033 nM), whereas 13 presented an
I2–IBS affinity (Ki = 18 11 nM) similar to the reference compound BU224.
ꢀ 2006 Elsevier Ltd. All rights reserved.
1. Introduction
possible new therapy that aims at reducing tolerance
developed along with the treatment by opioid drugs
led us to investigate a new series of dual acting drugs
that interact with l-opioid receptor and IBS simulta-
neously albeit separately. Among the advantages of
the multiple target ligand approach compared with the
co-administration of multiple drugs are a better patient
compliance, a more practical use in the clinical settings
and a lower risk of drug–drug interactions.9
Different studies have revealed that some agents that do
not interact directly with opioid receptors are able to
regulate the pharmacological actions of opioid ligands.
In particular, involvement of imidazoline binding site
(IBS) ligands1 in the modulation of analgesia2 and with-
drawal effects of morphine has been observed in mice
and rats.3–7 This is the case, for example, of the putative
endogenous IBS ligand agmatine (Chart 1) which
enhances analgesic action of morphine and inhibits tol-
erance and dependence to opioids.8 Our interest in a
The existence of a chemical entity displaying concomi-
tant opioid receptor and I2–IBS affinities was unknown
in the literature at the start of our studies.9,10 In our
previous work, two series of hybrid molecules incorpo-
rating a fentanyl core (i.e., opioid moiety) and guani-
dine-like moieties (i.e., IBS ligand) were synthesized
and pharmacologically evaluated (Chart 1).10,11 All of
these hybrid structures showed moderate to high affinity
towards l-opioid receptor, whereas only few of them
Keywords: I2–Imidazoline binding site affinity; l-opioid agonist; Alk-
ane guanidine; Analgesia; Dual acting drug; Opioid tolerance; Opioid
withdrawal; Human brain; BU224.
*
Corresponding author. Tel.: +34 91 5622900; fax: +34 91 5644853;
e-mail: dardonville@iqm.csic.es
0968-0896/$ - see front matter ꢀ 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2006.06.007

C. Dardonville et al. / Bioorg. Med. Chem. 14 (2006) 6570–6580
6571
probe further these findings, three new derivatives with
an alkyl spacer (3f, 5f and 7f)] were synthesized and
evaluated. For these series, an extensive modulation of
the chain length of the alkyl spacer between the two
pharmacophoric groups, fentanyl and guanidine, is pre-
sented (Scheme 1, Table 1). In addition, a more rigid
analogue with a m-xylene spacer (9f) was also prepared
and evaluated in order to explore possible structural
needs for these hybrid molecules.
O
N
N
Fentanyl
H
N
NH2
NH
H2N
N
N
Besides, we also wish to report the synthesis of new com-
pounds (2g, 4g, 6g and 8g) that display the combination
of the potent and selective I2–IBS ligand BU224 and the
l-opioid ligand fentanyl. BU2241 (Chart 1) is a selective
I2–IBS ligand which has also demonstrated interesting
spinal antinociceptive properties in vitro.12
HN
Agmatine
BU224
O
NH
spacer
N
NH
NH2
The affinity of the two series of compounds presented
here was evaluated at the I2–IBS and at the l-opioid
receptor. In addition, the functional activity of two
hybrid molecules was also investigated through
[35S]GTPcS assays on postmortem human frontal cortex
membranes. We chose to use membranes from human
brain which could be more relevant from a therapeuti-
cally vantage point. The analgesic potential of the most
promising dual ligand (4f) was evaluated in mice by hot
plate and writhing tests.
N
Spacer: phenyl, n-alkyl
Chart 1.
had high affinity for I2–IBS.11 Of special interest was the
structure–activity relationship for a series of agmatine-
like hybrid molecules, the methylene linker length of
which seemed to correlate well with I2–IBS affinity. To
O
N
H
X
N
NHBoc
X
H2N
NHR
ii
N
i
R= Boc
X
R= H
1a (83%)
2a (95%)
3a (87%)
4a (93%)
5a (99%)
6a (81%)
7a (77%)
8a (90%)
1b (55%)
2b (68%)
3b (56%)
4b (88%)
5b (60%)
6b (82%)
7b (50%)
8b (50%)
1: m = 2
2: m = 3
3: m = 4
4: m = 6
5: m = 7
6: m = 8
7: m = 9
8: m = 12
(CH2)m
9
9a (74%)
9b (37%)
O
X
N
NHR
N
iii
iv
R= H (. 2TFA)
R= Boc
1d: m = 2 (65%)
2d: m = 3 (64%)
3d: m = 4 (99%)
4d: m = 6 (91%)
5d: m = 7 (91%)
6d m = 8 (90%)
7d: m = 9 (75%)
8d: m = 12 (70%)
9d (99%)
1c (75%)
2c (53%)
3c (53%)
4c (73%)
5c (79%)
6c (56%)
7c (92%)
8c (84%)
9c (45%)
˚
Scheme 1. Reagents and conditions: (i) diamine (5 equiv), Boc2O (1 equiv), CHCl3, rt; (ii) NaBH3CN, molecular sieves (3 A), MeOH, rt; (iii)
(EtCO)2O, DMAP, pyridine, CH2Cl2, rt; (iv) CF3CO2H, CH2Cl2, 30 min.

6572
C. Dardonville et al. / Bioorg. Med. Chem. 14 (2006) 6570–6580
Table 1. Guanidine hybrid molecules. Affinity data [Ki (nM)] for
I2–IBS and l-opioid receptors
O
O
RO
MeO
b
59%
c
O
78%
N
NH
N
O
X
N
N
11
10: R = Me
H
NH2 . CF3CO2H
a
96%
N
R = H
O
O
Compound Spacer (X) [3H]-2BFI I2 [3H]-DAMGO l na
MeO
RO
d
Idazoxan
—
—
28 11
8593 738
437 228
1920 996
>10,000
ndb
5
4
3
3
3
3
3
3
3
3
6
47-75%
Fentanyl
2.9 1.5
433 83
23 4.5
N
N
CN
N
1fc
2fc
3f
4fc
5f
6fc
–(CH2)2–
–(CH2)3–
–(CH2)4–
–(CH2)6–
–(CH2)7–
–(CH2)8–
–(CH2)9–
HN
0.59 0.18
1.04 0.28
0.37 0.19
37 9.7
12
13: R = Me
14: R = H
e
67%
409 238
6627 3106
126 72
Scheme 2. Reagents and conditions: (a) SOCl2, MeOH; (b) 30% H2O2,
AcOH, 80 ꢁC; (c) KCN, BzCl, H2O, 0 ꢁC; (d) H2NCH2CH2NH2, P2S5
cat, PhMe, reflux; (e) LiOH, H2O–THF, rt.
7f
58 46
6.5 3.0
>10,000
26
6
477 75
8fc
9f
–(CH2)12
–
m-Xylene
0.0098 0.0033
a Values expressed as means standard error mean of n experiments.
b Not determined.
c From Ref. 11.
acid 14 with amines 2d, 4d, 6d and 8d in the presence
of Mukayama’s reagent afforded the corresponding hy-
brid molecules 2g, 4g, 6g and 8g, respectively. The diffi-
culty of the silica purification was evidenced by the
presence of multiple by-products. Isolation of a signifi-
cant quantity of 1-methyl-pyridinium adducts of the
amines (2d, 4d and 8d) and of further substitution prod-
ucts of the imidazoline nitrogen might account for the
low yield of these couplings (8–46%).
2. Results
2.1. Chemistry
The amines 1d–9d were synthesized in four steps starting
from the corresponding diamines 1–9 and following a
procedure previously described by us.11 In summary,
the reductive amination (NaBH3CN, pH 7–8) of 1-phen-
ethyl-4-piperidone with the mono-protected diamines
1a–9a afforded amines 1b–9b which were subsequently
acylated with propionic anhydride to afford the corre-
sponding propanamides 1c–9c. Treatment with
CF3CO2H afforded the amines 1d–9d as their trifluoro-
acetate salts (Scheme 1).
The NMR spectra of the propanamide compounds
showed characteristic duplicated signals corresponding
to the observation of the amide cis–trans isomerism
around the N–CO bond as previously reported.10
2.2. Pharmacology
2.2.1. Binding affinity. The binding affinity of the new
compounds (3f, 5f, 7f, 9f, 2g, 4g, 6g and 8g) and 1f,
2f, 4f, 6f and 8f11 was evaluated through competition
binding assays against the selective I2–IBS radioligand
[3H]-2-BFI and the selective l-opioid ligand [3H]DAM-
GO following the procedure previously described.10,15–17
The assays were performed in membranes from post-
mortem human frontal cortex, a brain area that shows
an important density of I2–IBS and l-opioid receptors.
Drug competition studies were performed with either
[3H]-2-BFI (1 nM) or [3H]-DAMGO (2 nM) in the ab-
sence or presence of various concentrations of compet-
ing drugs (10ꢀ11–10ꢀ3 M, nine concentrations). Non-
specific binding was estimated in the presence of
10ꢀ3 M idazoxan in experiments with [3H]-2-BFI, or
with 10ꢀ4 M naloxone in [3H]-DAMGO assays. The
opioid agonist fentanyl and the I2–IBS-selective ligand
idazoxan were used as references (Tables 1 and 2).
The BU224 moiety (14) was prepared in five steps from
the commercially available quinoline precursor (Scheme
2). Esterification of methylquinoline-6-carboxylate with
SOCl2/MeOH yielded 10 which was converted to the
cyano-derivative 12 in two steps applying the procedure
that Kaneko et al. used for the synthesis of methyl
2-cyanoquinoline 5-carboxylate.13 Formation of the
imidazoline nucleus was accomplished by heating the
cyanide 12 in the presence of ethylenediamine and a
catalytic amount of P2S5 in refluxing toluene to afford
13. Mild hydrolysis of the methyl ester 13 by treatment
with LiOH afforded 14. This reaction has to be carefully
monitored since hydrolysis of the imidazoline moiety is
a possible competitive reaction.
Two series of hybrid molecules were prepared starting
from the diamines 1d–9d (Scheme 3). The guanidines
were introduced with the classical Boc-protected thio-
urea reagent14 which allows the purification of the
Boc-protected guanidines by SiO2 chromatography.
Removal of the Boc-protecting group with CF3CO2H
afforded the guanidine hybrids 1f–9f. Coupling of the
Guanidine hybrid molecules. Table 1 shows the results of
binding of the guanidine hybrid series of the newly syn-
thesized compounds (3f, 5f, 7f and 9f) and previously
reported molecules (1f, 2f, 4f, 6f and 8f). Regarding

C. Dardonville et al. / Bioorg. Med. Chem. 14 (2006) 6570–6580
6573
O
NR
X
b
N
N
H
NHR
N
S
R= H (. 2TFA)
R= Boc
BocHN
NHBoc
a
X
1f: m = 2 (69%)
2f: m = 3 (98%)
3f: m = 4 (56%)
4f: m = 6 (62%)
5f: m = 7 (97%)
6f: m = 8 (74%)
7f: m = 9 (98%)
8f: m = 12 (60%)
1e (76%)
2e (50%)
3e (58%)
4e (57%)
5e (63%)
6e (95%)
7e (64%)
8e (84%)
(CH2)m
O
X
N
NH2 .2TFA
N
9f (99%)
9e (48%)
1d-9d
14
c
O
O
N
(CH2)m NH
N
N
2g: m = 3 (9%)
4g: m = 6 (8%)
6g: m = 8 (46%)
8g: m = 12 (8%)
NH
N
Scheme 3. Reagents and conditions: (a) HgCl2, Et3N, CH2Cl2, rt; (b) TFA, CH2Cl2; (c) 2-chloro-1-methylpyridinium iodide, Et3N, CH2Cl2, rt.
l-opioid receptor, all the molecules bound with affinities
in the nanomolar range; the best l-opioid ligands being
obtained with spacers of four (3f), six (4f) and seven (5f)
methylene units with affinity of 0.59, 1.04 and 0.37 nM,
respectively. Either increasing (m = 8, 9, 12) or decreas-
ing (m = 2, 3) the length of the spacer afforded less po-
tent ligands. On the contrary, increasing the length of
the methylene spacer improved I2–IBS binding affinity.11
Table 2. BU224 hybrid molecules. Affinity data [Ki (nM)] for I2–IBS and l-opioid receptors
O
N
(CH2)m NH
O
N
N
N
. CF3CO2H
N
H
Compound
–(CH2)m–a
[3H]-2BFI
I2
[3H]-DAMGO
nb
l
Idazoxan
Fentanyl
—
—
—
—
—
3
28 11
ndc
5
4
2
3
3
3
3
3
3
8593 738
9.8 0.3d
18 11
2.9 1.5
nd
BU224
13
14
>10,000
>10,000
6142 2123
2168 66
339 35
545 179
588 285
875 713
323 270
>10,000
2g
4g
6
6g
8g
8
12
547 316
a m, number of methylene units in the spacer.
b Values expressed as means standard error mean of n experiments.
c Not determined.
d From Ref. 18.

6574
C. Dardonville et al. / Bioorg. Med. Chem. 14 (2006) 6570–6580
Noteworthy is the excellent l-opioid affinity observed
for the hybrid compound 9f with the m-xylene spacer
(Ki = 0.0098 0.0033 nM). However, this compound
was not a l/I2 dual ligand since it lacked I2–IBS binding
affinity.
17.1 lM) in this assay. It should be noted that ‘affinities
of tritiated agonists for receptors in postmortem human
brain are usually several orders of magnitude higher
than the potencies of these agonists stimulating
[35S]GTPcS binding both in rat and in postmortem hu-
man brain membranes.’20
BU224 hybrid molecules. The affinity for the l-opioid
receptor was lower in this series, affording only low
micromolar binding for hybrid molecules with linkers
of 8 and 12 methylene units (6g and 8g, Table 2). Shorter
spacers (3 or 6 methylene units, 2g and 4g, respectively)
afforded low affinity l-opioid ligands. In this series, the
best affinities for I2–IBS were also observed for the lon-
gest linkers 4g and 8g (m = 6 and 12) with the exception
of 6g (m = 8) which did not bind to I2–IBS. The most
balanced I2/l binding affinity of this series was obtained
with the 12 methylene spacer hybrid molecule 8g
(Table 2).
Compound 8g showed a marked increase in nucleotide
binding compared to basal level (+125%). However, this
effect was not reverted in the presence of naloxone
10ꢀ5M (Fig. 1b). Several classes of GPCRs present in
membranes from postmortem human frontal cortex
(e.g., a2–adrenoceptors, CB1–cannabinoid, or 5HT1)
could potentially be activated by this compound and
explain the observed [35S]GTPcS binding activity.
2.2.3. In vivo assays. Because 4f proved to be a l-opioid
agonist in the binding experiments, we decided to evalu-
ate its analgesic potential in vivo in the hot plate and
writhing tests in mice (Table 3). Unfortunately, this
compound did not produce significant analgesia in both
assays up to 40 mg/kg ip. The experiments were discon-
tinued because of toxicity problems (27.3% death at
40 mg/kg). The lack of analgesic activity of 4f in the
hot plate and writhing tests may possibly relate to its
low efficacy in the functional assay. Alternatively, a
poor blood–brain barrier (BBB) penetration of this
dicationic compound and the consequent lack of activa-
tion of the opioid receptors in the CNS might also
explain the in vivo results.
Besides the binding studies performed on the BU224 hy-
brid molecules, two new analogues of BU224 (13 and
14) were evaluated as I2–IBS ligands. Noteworthy was
the excellent I2–IBS affinity displayed by 13, a new 6-
methyl ester analogue of BU224 (Ki = 18 11 nM vs
9.8 0.3 nM for BU22418). Interestingly, hydrolysis of
the ester group of 13 to give the 6-carboxylic acid ana-
logue (14) led to a 30-fold loss in binding affinity for
I2–IBS.
2.2.2. [35S]GTPcS functional assay. The therapeutic po-
tential of these new families of I2–IBS/l-opioid hybrid
compounds as possible l-opioid modulators depends
on their functional activity (i.e., agonist/antagonist) be-
cause l-opioid agonists would have analgesic properties,
whereas antagonists would not. Hence, we decided to
assess the functional activity of two of these new hybrid
molecules using the [35S]GTPcS assay in postmortem
human brain samples.19 This assay constitutes a func-
tional measure of the interaction of the receptor and
the G-protein, the first step in activation of G-protein
coupled receptors (GPCRs). It is a useful tool to distin-
guish between agonists (increasing the nucleotide bind-
ing), inverse agonists (decreasing the nucleotide
binding), and neutral antagonists (not affecting the
nucleotide binding) of GPCRs.
3. Discussion
Our approach towards the management of opioid-in-
duced tolerance developed during treatment with opioid
drugs like fentanyl led us to design and synthesize sever-
al hybrid molecules that are able to interact with both
I2–IBS and l-opioid receptors. The first goal of this
study was the development of new agmatine-like hybrid
analogues modulating the chain length of the methylene
spacer (1f–8f, Table 1) in order to find the optimum dis-
tance between the pharmacophoric units for a balanced
affinity on I2–IBS and l-opioid receptors. In this series,
we clearly found that the best spacer for l-opioid affinity
should have between four and seven methylene units (3f,
4f and 5f). Longer (6f–8f) or shorter (1f–2f) spacers lead
to a decrease in binding affinity for these receptors. On
the other hand, the best binding affinity for I2–IBS
was observed with the longest spacers (7f and 9f, 9
and 12 methylene units, respectively), confirming our
previous findings on this kind of alkaneguanidine deriv-
atives.11,15 We previously observed11 that the presence
of the phenethylpiperidyl propanamide pharmacophore
of the fentanyl structure could lead to a synergistic effect
for the binding of alkaneguanidinium moieties to I2–IBS
since alkaneguanidines alone (i.e., monoguanidinium)
show very poor I2–IBS affinity.15 This synergistic effect
seems to exist also for the binding to l-opioid receptors.
Accordingly, we observed that amines 2d and 3d (as
their TFA salts) showed no binding to these receptors
(Kil = 4689 2232 and >10,000 nM, respectively),
whereas their guanidinium derivatives 2f and 3f had Ki
In order to confirm the nature of the functional activity
of both families of dual compounds, we selected in each
family one molecule that presented the ‘best dual affini-
ty.’ Hence, the intrinsic activity of 4f and 8g, which
showed the highest l-opioid receptor along with an
acceptable I2–IBS affinity, was investigated with
[35S]GTPcS binding experiments on membranes of post-
mortem human frontal cortex.
The guanidine derivative 4f induced an increase in the
[35S]GTPcS binding of 25% compared to basal level
(Fig. 1a) revealing agonist properties. This effect was
reverted in the presence of naloxone 10ꢀ5M indicating
agonist properties at the human l-opioid receptor.
The apparent low potency of this compound
(EC50 = 4.21 0.57 lM) compares favourably with that
of the l-opioid agonist DAMGO (EC50 = 77.1

C. Dardonville et al. / Bioorg. Med. Chem. 14 (2006) 6570–6580
6575
a
200
150
100
50
4f
4f+naloxone
DAMGO
-17.5 -15.0 -12.5 -10.0 -7.5
-5.0
-2.5
Log [drug] M
300
200
100
0
b
8g
8g+naloxone
DAMGO
-17.5 -15.0 -12.5 -10.0 -7.5
-5.0
-2.5
Log [drug] M
Figure 1. Concentration–response curves of the stimulation of [35S]GTPcS-specific binding to postmortem human brain membranes by 4f (1a) or 8g
(1b) in the absence (j) or in presence (m) of the l-opioid antagonist naloxone, and by the selective l-opioid agonist DAMGO (d).
Table 3. Antinociceptive effect of compound 4f and morphine on hot
plate and acetic acid writhing tests in mice
spatial positioning of the guanidinium group to bind
to the l-opioid receptor. On the contrary, 9f was devoid
Compound Dose
(mg/kg, ip)
Hot plate
(% MPE)a
Writhing
of I2–IBS affinity (Ki > 10,000 nM) compared to its m-
phenyl analogue (Ki = 1890 499),10 indicating that
more conformational flexibility of the guanidine moiety
around the phenyl ring was detrimental for the binding
of these aromatic derivatives to I2–IBS.1 Compound 9f
is not a dual ligand but has impressive picomolar affinity
for the l-opioid receptor. Further in vitro and in vivo
studies of this compound are warranted even though
this goes beyond the present work focusing on dual act-
ing ligands.
n
(% inhibition)
n
4f
1
ꢀ5.6 3.8
ꢀ1.2 9.7
2.8 5.7
6
6
6
5
6
6
6
10
20
40
7.6 22.2
28.0 23.6
10
11
8.9 12.1
Morphine
2.5
25.5 18.0
51.1 12.6**
92.3 7.7***
5
10
100.0 0.0**
9
Values are expressed as means SEM from n mice per group.
a Values correspond to data obtained 30 min after compound 4f or
morphine administration.
The second family of hybrid molecules synthesized, the
BU224 derivatives 2g, 4g, 6g and 8g, only afforded
micromolar ligands for both receptors, the longest spac-
ers giving the best affinities for both receptors. These
trends may reveal the importance of steric factors for
the binding of these hybrid molecules to l-opioid recep-
tors and I2–IBS.
**p < 0.01; ***p < 0.001 versus saline (ANOVA followed by the
Student-Newman–Keuls test).
of 23 and 0.59 nM, respectively. These results altogether
indicate that in this series of hybrid molecules the pres-
ence of the guanidinium cation is crucial for the binding
to l-opioid receptors, while the fentanyl skeleton con-
fers I2–IBS affinity to the alkylguanidinium moiety.
The binding affinity observed for the new 6-substituted
analogues of BU224 (13 and 14) is especially interesting
since they demonstrate that the incorporation of a meth-
yl ester group in this position of the quinoline ring of
BU224 hardly affects its binding to I2–IBS (Table 2).
In contrast, a carboxylic acid moiety in that position
leads to a 60-fold decrease in binding affinity. Since
the carboxylate group is likely to be deprotonated under
physiological conditions, these differences may indicate
The remarkable l-opioid affinity of the m-xylene deriv-
ative 9f (Ki = 0.0098 0.0033 nM), which represents a
1000-fold increase compared to its m-phenyl analogue
(Ki = 7.8 2.5 nM),10 seems to indicate the needs of
more flexibility for this ligand to fit into the l-opioid
receptor active site. The m-xylene spacer may alter the

6576
C. Dardonville et al. / Bioorg. Med. Chem. 14 (2006) 6570–6580
that the presence of a negatively charged substituent in
the 6-position of the quinoline ring has a detrimental ef-
fect on the binding to I2–IBS.
over a 2 h period to a 0.25 M solution of diamine 1–9
(5 equiv; typical scale: 250–500 mmol) in CHCl3 cooled
with an ice-bath. The reaction mixture was stirred over-
night at room temperature and filtered. The filtrate was
concentrated under vacuum and the resulting oil dis-
solved in EtOAc (400 mL) was washed with half-satu-
rated brine (3· 150 mL), dried (MgSO4) and
concentrated under vacuum to afford pure mono-Boc-
protected diamine 1a–9a. Further purification by silica
chromatography was necessary for the diamine with
more than six methylene units and 1a.
4. Conclusion
The synthesis of new hybrid molecules incorporating the
fentanyl skeleton and a guanidine moiety linked with an
alkyl or aromatic spacer allowed to gain insights into the
SAR of this family of compounds. Hence, we verified
that the presence of the guanidine and the N-phenethyl
piperidyl propanamide core was necessary for these hy-
brid compounds to bind to both I2–IBS and l-opioid
receptors. The new family of BU224-derived dual com-
pounds only produced moderate affinity ligands for
these receptors, possibly due to a bigger steric hindrance
of the BU224 core.
N-(tert-Butoxycarbonyl)-1,6-hexanediamine (4a). Col-
ourless oil (93%). Rf = 0.59 (50% MeOH–NH3(satd) in
1
CH2Cl2); H NMR (CDCl3, 200 MHz) d 4.70 (br, 1H,
NH), 3.03 (m, 2H), 2.62 (t, 2H, J = 6.6 Hz), 1.38 (s,
9H), 1.50–1.20 (m, 10H); 13C NMR (CDCl3, 50 MHz)
d 155.7, 78.3, 41.6, 40.0, 33.2, 29.6, 28.0, 26.2, 26.1;
LRMS (ES+) m/z 217 [(M+H), 100].
To conclude, the guanidinoalkyl fentanyl derivatives
represent an interesting family of l-opioid/I2–IBS hy-
brid ligands that showed agonist properties (i.e., 4f) at
the human brain l-opioid receptor. Even though 4f
was devoid of analgesic action in vivo, possibly due to
its low efficacy and/or to unfavourable pharmacokinetic
properties, these hybrid molecules represent attractive
lead compounds.
N-(tert-Butoxycarbonyl)-1,3-phenylenedimethanamine
(9a). Purification by chromatography on silica with
CH2Cl2/MeOH–NH3 (90:10) yielded 9a as a colourless
oil (96%); 1H NMR (CDCl3): d 7.34–7.13 (m, 4H),
5.34 (br s, 1H, NH), 4.24 (m, 2H), 3.74 (s, 2H), 2.73
(br s, 2H), 1.42 (s, 9H); 13C NMR (CDCl3): d 161.9,
142.4, 139.3, 128.5 (2· CH), 126.0, 125.9, 79.1, 45.7,
44.3, 28.2; LRMS (ES+) m/z 237 [(M+H), 80%].
5. Experimental
5.1. Chemistry
B. Reductive amination. NaBH3CN (1.5 equiv) was add-
ed to a 0.2 M solution of 1-phenethyl-4-piperidone
(1 equiv; typical scale: 15–20 mmol), mono-Boc-protect-
˚
ed diamine 1a–8a (1.2 equiv) and 3 A molecular sieves in
Reactions were monitored by TLC using pre-coated sil-
ica gel 60 F254 plates. Chromatography was performed
either with silica gel 60 PF254 (particle size 40–63 lm) or
with 50 g SI Isolute prepacked columns. All reactions
requiring anhydrous conditions or an inert atmosphere
were performed under a positive pressure of N2. Dry
CH2Cl2 was obtained by distillation over CaCl2. 1H
and 13C NMR spectra were recorded on a Varian Gem-
ini 200, Bruker Advance 300 or Varian Inova 400 spec-
dry MeOH. The pH of the reaction was adjusted to 6–7
with AcOH during the course of the reaction. When the
reaction was complete (typically 2–5 h), the mixture was
filtered on a pad of Celite. The filtrate was concentrated
under vacuum and the crude product was purified by
chromatography.
tert-Butyl
[6-(1-phenethylpiperidin-4-ylamino)hex-
yl]carbamate (4b). Purification by chromatography with
CH2Cl2–Et3N (1%)/MeOH–NH3(satd) (100:0 ! 90:10)
1
trometer. Chemical shifts of the H NMR spectra were
1
afforded the product as an orange solid (64–88%). H
internally referenced to the residual proton resonance
of the deuterated solvents: CDCl3 (7.26 ppm), D2O
(d 4.6 ppm), CD3OD (3.49 ppm) and DMSO (d
2.49 ppm). Melting points were determined with a Reic-
hert-Jung Thermovar apparatus and are uncorrected.
Elemental analysis was performed on a Heraeus
CHN–O Rapid analyser. Analytical results were with-
in 0.4% of the theoretical values unless otherwise not-
ed. Electrospray mass spectra were recorded on an
MSD–Serie 1100 Hewlett Packard apparatus.
NMR (CDCl3, 200 MHz) d 7.00 (m, 5H), 4.85 (br,
1H), 2.90 (m, 2H), 2.75 (m, 2H), 2.66–2.52 (m, 2H),
2.50–2.20 (m, 5H), 2.11 (br, 1H), 1.96–1.60 (m, 4H),
1.40–1.04 (br m, 17H); 13C NMR (CDCl3, 75 MHz) d
155.7, 140.1, 128.3, 127.9, 125.5, 78.4, 60.2, 54.6, 52.1,
46.2, 40.1, 33.4, 32.1, 29.72, 29.66, 28.1, 26.3, 26.1;
LRMS (ES+) m/z 404 [(M+H), 100%]. Anal.
(C27H45N3O3) C, H, N.
tert-Butyl {3-[(1-phenethylpiperidin-4-ylamino)meth-
yl]benzyl}carbamate (9b). NaBH3CN (1.4 g, 3.6 mmol);
1-phenethyl-4-piperidone (331 mg, 2.4 mmol); mono-
Boc-protected amine 9a (681 mg, 2.9 mmol). Purifica-
tion by flash chromatography on silica with CH2Cl2/
MeOH–NH3 (100:0 ! 95:5)] yielded 9b as a colourless
oil (49%); 1H NMR (CDCl3): d 7.29–7.16 (m, 9H),
5.12 (br s, 1H, NH), 4.26 (m, 2H), 3.80 (br s, 2H),
3.06 (m, 2H), 2.84 (m, 2H), 2.70 (m, 2H), 2.28 (br t,
2H, J = 10.7 Hz), 1.99 (br d, 2H, J = 11.5 Hz), 1.61
Synthetic and spectroscopic data for 1a, 3a, 5a–8a, 1b,
3b, 5b–8b, 1c–3c, 5c–8c, 1d–3d, 5d–8d, 1e, 3e, 5e–8e,
1f, 3f, 5f–8f, 2g, 6g and 8g are available online as Sup-
plementary material.
5.2. General synthetic procedures
A. Mono-protection of diamines with Boc. A 0.5 M solu-
tion of Boc2O (1 equiv) in CHCl3 was added dropwise

C. Dardonville et al. / Bioorg. Med. Chem. 14 (2006) 6570–6580
6577
(dq, 2H, J = 2.9 Hz, J = 10.2 Hz), 1.43 (s, 9H); 13C
NMR (CDCl3): d 156.4, 140.3, 140.0, 139.7, 129.3,
129.1, 128.9, 128.8, 127.7, 127.6, 126.6, 80.0, 60.5,
52.3, 50.9, 46.9, 33.6, 31.9, 28.8, 9.41; LRMS (ES+) m/
z 424 [(M+H), 100%].
N-(6-Aminohexyl)-N-(1-phenethylpiperidin-4-yl)propi-
onamide (4d). Trifluoroacetate salt of 4d: brownish hygro-
scopic solid (91%); 1H NMR (CD3OD, 300 MHz) d 7.40–
7.20 (m, 5H), 4.34 (t, 0.5H, J = 11.9 Hz), 4.14 (t, 1H,
J = 11.1 Hz), 3.73 (br d, 2H, J = 11.1 Hz), 3.40–2.80 (m,
10H), 2.60–2.10 (m, 5H), 1.96 (m, 2H), 1.80–1.26 (m,
10H), 1.13 (br t, 3H, J = 7.7 Hz); 13C NMR (CD3OD,
75 MHz) d 176.6, 175.97, 137.6, 130.0, 129.8, 128.3,
59.0, 53.6, 53.1, 52.2, 46.2, 42.8, 40.7, 40.6, 31.7, 31.5,
30.1, 28.9, 28.45, 28.37, 27.9, 27.6, 27.5, 27.3, 27 .0, 26.9,
10.0; LRMS (ES+) m/z 360 [(M+H), 100%]; Anal.
(C22H37N3O 1.6 C2F3HO2) Calcd: C, 55.87; H, 7.11; N,
7.76. Found: C, 55.68; H, 6.99; N, 7.20.
C. Acylation with propionic anhydride. Propionic anhy-
dride (1.5 equiv) was added at room temperature to a
0.2 M solution of amine 1b–9b (1 equiv; typical scale:
5–15 mmol), DMAP catalytic (0.1 equiv) and pyridine
(5 equiv) in dry CH2Cl2. The reaction mixture was stir-
red at room temperature until all the starting material
was consumed. The volatiles were removed under vacu-
um and the crude oil dissolved in EtOAc was extracted
successively with satd NH4Cl aqueous solution and
brine, dried (Na2SO4) and concentrated. The purifica-
tion was carried out as outlined for each compound.
N-(3-(Aminomethyl)benzyl)-N-(1-phenethylpiperidin-
4-yl)propionamide (9d). Trifluoroacetate salt of 9d: yel-
1
lowish oil (99%); H NMR (CDCl3): d 7.68–7.40 (m,
9H), 4.85 (s, 0.8H), 4.81 (s, 1.2H), 4.49 (m, 0.6H), 4.32
(s, 1.2H), 4.27 (s, 0.8H), 3.84 (br d, 2H, J = 12.0 Hz),
3.50 (m, 2H), 3.47–3.28 (m, 4H), 2.86 (q, 0.8H,
J = 7.3 Hz), 2.52 (q, 1.2H, J = 7.3 Hz), 2.32 (q, 2H,
J = 12.5 Hz), 2.12 (br d, 2H, J = 12.9 Hz), 1.40 (t,
1.2H, J = 7.3 Hz), 1.28 (t, 1.8H, J = 7.3 Hz); 13C NMR
(CDCl3): d 177.3, 176.8, 141.5, 140.5, 135.3, 134.7,
130.8, 130.3, 129.9, 129.8, 129.1, 128.2, 127.7, 58.9,
53.6, 53.4, 53.1, 51.9, 45.3, 31.4, 29.0, 27.7, 27.5, 9.9,
9.8; LRMS (ES+) m/z 380 [(M+H), 100%]; Anal.
(C28H35F6N3O5Æ4H2O) Calcd: C, 52.56; H, 6.13; N,
6.99. Found: C, 52.09; H, 6.40; N, 6.51.
tert-Butyl {6-[(1-phenethylpiperidin-4-yl)propionylami-
no]hexyl}carbamate (4c). Following method C. Purifica-
tion by chromatography with CH2Cl2–Et3N (0.5%)/
MeOH–NH3(satd) (100:0 ! 98:2) afforded 4c as a colour-
less oil (73%). Rf = 0.34 (4% MeOH in CH2Cl2); Mix-
ture of two conformers as observed by NMR. 1H
NMR (CDCl3, 400 MHz) d 7.30–7.10 (m, 5H), 4.70–
4.40 (m, 1.5H), 3.57 (br, 0.5H), 3.32 (br d, 1.5H),
3.20–3.00 (m, 4H), 3.00–2.85 (m, 2.5H), 2.85–2.60 (m,
2H), 2.53 (br t, 1H), 2.30 (m, 2H), 2.25–1.65 (m, 4H),
1.60–1.20 (m, 17H), 1.10 (t, 3H, J = 7.3 Hz); 13C
NMR (CDCl3, 100 MHz) d 173.9, 172.9, 156.0, 137.6,
128.7, 128.6, 128.4, 126.7, 126.2, 77.9, 60.0, 59.1, 54.8,
53.0, 52.8, 50.1, 43.7, 41.9, 40.3, 33.3, 31.9, 31.3, 30.3,
29.9, 29.5, 28.3, 27.8, 26.8, 26.6, 26.3, 26.2, 9.6, 9.5;
LRMS (ES+) m/z 460 [(M+H), 100%].
E. Synthesis of the Boc-protected guanidines. HgCl2
(190 mg, 0.7 mmol, 1.2 equiv) was added to a solution
of amine 1d–8d (0.58 mmol, 1 equiv), N,N0-di(tert-butoxy-
carbonyl)thiourea(191 mg, 0.7 mmol, 1.2 equiv) and Et3N
(0.4 mL, 2.9 mmol, 5 equiv) in dry CH2Cl2, under a N2
atmosphere. After 3 h stirring at room temperature, the
darkgreyreactionmixturewasdilutedwithCH2Cl2 andfil-
tered on a pad of Celite. The filter cake was rinsed with
CH2Cl2 (2· 15 mL). The combined organic phases were
washed with water, dried (Na2SO4) and concentrated un-
der vacuum. The crude oil was chromatographed on a
short plug of silica (2.5 · 2.5 cm) eluting with cyclohex-
ane/EtOAc (50:50).
tert-Butyl {3-[(N-(1-phenethylpiperidin-4-yl)propion-
amido)methyl]benzyl} carbamate (9c). Method C. The
crude product was pre-purified on silica gel by flash
chromatography eluting with CH2Cl2/MeOH–NH3
(100:0 ! 98:2), then it was purified on silica gel eluting
with CH2Cl2/ Et3N (100:1) to give 9c as a colourless oil
1
(63%); H NMR (CDCl3): d 8.5 (br, 2H), 7.32–7.08 (m,
9H), 4.90–4.0 (m, 5H), 4.12 (br, 2H), 3.80–3.50 (m, 2H),
3.25–2.65 (m, 5H), 2.47 (m, 2H), 2.10–1.70 (m, 3H),
1.30–1.10 (m, 3H). 13C NMR (CDCl3): d 174.7, 155.8,
140.2, 139.6, 139.1, 128.9, 128.6, 128.4, 126.1, 126.0,
125.8, 124.5, 79.6, 60.4, 55.9, 53.1, 51.7, 46.3. 33.8,
29.7, 28.4, 27.1, 9.2; LRMS (ES+) m/z 480 [(M+H),
100%].
N-[N2,N3-(Bis-tert-butoxycarbonyl)guanidinohexyl]-N-
[1-phenethyl-4-piperidyl]propanamide (4e). Yellowish oil
1
(57%); H NMR (CDCl3, 300 MHz) d 11.43 (br, 1H),
8.22 (br, 1H), 7.30–7.05 (m, 5H), 4.40 (m, 0.5H), 3.50
(m, 0.5H), 3.34 (m, 2H), 3.20–2.95 (m, 4H), 2.74 (m, 2H,
1/2 A2B2), 2.54 (m, 2H, 1/2 A2B2), 2.27 (quint, 2H), 2.06
(m, 2H), 1.93–1.57 (m, 4H), 1.55–1.40 (m, 22H), 1.27 (br
m, 4H), 1.08 (m, 3H); 13C NMR (CDCl3, 75 MHz) d
173.6 (s), 172.9 (s), 164.4 (s), 155.9 (s), 153.16 (s), 153.10
(s), 139.8 (s), 128.5 (d), 128.2 (d), 125.9 (d), 82.88 (s),
82.76 (s), 79.06 (s), 78.95 (s), 60.2 (t), 55.0 (d), 53.05 (t),
52.98 (t), 51.1 (d), 43.2 (t), 41.9 (t), 40.7 (t), 40.5 (t), 33.5
(t), 33.4 (t), 31.3 (t), 30.5 (t), 29.5 (t), 29.4 (t), 28.8 (t),
28.4 (t), 28.1 (q), 27.9 (q), 26.72 (t), 26.67 (t), 26.56 (t),
26.49 (t), 26.37 (t), 9.62 (q), 9.46 (q); LRMS (ES+) m/z
602.5 [(M+H)]; Anal (C33H55N5O5) C, H, N.
D. Removal of the Boc-protecting groups with TFA. A
solution of the Boc-protected amine 1c–9c in 50%
TFA/CH2Cl2 was stirred for 30 min at room tempera-
ture. The volatiles were removed under vacuum. Et2O
was added to the crude residue and rotatory evaporation
was repeated. The crude oil was dissolved in a little
quantity of EtOH (ca 5 mL) and Et2O was added until
a precipitate formed. The product was allowed to stand
in the freezer for 2 h. The mother liquor was discarded
and the oily residue that settled at the bottom of the
flask was collected and dried under high-vacuum,
affording the TFA salt of the product as a highly hygro-
scopic solid.
N-[Guanidinohexyl]-N-[1-phenethyl-4-piperidyl]pro-
panamide (4f).11 Following method D (1 h at room

6578
C. Dardonville et al. / Bioorg. Med. Chem. 14 (2006) 6570–6580
temperature). Trifluoroacetate salt of 4f: brownish
hygroscopic solid (62%); Anal. (C27H39F6N5O5) C,
H, N.
40 mL). The organic extracts were dried (MgSO4) and
concentrated to dryness under reduced pressure to yield
11 as a crude off-white solid (1.3 g, 59%). Rf = 0.11 (5%
MeOH in Et2O); mp = 155–156 ꢁC; H NMR (CDCl3,
300 MHz) d 8.76 (d, 1H, J = 9.2 Hz), 8.54 (m, 2H), 8.27
(dd, 1H, J = 1.5 and 9 Hz), 7.78 (d, 1H, J = 8.5 Hz), 7.32
(dd, 1H, J = 6.2 and 8.3 Hz), 3.96 (s, 3H); 13C NMR
(CDCl3, 75 MHz) d 166.1, 143.6, 137.5, 131.4, 130.8,
130.3, 130.2, 126.9, 122.3, 120.8, 53.1; LRMS (ES+) m/z
204 [(M+H), 100%]. Anal. (C9H11NO3) C, H, N.
1
N-[N2,N3-((Bis-tert-butoxycarbonyl)guanidinomethyl)-
benzyl]-N-[1-phenethyl-4-piperidyl]propanamide (9e).
Pale yellow solid; 1H NMR (CDCl3) d 11.55 (br,
1H), 8.56 (br, 1H), 7.34–7.09 (m, 9H), 4.80–4.40 (m,
4H), 3.03 (m, 2H), 2.74 (m, 2H), 2.56–2.45 (m, 2H),
2.22 (m, 2H), 1.80–1.50 (br m, 5H), 1.48 (m, 18H),
1.26–1.21 (m, 2H), 1.11 (t, 3H, J = 7.3 Hz); LRMS
(ES+) m/z 622 [(M+H), 100%].
Methyl 2-cyanoquinoline-6-carboxylate (12). The N-
oxide 11 (1.25 g, 6.2 mmol) was dissolved in an aqueous
solution (25 mL) containing KCN (720 mg, 11 mmol).
The solution was cooled with an ice-bath and benzoyl
chloride (1.5 mL, 12.9 mmo.l) was added in small por-
tions. The reaction mixture was stirred [note: an efficient
stirring is needed due to formation of a pasty precipi-
tate] at 0 ꢁC for 1 h. The resulting precipitate was col-
lected by filtration and washed thoroughly with 1 M
aqueous K2CO3 and finally with water. The crude prod-
uct was crystallized from MeOH/CH2Cl2 (2:1), rinsed
with Et2O and dried in vacuo, yielding 12 as a pale pink
solid (1.02 g, 78%). Rf = 0.66 (100% in Et2O); mp = 195–
196 ꢁC; 1H NMR (CDCl3, 200 MHz) d 8.64 (s, 1H), 8.41
(m, 2H), 8.22 (br d, 1H, J = 8.8 Hz), 7.76 (br d, 1H,
J = 8.4 Hz), 4.0 (s, 3H); 13C NMR (CDCl3, 75 MHz) d
165.9, 149.8, 139.0, 135.7, 130.8, 130.7, 130.4, 127.9,
124.1, 117.3, 52.9; LRMS (ES+) m/z 213 [(M+H),
100%]. Anal. (C12H8N2O2Æ 0.1H2O) Calcd: C, 67.39;
H, 3.82; N, 13.10. Found: C, 67.12; H, 3.86; N, 12.91.
N-[(3-Guanidinomethyl)benzyl]-N-[1-phenylethyl-4-pip-
eridyl]propanamide (9f). Following method D (2 h at
room temperature). Trifluoroacetate salt of 9f: yellow
hygroscopic solid (100%); 1H NMR (CD3OD,
300 MHz) d 7.70–7.33 (m, 9H), 5.25 (br, overlap with
solvent), 5.00–4.40 (m, 6H), 3.86 (br d, 2H), 3.50 (m,
overlap with solvent), 3.42–3.15 (m, 4H), 2.84 (m, 1H),
2.56 (q, 2H, J = 7.3 Hz), 2.50–2.05 (m, 5H), 1.40 (t,
1.5H, J = 7.3 Hz), 1.29 (t, 1.5H, J = 7.4 Hz), 1.1 (br,
1H); 13C NMR (CDCl3) d 160.7, 160.3, 157.2, 137.4,
136.2, 134.7, 129.9, 129.2, 128.4, 127.8, 127.2, 125.5,
124.5, 58.7, 53.0, 49.8, 47.2, 44.9, 30.4, 27.2, 26.7, 9.5;
LRMS (ES+) m/z 422 [(M+H), 100%].
F. Synthesis of the BU224 moiety
Methyl quinoline-6-carboxylate (10). Thionyl chloride
(4 mL, 55 mmol) was added dropwise at 0 ꢁC to a sus-
pension of commercial 6-quinoline carboxylic acid
(3.17 g, 18.3 mmol) in MeOH (100 mL). A thick precip-
itate formed after completion of the addition. The ice-
bath was removed and the reaction mixture was stirred
at room temperature for 1 h. The stopper of the flask
was replaced by a reflux condenser and the reaction mix-
ture was refluxed for 1 h. After stirring overnight at
room temperature, the solvent was removed under re-
duced pressure. The solid was taken up in a 1 M
K2CO3 solution (50 mL) and the aqueous phase was
extracted with CH2Cl2 (2· 100 mL). The organic ex-
tracts were washed with brine, dried (MgSO4) and con-
centrated to dryness, affording 10 as an off-white solid
(3.28 g, 96%). Rf = 0.22 (33% EtOAc in hexane);
mp = 86–88 ꢁC; 1H NMR (CDCl3, 300 MHz) d 8.96
(dd, 1H, J = 1.8 and 4.2 Hz), 8.54 (d, 1H, J = 1.8 Hz),
8.27 (dd, 1H, J = 1.8 and 9 Hz), 8.21 (dt, 1H,
J = 8.7 Hz), 7.42 (dd, 1H, J = 4.2 and 8.1 Hz), 3.96 (s,
3H); 13C NMR (CDCl3, 75 MHz) d 166.9, 152.9,
150.4, 137.7, 131.4, 130.2, 129.3, 128.5, 127.8, 122.2,
52.8; LRMS (ES+) m/z 188 [(M+H), 100%]; Anal.
(C11H9NO2 0.1 H2O) Calcd: C, 69.95; H, 4.86; N,
7.42. Found: C, 69.93; H, 4.90; N, 7.33.
Methyl 2-(4,5-dihydro-1H-imidazol-2-yl)quinoline-6-
carboxylate (13). A mixture of cyanide 12 (706 mg,
3.3 mmol), ethylenediamine (2.2 mL, 33 mmol) and
P2S5 (73 mg, 0.3 mmol) in dry toluene (10 mL) was stir-
red at room temperature for 2 days under a N2 atmo-
sphere. The mixture was then refluxed for 3 h to lead
the reaction to completion. The solvent was removed
under reduced pressure and the crude residue was parti-
tioned between water and CH2Cl2. The aqueous phase
was extracted with CH2Cl2. The combined organic ex-
tracts were washed with brine, dried and concentrated
to yield 13 as an off-white powder (398 mg, 47%).
Rf = 0.08 (10% MeOH in Et2O); mp = 242–244 ꢁC
1
(180–220 ꢁC sublimation); Free base of 13: H NMR
(CDCl3, 300 MHz) d 8.55 (s, 1H), 8.27 (m, 3H), 8.06
(d, 1H, J = 8.8 Hz), 6.20 (br, 1H, NH), 4.10 (br, 2H),
3.99 (s, 3H), 3.69 (br, 2H); 13C NMR (CDCl3, 75 MHz)
d 167.4 (s), 166.8 (s), 150.9 (s), 149.3 (s), 138.0 (d), 137.6
(d), 131.0 (d), 130.7 (d), 129.6 (d), 129.1 (s), 128.1 (s),
120.6 (d), 57.0 (br), 52.9 (q), 45.4 (br), 43.0 (t, weak),
41.7 (t, weak); LRMS (ES+) m/z 256 [(M+H), 100%].
1
TFA salt of 13: H NMR (CDCl3, 300 MHz) d 10.60
Methyl quinoline-1-oxide-6-carboxylate (11). To the
ester 10 (2.04 g, 10.9 mmol) dissolved in AcOH
(50 mL) was added 30% aqueous H2O2 solution (7 mL,
55 mmol). The reaction mixture was heated at ca.
80 ꢁC for 2 h and stirred at room temperature overnight.
After the repeated addition of water (50 mL) followed
by concentration to the original volume under reduced
pressure (3·), the residue was basified with aqueous
K2CO3 solution and extracted with CH2Cl2 (3·
(br, 1H), 8.50 (s, 1H), 8.50–8.20 (m, 3H), 8.09 (d, 1H,
J = 8.8 Hz), 4.24 (s, 4H), 3.99 (s, 3H); 13C NMR (CDCl3,
75 MHz) d 167.3, 164.8, 149.3, 141.2, 131.6, 131.4, 131.3,
130.7, 129.6, 120.2, 53.7, 45.8.
2-(4,5-Dihydro-1H-imidazol-2-yl)quinoline-6-carbox-
ylic acid (14). Lithium hydroxide monohydrate (55 mg,
1.37 mmol) was added to a stirred solution of 13
(140 mg, 0.55 mmol) in a mixture of THF (5 mL) and

C. Dardonville et al. / Bioorg. Med. Chem. 14 (2006) 6570–6580
6579
water (2 mL). The ester deprotection was complete
after 45 min at room temperature. The solution
was acidified to pH ꢁ3 with 1 M HCl and diluted with
acetone (100 mL). The flask was allowed to stand for
1 h at room temperature. The white precipitate was
collected by filtration and dried in vacuo. Hydrochlo-
ride salt of 14: colourless solid (102 mg, 67%).
mp = 310–312 ꢁC; 1H NMR (D2O, 200 MHz) d 8.00
(m, 2H), 7.86 (d, 1H, J = 10 Hz), 7.49 (d, 1H,
J = 8.8 Hz), 7.38 (d, 1H, J = 8.4 Hz), 4.10 (s, 4H);
13C NMR (D2O, 50 MHz) d 168.9, 163.3, 147.7,
141.5, 140.2, 130.9, 130.3, 130.2, 130.1, 127.9, 119.4,
45.2; LRMS (ES+) m/z 242 [(M+H), 100%]. Anal.
(C13H12ClN3O2) C, H, N.
chased from Sigma (St. Louis, USA). Morphine HCl
was obtained from Alcaliber S.A. (Madrid, Spain). All
other chemicals were of the highest purity commercially
available.
5.3.1. Preparation of membranes. Postmortem human
frontal cortex samples of each subject (ꢁ1 g) were
homogenized using a Teflon–glass grinder (10 up-and-
down strokes at 1500 rpm) in 30 volumes of homogeni-
zation buffer (1 mM EGTA, mM MgCl2, 1 mM DTT
and 50 mM Tris–HCl, pH 7.4) supplemented with
0.25 M sucrose. The crude homogenate was centrifuged
for 5 min at 1000g (4 ꢁC) and the supernatant was cen-
trifuged again for 10 min at 40,000g (4 ꢁC). The resul-
tant pellet was washed twice in 20 volumes of
homogenization buffer and recentrifuged in similar con-
ditions. Aliquots of 1 mg protein were stored at ꢀ70 ꢁC
until assay. Protein content was measured according to
the method of Bradford using BSA as standard, and
was similar in the different brain samples.
G. Synthesis of the BU224 hybrid molecules. To a mix-
ture of the trifluoroacetate salt of the amine 1d–8d
(0.15 mmol, 1 equiv) and 14 (0.15 mmol, 1 equiv) in
dry CH2Cl2 (3 mL) were added Et3N (0.1 mL, 5 equiv)
and 2-chloro-1-methylpyridinium iodide (0.17 mmol,
1.2 equiv). The reaction mixture was stirred at room
temperature and quenched when all the acid starting
material had been consumed (1–4 days). Workup A:
the crude slurry was poured into diluted HCl (10 mL)
and the aqueous phase was extracted with EtOAc (3·
10 mL). The aqueous phase was basified to pH ꢁ9 with
10% aqueous NaOH solution and extracted with EtOAc
(3· 10 mL). The combined organic phases were washed
with brine, dried (Na2SO4) and concentrated to give a
crude yellow oil. Workup B: the crude reaction mixture
was partitioned between EtOAc and water. The organic
phase was collected and the aqueous phase was basified
with 10% aqueous NaOH solution. The aqueous phase
was extracted with CH2Cl2. The combined organic ex-
tracts were washed with brine, dried and concentrated.
5.3.2. [35S]GTPcS binding assays. The incubation buffer
for measuring [35S]GTPcS binding to brain membranes
contained, in a total volume of 500 lL, 1 mM EGTA,
3 mM MgCl2, 100 mM NaCl, 50 mM GDP, 50 mM
Tris–HCl at pH 7.4 and 0.5 nM [35S]GTPcS. Protein ali-
quots were thawed and re-suspended in the same buffer.
The incubation was started by addition of the mem-
brane suspension (40 lg of membrane proteins) to the
previous mixture and was performed at 30 ꢁC for
120 min with shaking. In order to evaluate the influence
of the compounds on [35S]GTPcS binding, eight concen-
trations (10ꢀ10–10ꢀ3 M) of the different drugs were added
to the assay. Incubations were terminated by adding 3 mL
of ice-cold re-suspension buffer followed by rapid filtra-
tion through Whatman GF/C filters pre-soaked in the
same buffer. The filters were rinsed twice with 3 mL of
ice-cold re-suspension buffer, transferred to vials contain-
ing 5 mL of OptiPhase HiSafe II cocktail (Wallac, UK)
and the radioactivity trapped was determined by liquid
scintillation spectrometry (Packard 2200CA). The
[35S]GTPcS bound was about 7–14% of the total
[35S]GTPcS added. Non-specific binding of the radioli-
gand was defined as the remaining [35S]GTPcS binding
in the presence of 10 lM unlabelled GTPcS.
2-(4,5-Dihydro-1H-imidazol-2-yl)-N-(6-(N-(1-phen-
ethylpiperidin-4-yl)propionamido)hexyl)quinoline-6-car-
boxamide (4g). Following workup B. Chromatography
(2 g SI) with CH2Cl2/MeOH–NH3(satd) (100:0 ! 85:15)
yielded the hybrid molecule 4g as a yellowish oil (8%).
1
Rf = 0.2 (10% MeOH–NH3(satd) in CH2Cl2); H NMR
(CDCl3, 300 MHz) d 8.40–8.25 (m, 3H), 8.12 (m, 2H),
7.35–7.15 (m, 5H), 6.79 (br m, 0.5H), 6.52 (br m,
0.5H), 3.92 (br s, 4H), 3.52 (m, 2H), 3.30–3.05 (m,
4H), 2.83 (m, 2H, 1/2 A2B2), 2.67 (m, 2H, 1/2 A2B2),
2.45–1.20 (m, 18H), 1.15 (t, 1.5H, J = 7.3 Hz), 1.12
(t, 1.5H, J = 7.3 Hz); 13C NMR (CDCl3, 75 MHz) d
173.8, 173.2, 166.7, 164.3 (2· C), 149.7, 149.5, 148.2,
140.0, 139.7, 137.4 (2· C), 133.6, 130.1, 130.0, 128.6,
128.4 (br), 128.0, 127.9, 127.7, 127.2, 126.2 (2· C),
120.3, 120.2, 60.4, 60.3, 55.4, 53.2, 51.1, 43.3, 41.4,
40.2, 39.9, 33.8, 33.5, 31.5, 30.9, 30.3, 29.7 (m), 29.4,
29.2, 27.3 (m), 26.9 (m), 26.7 (m), 26.4, 26.0, 9.8 (2·
C); LRMS (ES+) m/z 583.5 [(M+H)], 292.3 [(M+2H),
100]; Anal. (C35H46N6O2) C, H, N.
5.3.3. In vivo analgesia
Animals. Male Swiss mice (final weight 26–32 g) were
housed 6 per cage on a 12 h light–dark cycle at 22 ꢁC with
free access to food and water. The animals were allowed
to habituate to the laboratory environment for at least 2 h
before any experiment was initiated; they were used only
once and killed by cervical dislocation after testing.
Writhing test. The abdominal constrictor test was
performed by the intraperitoneal (ip) injection of acetic
acid (2%; 10 mL/kg) to produce the typical writhing
reaction that mimics acute visceral pain (arching of
back, development of tension in the abdominal muscles,
elongation of the body and extension of the hind limbs).
Immediately after the acetic acid administration, the
animals were placed in individual transparent cages
and the number of writhes was counted during a
5.3. Pharmacology, materials and methods
[35S]GTPcS (1250 Ci/mmol) was purchased from Du-
Pont NEN (Brussels, Belgium), [D-Ala2, N-Me-Phe4,
Gly5-ol]-enkephalin
(DAMGO),
DL-dithiothreitol
(DTT), GDP, GTP, GTPcS and naloxone were pur-

6580
C. Dardonville et al. / Bioorg. Med. Chem. 14 (2006) 6570–6580
20 min period. The test was carried out twenty min after
the ip injection of compound 4f (20 and 40 mg/kg,
n = 10–11) or the same volume of saline (5 mL/kg,
n = 12). Morphine (10 mg/kg, ip, n = 9) was used as a
positive control. The analgesic activity was expressed
as the percent inhibition of the number of writhes ob-
served in saline control animals (16.7 3.5, n = 12).
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/
j.bmc.2006.06.007.
References and notes
1. Dardonville, C.; Rozas, I. Med. Res. Rev. 2004, 24, 639.
2. Sanchez-Blazquez, P.; Boronat, M. A.; Olmos, G.; Garcia-
Sevilla, J. A.; Garzon, J. Br. J. Pharmacol. 2000, 130, 146.
3. Ruiz-Durantez, E.; Torrecilla, M.; Pineda, J.; Ugedo, L.
Br. J. Pharmacol. 2003, 138, 494.
4. Boronat, M. A.; Olmos, G.; Garcia-Sevilla, J. A. Br. J.
Pharmacol. 1998, 125, 175.
5. Boronat, M. A.; Olmos, G.; Garcia-Sevilla, J. A. Ann.
N.Y. Acad. Sci. 1999, 881, 359.
6. Morgan, A. D.; Campbell, U. C.; Fons, R. D.; Carroll, M.
E. Pharmacol. Biochem. Behav. 2002, 72, 873.
7. Miralles, A.; Esteban, S.; Sastre-Coll, A.; Moranta, D.;
Asensio, V. J.; Garcia-Sevilla, J. A. Eur. J. Pharmacol.
2005, 518, 234.
8. Su, R. B.; Li, J.; Qin, B. Y. Acta Pharmacol. Sin. 2003, 24,
631.
9. Morphy, R.; Rankovic, Z. J. Med. Chem. 2005, 48, 6523.
10. Montero, A.; Goya, P.; Jagerovic, N.; Callado, L. F.;
Meana, J. J.; Giron, R.; Goicoechea, C.; Martin, M. I.
Bioorg. Med. Chem. 2002, 10, 1009.
Hot-plate test. Mice were placed in a hot plate
(LE7406, Panlab S.L.; Spain) at 55 0.1 ꢁC as a nocicep-
tive stimulus and the time of latency until jumping was
taken as the end point, with a cut-off time of 120 s. The
time of latency was determined 30 min before and at dif-
ferent times (5–120 min) after the ip injection of separated
doses of compound 4f (1, 10, 20 and 40 mg/kg, n = 6) or
the same volume of saline (5 mL/kg, n = 10). Morphine
(2.5, 5 and 10 mg/kg, ip) was used as a positive control.
Antinociceptive activity was expressed as the percentage
of the maximal possible effect (% MPE), which was calcu-
lated as follows:
% MPE = 100 · (latency of treated ꢀ latency of saline)/
(cut-off time ꢀ latency of saline). The latency of saline
was 25 2s (n = 10).
11. Dardonville, C.; Jagerovic, N.; Callado, L. F.; Meana, J.
J. Bioorg. Med. Chem. Lett. 2004, 14, 491.
Acknowledgments
12. Diaz, A.; Mayet, S.; Dickenson, A. H. Eur. J. Pharmacol.
1997, 333, 9.
13. Kaneko, C.; Yamamoto, A.; Hashiba, M. Chem. Pharm.
Bull. 1979, 27, 946.
14. Poss, M. A.; Iwanowicz, E.; Reid, J. A.; Lin, J.; Gu, Z.
Tetrahedron Lett. 1992, 33, 5933.
15. Dardonville, C.; Rozas, I.; Callado, L. F.; Meana, J. J.
Bioorg. Med. Chem. 2002, 10, 1525.
16. Miralles, A.; Olmos, G.; Sastre, M.; Barturen, F.; Martin,
I.; Garcia-Sevilla, J. A. J. Pharmacol. Exp. Ther. 1993,
264, 1187.
17. Gabilondo, A. M.; Meana, J. J.; Barturen, F.; Sastre, M.;
Garcia-Sevilla, J. A. Psychopharmacology (Berl.) 1994,
115, 135.
This work was supported by the Spanish Grant SAF00–
0114–C02, the Basque Government (ETORTEK Pro-
gram) and the Bizkaiko Foru Aldundia (EKINBERRI
7/12/EK/2005/65). C.D. was a recipient of a postdoctoral
´
fellowship from the Spanish Ministerio de Educacion
Cultura y Deporte (SB2001–0174). CFF is a recipient
of a postgraduate fellowship from the CSIC (I3P–
BPD2004). We thank the staff members of the Instituto
Vasco de Medicina Legal for their cooperation in the
study and Ms. Itxazne Rodil for her excellent technical
assistance.
18. Martin-Gomez, J. I.; Ruiz, J.; Callado, L. F.; Garibi, J.
M.; Aguinaco, L.; Barturen, F.; Javier Meana, J. Neuro-
Report 1996, 7, 1393.
Supplementary data
19. Gonzalez-Maeso, J.; Rodriguez-Puertas, R.; Gabilondo,
A. M.; Meana, J. J. Eur. J. Pharmacol. 2000, 390, 25.
20. Gonzalez-Maeso, J.; Rodriguez-Puertas, R.; Meana, J. J.
Eur. J. Pharmacol. 2002, 452, 21.
Synthetic and spectroscopic data for 1a, 3a, 5a–8a, 1b,
3b, 5b–8b, 1c–3c, 5c–8c, 1d–3d, 5d–8d, 1e, 3e, 5e–8e,
1f, 3f, 5f–8f, 2g, 6g and 8g. Combustion analysis data.