Design, synthesis and biological evaluation of N-phenylalkyl-substituted tramadol derivatives as novel μ opioid receptor ligands

Article abstract of DOI:10.1038/aps.2014.171

Aim: Tramadol is an atypical opioid analgesic with low potential for tolerance and addiction. However, its opioid activity is much lower than classic opiates such as morphine. To develop novel analgesic and further explore the structure activity relations

Full text of DOI:10.1038/aps.2014.171

Acta Pharmacologica Sinica 2015: 1–8
© 2015 CPS and SIMM All rights reserved 1671-4083/15
www.nature.com/aps
npg
Original Article
Design, synthesis and biological evaluation of
N-phenylalkyl-substituted tramadol derivatives as
novel µ opioid receptor ligands
2,
1,
Qing SHEN1, #, Yuan-yuan QIAN1, #, Xue-jun XU2, Wei LI1, , Jing-gen LIU , Wei FU
*
*
*
1Department of Medicinal Chemistry, School of Pharmacy, Fudan University, Shanghai 201203, China; 2Shanghai Institute of Materia
Medica, Chinese Academy of Sciences, Shanghai 201203, China
Aim: Tramadol is an atypical opioid analgesic with low potential for tolerance and addiction. However, its opioid activity is much lower
than classic opiates such as morphine. To develop novel analgesic and further explore the structure activity relationship (SAR) of
tramadol skeleton.
Methods: Based on a three-dimensional (3D) structure superimposition and molecular docking study, we found that M1 (the active
metabolite of tramadol) and morphine have common pharmacophore features and similar binding modes at the μ opioid receptor
in which the substituents on the nitrogen atom of both compounds faced a common hydrophobic pocket formed by Trp2936.48 and
Tyr3267.43. In this study, N-phenethylnormorphine was docked to the μ opioid receptor. It was found that the N-substituted group of
N-phenethylnormorphine extended into a hydrophobic pocket formed by Trp2936.48 and Tyr3267.43. This hydrophobic interaction
may contribute to the improvement of its opioid activities as compared with morphine. The binding modes of M1, morphine and
N-phenethylnormorphine overlapped, indicating that the substituent on the nitrogen atoms of the three compounds may adopt
common orientations. A series of N-phenylalkyl derivatives from the tramadol scaffold were designed, synthesized and assayed in
order to generate a new type of analgesics.
Results: As a result, compound 5b was identified to be an active candidate from these compounds. Furthermore, the binding modes of
5b and morphine derivatives in the μ opioid receptor were comparatively studied.
Conclusion: Unlike morphine-derived structures in which bulky N-substitution is associated with improved opioid-like activities, there
seems to be a different story for tramadol, suggesting the potential difference of SAR between these compounds. A new type of
interaction mechanism in tramadol analogue (5b) was discovered, which will help advance potent tramadol-based analgesic design.
Keywords: μ-opioid receptor; tramadol; morphine; molecular docking
Acta Pharmacologica Sinica advance online publication, 8 Jun 2015; doi: 10.1038/aps.2014.171
and chronic[1]. In addition, it has also been used to treat
Introduction
Opioids are narcotic analgesics widely prescribed to relieve depression, postherpetic neuralgia, diabetic neuropathy pre-
moderate-to-severe pains and most of these agents exert their mature ejaculation[2-8]. Tramadol displays some weak side
opioid-like effects through opioid receptors (eg, μ, δ, κ, and effects, including nausea, dizziness, indigestion and abdomi-
ORL1 receptors). However, most clinically used analgesics nal pain in individual patients[9, 10]. It was identified to be an
are restricted to μ opioid agonists, which are associated with atypical opioid, both structurally and pharmacologically. It
respiratory depression, constipation, addiction, physical acts as a weak μ opioid receptor agonist and serotonin reup-
dependence and other notorious adverse effects. Tramadol, take and norepinephrine reuptake inhibitor[11, 12]. Its O-des-
which was launched in 1977, is a fully synthetic opioid pain methyl metabolite (M1) is much more potent on the μ opioid
medication used to treat moderate to severe pain, both acute receptor[13] (Figure 1). However, few studies on structure and
activity relationship of tramadol analogues and the binding
# These authors contributed equally to this work.
To whom correspondence should be addressed.
E-mail wfu@fudan.edu.cn (Wei FU);
wei-li@fudan.edu.cn (Wei LI);
jgliu@mail.shcnc.ac.cn (Jing-gen LIU)
Received 2014-11-16 Accepted 2014-12-29
mode between tramadol analogues and μ opioid receptor were
*
performed. In this study, we designed a series of N-phenylal-
kyl substituted tramadol derivatives to explore their structure
activity relationship for developing novel opioid analgesics
and comparatively discussed the difference in binding modes

www.nature.com/aps
npg
Shen Q et al
2
was added into the reaction with a syringe. When the mixture
was heated to 70°C, 3-bromoanisole (5 mmol) solved in anhy-
drous THF was added drop-wise to the mixture. After all
the magnesium chips were solved, the reaction was cooled to
room temperature. Then, the Mannich base (3a–3d) (1 mmol)
solved in anhydrous THF was added drop-wise to the reac-
tion. After stirring for 3 h, the reaction was quenched with a
saturated NH4Cl aqueous solution. The mixture was diluted
with water and extracted with CHCl3 (3×15 mL). Combined
organic extracts were dried over anhydrous sodium sulfate,
filtered, and evaporated under reduced pressure. The residue
was purified with column chromatography on silica gel using
CH2Cl2/MeOH (20:1) to yield a yellow solid (4a–4d). The base
(4a–4d) was transformed into hydrochlorides in diethylether
by adding HCl in diethylether.
Figure 1. Structure of (±)-tramadol 1, metabolite (±)-M1, codeine, and
morphine.
General procedure for the preparation of (5a–5d)
A solution of (4a–4d) (0.59 mmol) in anhydrous DCM (15 mL)
was cooled to -40°C. Boron tri-bromide in DCM (1.5 mmol)
was added drop-wise into the solution. After being stirred
for 1 h, the reaction mixture was allowed to adjust to room
temperature. The reaction mixture was quenched by the addi-
tion of drops of ice-cold water. After dilution with water, the
crude product was extracted with CHCl3 (3×15 mL). Com-
bined organic extracts were dried over anhydrous sodium
sulfate, filtered, and evaporated under reduced pressure. The
crude product was purified by column chromatography over
silica using CH2Cl2/MeOH (20:1) to yield yellow oil (5a–5d).
The base (5a–5d) was transformed into hydrochlorides in
diethylether by adding HCl in diethylether.
of these compounds and morphine derivatives.
Materials and methods
Experimental section
All chemicals and solvents were supplied by Tansoole and
were used without further purification. 1H and 3C NMR spec-
tra were recorded on a Bruker AMX-400 instrument. Proton-
coupling patterns were described as singlet, doublet, triplet,
quartet, multiplet, and broad. Mass spectra were generated
with electric ionization (ESI) produced by an HP5973N ana-
lytical mass spectrometer. High-resolution mass spectrom-
etry (HRMS) spectra were recorded with an AB 5600+ Q TOF
instrument.
2-{[Benzyl(methyl)amino]methyl}-1-(3-methoxyphenyl)
cyclohexanol-HCl (4a)
1
White powder, 55 mg 47% H NMR (400 MHz, d6-DMSO) δ
General procedure for the syntheses of N-methyl-N-phenylalkyl-
aminomethyl-cyclohexanones (3a–3d)
10.18 (s, 1H), 7.62–7.53 (m, 1H), 7.47–7.19 (m, 5H), 7.04 (d, J=7.9
Hz, 2H), 6.86–6.72 (m, 1H), 5.06 (d, J=46.2 Hz, 1H), 4.20 (ddd,
J=31.9, 13.0, 3.6 Hz, 1H), 3.93 (ddd, J=18.3, 13.0, 6.1 Hz, 1H),
3.76 (d, J=7.1 Hz, 3H), 2.84–2.61 (m, 1H), 2.47–2.27 (m, 4H),
2.25–2.02 (m, 2H), 1.89–1.07 (m, 7H) ppm. 13C NMR (151 MHz,
DMSO) δ 159.61, 150.36, 131.85, 131.61, 130.79, 129.81, 129.53,
129.11, 128.95, 117.71, 112.04, 74.44, 60.26, 57.57, 56.78, 55.46,
42.23, 26.64, 24.89, 21.55. ESI-MS m/z 340.2 [M+H]+ HRMS
m/z calculated for C22H30NO2 [M+H]+, 340.2271; observed,
340.2281.
A mixture of cyclohexanone (1 g, 1.06 mmol), N-methylphe-
nylalkyl-amine (1.06 mmol) and 1/5th of paraformaldehyde
(2 mmol) in isopropanol (20 mL) was stirred and concentrated
HCl was added drop-wise to adjust the solution to pH 4. The
reaction mixture was then heated in an oil bath at 90–95 °C for
30 min with stirring. The other four portions of paraformal-
dehyde were added at 15 min intervals. The reaction mixture
was further refluxed for 4 h and the solvent was distilled
off. The residues were washed with hexane, adjusted to an
alkaline pH with a sodium bicarbonate solution and extracted
with ethyl acetate. Combined organic extracts were dried over
Na2SO4 and distilled to afford the Mannich bases (3a–3d) in
good yield.
2-{[Methyl(phenethyl)amino]methyl}-1-(3-methoxyphenyl)-
cyclohexanol-HCl (4b)
White powder 27 mg 36% 1H NMR (400 MHz, DMSO) δ 10.14
(d, J=32.5 Hz, 1H), 7.37–7.04 (m, 8H), 6.78 (t, J=6.4 Hz, 1H), 5.12
(s, 1H), 3.73 (d, J=18.1 Hz, 3H), 3.15–2.74 (m, 4H), 2.74–2.53 (m,
3H), 2.44 (dd, J=18.4, 8.9 Hz, 2H), 2.35–2.07 (m, 2H), 1.92–1.35
(m, 7H). 13C NMR (151 MHz, DMSO) δ 159.64, 150.44, 137.50,
137.19, 129.56, 129.14, 129.00, 127.22, 127.15, 117.80, 112.04,
74.54, 58.22, 55.45, 54.26, 42.54, 30.01, 28.58, 27.05, 25.02, 21.66.
ESI-MS m/z 354.3 [M+H]+ HRMS m/z calculated for C23H32NO2
[M+H]+, 354.2428; observed, 354.2438.
General procedure of the Grignard reaction for the preparation of
4a–4d
A 100-mL, three-necked, round-bottomed flask equipped with
a magnetic stirring bar, dropping funnel and reflux condenser
was charged with magnesium chips (5 mmol) and iodine (5
mg) under argon. Anhydrous tetrahydrofuran (THF, 15 mL)
Acta Pharmacologica Sinica

www.chinaphar.com
Shen Q et al
npg
3
1-(3-Methoxyphenyl)-2-{[methyl(3-phenylpropyl)amino]methyl} J=4.6 Hz, 2H), 2.48 (d, J=6.9 Hz, 2H), 2.41 (d, J=4.7 Hz, 2H),
cyclohexanol-HCl (4c)
2.17–1.29 (m, 11H). 13C NMR (101 MHz, DMSO) δ 157.67,
1
White powder 100 mg 60% H NMR (400 MHz, DMSO) δ 150.12, 140.97, 129.43, 128.84, 128.69, 126.56, 116.02, 113.74,
10.07 (d, J=57.6 Hz, 1H), 7.36–7.01 (m, 8H), 6.85–6.66 (m, 1H), 112.84, 74.19, 57.61, 57.11, 52.55, 42.32, 32.39, 27.02, 25.65, 25.07,
5.10 (s, 1H), 3.74 (d, J=7.2 Hz, 3H), 2.84 (t, J=10.9 Hz, 2H), 2.68 23.83, 21.58. ESI-MS m/z 354.3 [M+H]+ HRMS m/z calculated
(m, 1H), 2.54 (d, J=4.7 Hz, 2H), 2.49–2.43 (m, 2H), 2.43–2.31 for C23H32NO2 [M+H]+, 354.2427; observed, 354.2435.
(m, 2H), 2.29–2.02 (m, 2H), 1.89–1.27 (m, 9H). 13C NMR (151
MHz, DMSO) δ 159.64, 150.41, 140.94, 129.58, 128.82, 128.72, 3-{1-hydroxy-2-{[methyl(4-phenylbutyl)amino]methyl}cyclohexyl}
128.49, 126.88, 126.56, 117.76, 111.72, 74.52, 57.10, 55.46, 52.56, phenol-HCl (5d)
1
42.35, 32.16, 27.01, 25.56, 24.99, 23.73, 21.59. ESI-MS m/z 368.3 White powder 80 mg 68% H NMR (400 MHz, DMSO) δ 9.15
[M+H]+ HRMS m/z calculated for C24H34NO2 [M+H]+, 368.2584; (s, 1H), 7.32–7.23 (m, 2H), 7.16 (d, J=10.0, 4.0 Hz, 3H), 7.06 (t,
observed, 368.2593.
J=7.8 Hz, 1H), 6.90 (s, 1H), 6.82 (d, J=7.6 Hz, 1H), 6.55 (dd,
J=7.9, 2.0 Hz, 1H), 4.83 (s, 1H), 2.54 (d, J=7.7 Hz, 2H), 2.26–
1-(3-Methoxyphenyl)-2-{[methyl(4-phenylbutyl)amino]methyl} 2.04 (m, 2H), 2.02–1.84 (m, 4H), 1.83–1.18 (m, 14H). 13C NMR
cyclohexanol-HCl (4d)
(101 MHz, DMSO) δ 156.79, 151.29, 142.27, 128.44, 128.19,
1
White powder 78 mg 67% H NMR (400 MHz, DMSO) δ 9.86 125.56, 115.50, 112.41, 112.19, 74.38, 58.44, 57.44, 43.26, 42.65,
(d, J=74.2 Hz, 1H), 7.34–7.14 (m, 6H), 7.07 (dd, J=17.0, 9.9 Hz, 41.03, 34.96, 28.67, 26.46, 26.15, 25.70, 21.75. ESI-MS m/z 368.3
2H), 6.89–6.77 (m, 1H), 5.14 (dd, J=25.6, 7.9 Hz, 1H), 3.76 (d, [M+H]+ HRMS m/z calculated for C24H34NO2 [M+H]+, 368.2584;
J=1.9 Hz, 3H), 2.91–2.76 (m, 2H), 2.68 (dt, J=17.4, 11.8 Hz, 1H), observed, 368.2588.
2.58–2.52 (m, 3H), 2.49–2.30 (m, 3H), 2.30–2.01 (m, 2H), 1.86–
0.95 (m, 11H). 13C NMR (151 MHz, DMSO) δ 159.65, 150.43, Radioligand binding assay
142.08, 140.32, 129.58, 128.77, 128.66, 126.88, 126.33, 117.87, Chinese hamster ovary (CHO) cells stably transfected with
112.04, 74.43, 57.65, 57.09, 55.46, 52.73, 42.27, 34.93, 28.39, 27.12, the human κ-opioid receptor and the δ-opioid receptor were
24.98, 23.47, 21.63. ESI-MS m/z 382.3 [M+H]+ HRMS m/z calcu- obtained from SRI International (Palo Alto, CA, USA), and
lated for C25H36NO2 [M+H]+, 382.2741; observed, 382.2742.
those transfected with the μ-opioid receptor were obtained
from George Uh1 (NIDA Intramural Program, Bethesda, MD,
3-{2-{[Benzyl(methyl)amino]methyl}-1-hydroxycyclohexyl}phenol- USA). The cells were grown in 100-mm dishes in Dulbecco’s
HCl (5a)
modified Eagle’s media (DMEM) supplemented with 10% fetal
1
White powder 98 mg 49% H NMR (400 MHz, DMSO) δ 9.69 bovine serum (FBS) and penicillin-streptomycin (10000 U/mL)
(s, 1H), 9.36 (d, J=13.1 Hz, 1H), 7.56–7.19 (m, 5H), 7.13 (dt, at 37 °C under 5% CO2 atmosphere. The affinity and selec-
J=11.0, 7.9 Hz, 1H), 6.99–6.80 (m, 2H), 6.64 (ddd, J=18.3, 8.0, tivity of the compounds for multiple opioid receptors were
2.1 Hz, 1H), 4.98 (d, J=43.5 Hz, 1H), 4.31–4.12 (m, 1H), 3.95 determined by incubating the membranes with radiolabeled
(ddd, J=18.5, 12.9, 6.0 Hz, 1H), 2.85–2.54 (m, 2H), 2.47–2.22 (m, ligands and 12 different concentrations of the compounds at
4H), 2.13–1.13 (m, 9H). 13C NMR (101 MHz, DMSO) δ 157.64, 25°C in a final volume of 1 mL of 50 mmol/L Tris-HCl, pH
150.10, 131.55, 129.90, 129.42, 129.19, 129.04, 115.99, 113.69, 7.5. Incubation times of 60 min were used for the μ-selective
112.80, 74.27, 60.37, 57.07, 56.60, 42.10, 29.47, 26.49, 24.95, 21.52. peptide [3H]DAMGO, the κ-selective ligand [3H]U69593 and
ESI-MS m/z 326.2 [M+H]+ HRMS m/z calculated for C21H28NO2 the δ-selective antagonist [3H]DPDPE.
[M+H]+, 326.2115; observed, 326.2118.
[35S]GTP-γ-S functional assay
3-{1-Hydroxy-2-{[methyl(phenethyl)amino]methyl}cyclohexyl} In a final volume of 0.5 mL, various concentrations of each
phenol-HCl (5b)
tested compound were incubated with 7.5 mg of CHO cell
1
White powder 118 mg 57% H NMR (400 MHz, DMSO) δ membranes that stably expressed the human μ opioid recep-
9.32 (s, 1H), 8.80 (d, J=58.1 Hz, 1H), 7.40–7.07 (m, 6H), 6.94 (d, tor. The assay buffer consisted of 50 mmol/L Tris-HCl, pH
J=12.7 Hz, 2H), 6.62 (d, J=7.2 Hz, 1H), 5.06 (s, 1H), 3.23–2.53 7.4, 3 mmol/L MgCl2, 0.2 mmol/L EGTA, 3 mmol/L GDP, and
(m, 8H), 2.16 (s, 1H), 1.94–1.32 (m, 9H). 13C NMR (101 MHz, 100 mmol/L NaCl. The final concentration of [35S]GTP-γ-S
DMSO) δ 157.30, 151.79, 140.91, 129.05, 128.97, 128.60, 126.16, was 0.08 nmol/L. Nonspecific binding was measured by the
116.02, 112.93, 112.69, 74.84, 59.90, 58.52, 43.81, 43.26, 41.50, inclusion of 10 mmol/L GTP-γ-S. Binding was initiated by the
33.09, 26.88, 26.16, 22.23. ESI-MS m/z 340.3 HRMS m/z calcu- addition of the membranes. After an incubation of 60 min at
lated for C22H30NO2 [M+H]+, 340.2271; observed, 340.2282.
30°C, reactions were terminated by rapid filtration and radio
activity was determined by liquid scintillation counting.
3-{1-Hydroxy-2-{[methyl(3-phenylpropyl)amino]methyl}cyclohexyl}
phenol-HCl (5c)
Molecular simulation
White powder 120 mg 66% 1H NMR (400 MHz, DMSO) δ 9.59 The 3D structure of tramadol, M1, morphine, and codeine
(d, 1H), 9.38 (d, 1H), 7.36–7.26 (m, 2H), 7.15 (ddt, J=15.8, 10.2, were built using the SYBYL6.9 program and optimized by
7.6 Hz, 4H), 6.99–6.79 (m, 2H), 6.63 (d, J=7.9 Hz, 1H), 5.01 (s, a Gaussian program with the same method used in a previ-
1H), 2.85 (dd, J=13.8, 8.8 Hz, 2H), 2.77–2.60 (m, 1H), 2.57 (d, ous study[14]. The 3D structures of the compounds were also
Acta Pharmacologica Sinica

www.nature.com/aps
npg
Shen Q et al
4
superimposed using the software packages in SYBYL6.9.
structure of opioid receptors[19-21]. M1 maintained the classic
Molecular docking was conducted using GOLD 5.0.1[15]. The interactions of morphinans with the μ opioid receptor. The
binding site was defined to include all residues within a 15.0 superimposition and the docking studies showed that M1 and
Å radius of the conserved D3.32Cγ carbon atom. A hydrogen- morphine have the same pharmacophore features and similar
bond constraint was set between the protonated nitrogen binding mode at the μ opioid receptor.
atom (N1) of ligand and D3.32 side chain. Ten conformations
In carefully examining the binding modes of M1 and mor-
were produced for each ligand and Gold-Score was used as phine, it was worth noting that the substituent on the nitro-
the scoring function. Other parameters were set as standard gen atom of both compounds faced a common hydrophobic
default. High-scoring complexes were inspected visually to pocket formed by Trp2936.48 and Tyr3267.43. It is well known
select the most reasonable solution.
that the introduction of N-arylalkyl substitutents, such as an
N-phenylethyl group, is associated with significant improve-
ment of opioid-like activities for classic morphinans. Intro-
duction of an N-phenylethyl substituent on morphine signifi-
Results
Design rationality
In previous studies over the past several years, tramadol is cantly enhances the activity of morphine at the μ opioid recep-
usually considered to be structurally related with codeine[16–18]
.
tor. The binding mode of N-phenethylnormorphine was also
In our initial study, the three dimensional structures of tra- examined by molecular docking (Figure 3C). The N-pheneth-
madol and codeine were superimposed (Figure 2). It was ylnormorphine maintained the common binding interac-
found that the nitrogen atoms and 3-methoxylphenyl groups tions of M1 and morphine, while the N-phenylethyl portion
in both compounds were located in the same position, which extended into the hydrophobic pocket formed by Trp2936.48
showed that tramadol and codeine contained common phar- and Tyr3267.43. This additional interaction may contribute to
macophore features. As tramadol displays μ opioid activity the improvement of N-phenethylnormorphine opioid activity
primarily through its O-desmethyl metabolite (M1), morphine compared with morphine. When the binding modes of M1,
is also a more potent O-desmethyl metabolite of codeine. The morphine and N-phenethylnormorphine were superimposed,
3D structure of M1 and morphine were also superimposed we found that protonated nitrogen atoms in the three com-
(Figure 2), which indicated that the two compounds contain pounds formed salt bridges with Asp1473.32, while their phenol
the same pharmacophore features. Then, M1 and morphine groups formed hydrogen binding interactions with water
were docked to a crystal structure of the μ opioid receptor molecules and His2976.52 (Figure 3D). The N-substitution of
(PDB code 4DKL) using the program GOLD 5.0.1. It was the three compounds also faced the same hydrophobic pocket
found that protonated nitrogen atoms in both compounds (Figure 3D, 3E). Because N-phenylethyl substitution can
formed a salt bridge with Asp1473.32, while the phenol groups improve the opioid activity of morphine, we proposed that the
of the two compounds formed hydrogen binding interactions introduction of a phenylalkyl group, such as phenylethyl, to
with water molecules (Figure 3A, 3B). These binding modes the nitrogen atom of M1 may enhance the binding affinity at
were consisted with morphinans’ binding modes in the crystal the μ opioid site as it does to morphine. Thus, we designed a
series of N-phenylalkyl derivatives of tramadol (Figure 4) and
M1 to investigate if this modification could improve the activ-
ity of the μ opioid receptor with the aim of developing a novel
class of opioid ligands.
Synthesis
The synthesis of 4a–5d was described in Scheme 1. Cyclo-
hexanone was condensed with paraformaldehyde and
N-methylalkylphenylamine to afford aminomethylhexanones
3a–3d, which were further reacted with the Grignard reagent
prepared from 3-bromoanisole to yield 4a–4d. Specifically,
addition of the Grignard reagent to ketones 3a–3d provided
crude 4a–4d as mixtures of diastereomers (76% cis for 4a–4d)
[16]. The abundance of the cis-isomer could be improved to 95%
by flash chromatography (DCM/MeOH, 20:1). O-desmethyl
derivatives were prepared by removal of the O-methyl group
of 5a–5d.
Binding affinity and functional activity
Similar to tramadol, all methoxyl derivatives, 4a–4d, did not
display binding affinities for opioid receptors, while the phe-
nolic hydroxyl derivatives, 5a–5d, exhibited higher affinities
Figure 2. Superimposed 3D structure of tramadol and codeine; M1 and
morphine.
and selectivity against the μ opioid receptor. The binding
Acta Pharmacologica Sinica

www.chinaphar.com
Shen Q et al
npg
5
Figure 3. Ligand binding modes at the μ opioid receptor. (A) The binding mode of M1 at μ opioid receptor is shown in green; (B) The binding mode
of morphine at μ opioid receptor is shown in yellow; (C) The binding mode of N-phenethylnormorphine at μ opioid receptor is shown in blue; (D) The
superimposition of the binding modes of M1 (green), morphine (yellow) and N-phenethylnormorphine (blue); (E) The superimposition of the binding
modes of M1 (green), morphine (yellow) and N-phenethylnormorphine (blue), in which the μ opioid receptor is presented by its potential surface.
Figure 4. Designed N-arylalkylaminomethylenecyclohexanes analogues.
affinities of 5a–5d ranged from 99.7 nmol/L to 1297 nmol/L
(Table 1). We found that the μ opioid binding affinities of
5a–5d were associated with the length of the linker (number
of carbon atoms between the nitrogen atom and the intro-
duced phenyl group). When the length of the linker equaled
Figure 5. The plot of GTPγS binding assay of 5b at the µ opioid receptor.
2, the corresponding target compound 5b displayed the high- lic hydroxyl-substituted compounds were much more potent
est binding affinity at the μ opioid receptor (Ki=99 nmol/L) that methoxyl substituted compounds. For N-substitution, the
among all the aminomethylenecyclohexane analogues. The length of the linker between the nitrogen atom and the intro-
binding affinity of 5b was slightly weaker than that of M1 duced phenyl group had a substantial impact on opioid activ-
(Ki=13 nmol/L), but maintained agonistic activity (EC50=258 ity. The phenylethyl-substituted derivative 5b was the most
nmol/L) (Figure 5) equal to M1 (EC50=244.7 nmol/L), as dem- potent compound and displayed agonist activity equal to M1.
onstrated in the [35S] GTP-γ-S binding assays (Table 2). When
the length of the linker increased to 3 and 4, the agonistic model and binding mode (Figure 6A); however, the ques-
Morphine and tramadol have the same pharmocophore
activities were decreased 1-fold as compared with that of 5b.
tion remains as to whether their derivatives maintain similar
properties? To answer this question, the binding modes of
N-phenylethyl-substituted compound 5b and N-phenylethyl-
Discussion
In this study, the target compounds displayed similar struc- morphine were compared (Figure 6B and 6C). In terms of
ture and activity relationships with tramadol and that pheno- the binding modes of morphine and its analogues, N-phenyl-
Acta Pharmacologica Sinica

www.nature.com/aps
npg
Shen Q et al
6
Scheme 1. Reagents and conditions: (a) paraformaldehyde, HCl, isopropanol, 90°C; (b) 3-bromoanisole, Mg, anhydrous THF, room temperature (RT); (c)
BBr3, anhydrous DCM, -40°C.
Table 1. Binding affinities of 4a-5d compounds for μ/κ/δ (Ki or percentage displacement of radio-labeled ligand at 1 μmol/L).
Inhibition (%) or Ki (mean±SEM, nmol/L)
Compound
n
R
μa
κb
δc
(±)Tramadol
(±)M1
4a
4b
4c
4d
5a
5b
5c
9.6%±0.4%
13.0%±0.5
5.5%±0.8%
6.0%±0.1%
9.1%±0.1%
NA
1297±167.0
99.7±5.9
359.5±10.4
330.1±25.5
NAd
NA
20.2%±0.1%
3.6%±0.4%
8.1%±0.9%
28.1%±0.9%
21.3%±1.8%
1814±61.5
4781±74.5
2464±57.5
1654±89.0
19.0%±2.3%
18.1%±0.9%
21.9%±0.6%
13.0%±1.3%
11.9%±1.4%
1888±26.5
4234±303.0
NA
1
CH3
CH3
CH3
CH3
H
H
H
H
2
3
4
1
2
3
4
5d
1388±61.5
a Displacement of [3H]DAMGO from CHO cell membranes expressing human μ opioid receptor. b Displacement of [3H]U69593 from CHO cells expressing
human κ opioid receptor. c Displacement of [3H]DPDPE from CHO cell membranes expressing human δ opioid receptor. dNA indicates the binding
affinity on opioid receptor was not available.
ethylmorphine maintained the binding mode of morphine
(Figure 6B). However, N-phenylethyl tramadol 5b changed
the binding orientation of tramadol (Figure 6C). The cationic
amines of 5b still formed a salt bridge with the carboxyl group
of Asp1473.32, but the relative position of cationic amines were
Table 2. [35S]GTP-γ-S binding assays for μ opioid receptor.
Compound
Efficacy (% Emax±SD)
EC50 (nmol/L)
a
(±)Tramadol
–
–
shifted to the downside of the carboxyl group in Asp1473.32
.
(±)M1
DAMGO
5a
225.7±9.00
225.3±0.70
–
244.7±23.9
29.9±2.03
–
The N-phenylethyl group in 5b did not extend into the hydro-
phobic binding pocket formed by Trp2936.48 and Tyr3267.43, as
was observed for the phenyl group of N-phenethylnormor-
phine. Instead, it was extended into a new pocket formed
by the residues Ile1443.29, Val1433.28, and Leu219 in the second
extracellular loop (ECL2). This new interaction resulted in the
downside movement of the ligand and the phenol group in 5b
5b
185±1
258±46
5c
5d
168.9±3.90
189.7±2.10
467.0±93.7
502.7±14.8
a
[
35S]GTP-γ-S binding activity was not detected.
Acta Pharmacologica Sinica

www.chinaphar.com
Shen Q et al
npg
7
Figure 6. Ligand binding modes at the μ opioid receptor. (A) Superimposed binding modes of M1 (green) and morphine (yellow) at the μ opioid
receptor; (B) Superimposed binding modes of morphine (yellow) and N-phenethylnormorphine (blue) at the μ opioid receptor; (C) Superimposed binding
modes of M1 (green) and 5b (pink) at μ the opioid receptor. (D) The binding mode of 5b is shown in pink.
failed to form the hydrogen binding network composed of two Leu219 in ECL2. Our study indicated that the introduction of
water molecules and His2976.52 (Figure 6C and 6D). The results a new hydrogen bond donor in the phenol group of 5b may
showed that the introduction of an N-phenylethyl group restore the original water bridge hydrogen binding network
changed the original binding orientation of tramadol and mor- and increase its activity with the μ receptor.
phine; the activity of N-phenylethyl tramadol 5b decreased
8-fold, but maintained similar agonistic activity (EC50=258
Acknowledgements
nmol/L and 244 nmol/L for 5b and M1, respectively). Due to This work was supported by the Science and Technology
the downside movement of 5b in the active site, the key water- Commission of Shanghai Municipality (No 14431990500)
mediated interaction between the phenol group in M1 and and the National Natural Science Foundation of China (No
His2976.52 of the μ opioid receptor disappeared and a space 81473136 and 30901857).
was created. The introduction of a new hydrogen bond donor
in the phenol group of 5b may restore the hydrogen binding
network composed of two water molecules and His2976.52 and Wei FU, Wei LI, and Qing SHEN designed the research; Qing
increase its activity with the μ receptor. SHEN and Yuan-yuan QIAN synthesized the compounds;
Author contribution
In this study, we found that tramadol active metabolite Xue-jun XU and Jing-gen LIU performed the pharmacological
M1 and morphine had common pharmacophore features assay; and Qing SHEN and Wei FU wrote the paper.
and similar binding modes at the μ opioid receptor using 3D
structure superimposition and a molecular docking technique.
The attachment of N-phenylethyl to morphine introduced
hydrophobic interactions with Trp2936.48 and Tyr3267.43 and
improved its opioid activity. Then, a series of N-phenylalkyl
substituted derivatives of tramadol were designed, synthe-
sized and evaluated for opioid activity. The N-phenylethyl
substituted compound 5b displayed equivalent activity in
functional assays in comparison to M1. Further, a molecular
docking study with 5b identified that 5b adopted a novel
binding orientation, unlike N-phenylethyl morphine, where
the N-phenylethyl group of 5b extended into another hydro-
phobic pocket formed by the residues Ile1443.29, Val1433.28, and
References
1
Grond S, Sablotzki A. Clinical pharmacology of tramadol. Clin
Pharmacokinet 2004; 43: 879–923.
2
Harati Y, Gooch C, Swenson M, Edelman S, Greene D, Raskin P, et al.
Double-blind randomized trial of tramadol for the treatment of the
pain of diabetic neuropathy. Neurology 1998; 50: 1842–6.
Harati Y, Gooch C, Swenson M, Edelman SV, Greene D, Raskin P, et al.
Maintenance of the long-term effectiveness of tramadol in treatment
of the pain of diabetic neuropathy. J Diabetes Complications 2000;
14: 65–70.
3
4
5
Barber J. Examining the use of tramadol hydrochloride as an anti-
depressant. Exp Clin Psychopharmacol 2011; 19: 123–30.
Gobel H, Stadler T. Treatment of post-herpes zoster pain with
Acta Pharmacologica Sinica

www.nature.com/aps
npg
Shen Q et al
8
tramadol. Results of an open pilot study versus clomipramine with or 14 Yu Z, Ma YC, Ai J, Chen DQ, Zhao DM, Wang X, et al. Energetic
without levomepromazine. Drugs 1997; 53: 34–9.
factors determining the binding of type I inhibitors to c-Met kinase:
experimental studies and quantum mechanical calculations. Acta
Pharmacol Sin 2013; 34: 1475–83.
6
7
Boureau F, Legallicier P, Kabir-Ahmadi M. Tramadol in post-herpetic
neuralgia: a randomized, double-blind, placebo-controlled trial. Pain
2003; 104: 323–31.
Wu T, Yue X, Duan X, Luo DY, Cheng Y, Tian Y, et al. Efficacy and safety
of tramadol for premature ejaculation: a systematic review and meta-
analysis. Urology 2012; 80: 618–24.
15 Jones G, Willett P, Glen RC, Leach AR, Taylor R. Development and
validation of a genetic algorithm for flexible docking. J Mol Biol 1997;
267: 727–48.
16 Shao L, Abolin C, Hewitt MC, Koch P, Varney M. Derivatives of
tramadol for increased duration of effect. Bioorg Med Chem Lett
2006; 16: 691–4.
8
9
Wong BLK, Malde S. The use of tramadol “on-demand” for premature
ejaculation: a systematic review. Urology 2013; 81: 98–103.
Langley PC, Patkar AD, Boswell KA, Benson CJ, Schein JR. Adverse 17 Shao L, Wang F, Hewitt MC, Barberich TJ. mu-Opioid/5-HT4 dual
event profile of tramadol in recent clinical studies of chronic osteo-
arthritis pain. Curr Med Res Opin 2010; 26: 239–51.
10 Keating GM. Tramadol sustained-release capsules. Drugs 2006; 66:
223–30.
11 Gobbi M, Moia M, Pirona L, Ceglia I, Reyes-Parada M, Scorza C, et al.
p-Methylthioamphetamine and 1-(m-chlorophenyl)piperazine, two non-
pharmacologically active agents-efforts towards an effective opioid
analgesic with less GI and respiratory side effects (Part I). Bioorg Med
Chem Lett 2009; 19: 5679–83.
18 Buschmann H, Christoph T, Maul C, Sundermann B, editors.
Analgesics: From Chemistry and Pharmacology to Clinical Application.
Germany: Heppenheim; 2002.
neurotoxic 5-HT releasers in vivo, differ from neurotoxic amphetamine 19 Granier S, Manglik A, Kruse AC, Kobilka TS, Thian FS, Weis WI, et al.
derivatives in their mode of action at 5-HT nerve endings in vitro.
J
Structure of the delta-opioid receptor bound to naltrindole. Nature
Neurochem 2002; 82: 1435–43.
2012; 485: 400–4.
12 Reimann W, Schneider F. Induction of 5-hydroxytryptamine release by 20 Manglik A, Kruse AC, Kobilka TS, Thian FS, Mathiesen JM, Sunahara
tramadol, fenfluramine and reserpine. Eur J Pharmacol 1998; 349:
RK, et al. Crystal structure of the micro-opioid receptor bound to a
199–203.
morphinan antagonist. Nature 2012; 485: 321–6.
13 Gillen C, Haurand M, Kobelt DJ, Wnendt S. Affinity, potency and 21 Wu H, Wacker D, Mileni M, Katritch V, Han GW, Vardy E, et al.
efficacy of tramadol and its metabolites at the cloned human mu-
opioid receptor. Naunyn Schmiedebergs Arch Pharmacol 2000; 362:
116–21.
Structure of the human kappa-opioid receptor in complex with JDTic.
Nature 2012; 485: 327–32.
Acta Pharmacologica Sinica