Supporting Information

 Wiley-VCH 2013
69451 Weinheim, Germany

Occurrence of the Synthetic Analgesic Tramadol in an African

Medicinal Plant**
Ahcne Boumendjel, Germain Sotoing Tawe,* Elisabeth Ngo Bum, Tanguy Chabrol,
Chantal Beney, Valrie Sinniger, Romain Haudecoeur, Laurence Marcourt, Soura Challal,
Emerson Ferreira Queiroz, Florence Souard, Marc Le Borgne, Thierry Lomberget,
Antoine Depaulis, Catherine Lavaud, Richard Robins, Jean-Luc Wolfender, Bruno Bonaz, and
Michel De Waard*

anie_201305697_sm_miscellaneous_information.pdf
 Supporting information

Table of contents

Experimental Procedures S-1

Materials S-1
Bio-guided extraction of tramadol S-2
Acetic acid-induced abdominal constriction test S-2
Anti-nociceptive activity using formalin-induced nociception S-3
Anti-nociceptive activity using the hotplate test S-3
Tail-flick test S-3
Glutamate-induced nociception S-4
Chiral HPLC analysis S-4
UHPLC-PDA-ESI-TOF-MS analyses S-4
Isotope ratio measurement by mass spectrometry S-5
NMR analysis of the crude extract and quantification of the natural tramadol S-5
Crystal X-Ray diffraction and structure analysis S-6
Results S-6

Experimental Procedures

Materials
For structural elucidation, NMR spectra were recorded on a Bruker AC-400 instrument (1H 400 MHz).
For quantification and NMR fingerprinting, the NMR spectroscopic data were recorded on a 500 MHz
Varian Inova spectrometer. Chemical shifts are reported in parts per million (δ), with use of the
residual CD3OD signal (δH 3.31, δC 49.0) as internal standards for 1H and 13
C NMR. Coupling
constants (J) are given in Hz.
Electrospray ionization (ESI) mass spectra were acquired by the Analytical Department of Grenoble
University of Joseph Fourier on an Esquire 300 Plus Bruker Daltonis instrument with a nanospray
inlet. The crystal structure was investigated with a Kappa CCD nonius diffractometer. HR-ESI-MS
data were obtained on a Micromass-LCT Premier Time-Of-Flight Mass Spectrometer from Waters
with an electrospray interface. Semi-preparative HPLC was carried out with a Shimadzu LC-8A pump
equipped with a UV detector using an X-Bridge C18 column (5 mm, 150 × 4.9 mm, Waters). UHPLC-
PDA-TOF-MS analysis was performed on a Water Acquity UHPLC system coupled to a Waters
Micromass LCT Premier Time-Of-Flight Mass Spectrometer (Milford, MA, USA), which was

S-1
equipped with an electrospray interface. Commercial tramadol salt (HCl) used for comparison
purposes was purchased from Sigma-Aldrich.
Three independent samples of the root bark of N. latifolia were collected in the National Park of
Benoué (North Cameroon) in the dry season (April 2009, March and July 2010, February 2011). Plant
collection were performed by different people (Master 2 and PhD students and by G.S.T. himself)
from three different laboratories and three different universities (University of Buea, University of
Ngaoundere and University of Yaoundé) in order to fully exclude any source of contamination. The
plant was identified at the national herbarium (Yaoundé, Cameroon) where a voucher specimen (No.
20144/SRF/Cam) has been deposited. Samples of the plant will be made available free of charge to
any scientist that may want to confirm our findings.

Bio-guided extraction of tramadol

The freshly-collected root bark was dried in laboratory oven at 65°C and ground to a powder.
Powdered plant material (20 g) was mixed with 300 mL of methanol for 72 hrs. The extract was
filtered and the solvent was evaporated in a rotary evaporator under reduced pressure at 40°C to give 2
g of the methanol extract. The methanolic extract of N. latifolia (20 mg) was first fractionated using
semi-preparative HPLC with a reverse phase C18 column (250 × 10 mm, 10 µm; Vydac) eluted with
MeOH and H2O in a linear gradient mode from 10 to 60% MeOH in 40 min. The flow rate was 10
mL/min with UV detection at 215 nm. The semi-preparative HPLC separation yielded tens of
fractions. The fractions were dried and evaluated for their anti-nociceptive activity using the acetic
acid-induced abdominal constriction test. The most potent anti-nociceptive fraction was analysed by
thin layer chromatography (TLC) demonstrating the presence of a single compound. The preliminary
analyses by NMR and mass spectrometry, performed on the compound obtained suggested
unambiguously the presence of the 2-[(dimethylamino)methyl]-1-(3-methoxyphenyl)cyclohexanol
skeleton. The presence of a basic nitrogen atom in the compound structure prompted us to establish a
more appropriate extraction procedure, routinely used in the extraction of naturally occurring
alkaloids. For this, roots of N. latifolia (20 g of powder) were mixed with absolute ethanol (300 mL)
and refluxed for 24 hrs (Soxhlet system). Ethanol was evaporated and the residue was diluted in HCl
(5%) and extracted with dichloromethane. The aqueous solution was treated with a saturated aqueous
solution of NaHCO3 (pH 6-7). The solution was extracted with dichloromethane, which was dried over
MgSO4 and evaporated. The residue obtained was triturated with diethyl ether and filtered. The
residue was collected and dried under vacuum to yield 200 mg of pure tramadol as an oily yellow
liquid. This corresponded to a yield of 1%. Purity of the product obtained by this protocol was > 95%.

Acetic-acid-induced abdominal constriction test

Experiments in mice were performed on C57Bl/6 male mice (Janvier, Le-Genest-St-Isle, France)
weighing 26-35 g, housed in individual cages with food and water ad libidum and kept on a 12 h/12 h

S-2
light-dark cycle. All animal experimentations were carried out in accordance with the rules of the
European Committee Council Directive of November 24, 1986 (86/609/EEC) and all procedures were
approved by the local department of the veterinarian services for the use and care of animals
(agreement #380612) as well as by the ethical committee of the Grenoble Institute of Neuroscience.
All efforts were made to minimize animal suffering and moderate the number of animals used in each
series of experiments.
The methanolic fraction of N. latifolia (16, 40 or 80 mg/kg, per os) or HPLC fractions (8, 16 or 32
mg/kg, p.o.), purified tramadol isolated from N. latifolia (8, 16 or 32 mg/kg, p.o.), aspirin (150 mg/kg,
p.o.), morphine (5 mg/kg, s.c.), naloxone + methanolic fraction (1 mg/kg, intraperitoneally (i.p.) + 80
mg/kg, p.o.), naloxone + HPLC fraction (1 mg/kg, i.p. + 32 mg/kg, p.o.) or 0.9% NaCl (p.o.) were
administered to mice. One hour after oral administration of these substances, each animal was injected
i.p. with 0.6% acetic acid in a volume of 10 mL/kg. After acetic acid injection, the number of
stretching or writhing responses per animal was recorded during 30 min after a latency period of 5
min. Inhibition was expressed in percentage relative to the saline-treated control [1, 2].

Anti-nociceptive activity using the formalin-induced nociception

The formalin test was carried out as described by Hunskaar and Hole (1987) with slight modifications
[3]
. The negative control was treated with 0.9% NaCl. The positive control received indomethacin (10
mg/kg, p.o.) or morphine (5 mg/kg, s.c.), two reference analgesic compounds. Other groups of mice
were treated with methanolic fractions of N. latifolia fractions (16, 40 or 80 mg/kg, p.o.), or purified
tramadol isolated from N. latifolia (8, 16 and 32 mg/kg, p.o.). Pain was induced by injecting 50 µL of
2.5% formalin (40% formaldehyde) in distilled water in the right hind paw pad. Mice were given
different substances 1 h prior to formalin injection. Animals were individually placed in a transparent
Plexiglas observation chamber (27 × 20 × 18 cm). The amount of time spent licking and biting the
injected paw was indicative of pain and was recorded during the first 0 – 5 min (first phase), followed
by the 15 – 30 min period (second phase).

Anti-nociceptive activity using the hotplate test

The hot plate test was used to measure response latencies according to the method previously
[4]
described , with minor modifications. In these experiments, the hot plate was maintained at 55 ±
0.5°C. Each mouse (six per group) acted as its own control. One hour before treatment, the reaction
time of each mouse (licking of the forepaws or jumping response) was measured twice with a 10 min
interval. The average of the two readings was defined as the initial reaction time. The reaction time
following the administration of purified tramadol isolated from N. latifolia (8, 16 or 32 mg/kg, p.o.),
aspirine (150 mg/kg, p.o.), morphine (5 mg/kg, s.c.), naloxone + HPLC fraction (1 mg/kg, i.p. + 32
mg/kg, p.o.) and 0,9% NaCl (p.o.) was measured at 0.5, 1, 5 and 6 hrs after a latency period of 30 min.
[5]

S-3
Tail-flick test
[6]
The tail-flick test was carried out according to the method described by D’Amour and Smith . This
involved immersing the extreme 3 cm of the tail of each mouse in a water bath containing water at a
temperature of 55 ± 0.5°C. Within a few second, the mice reacted by withdrawing the tail. The
reaction time was recorded with a stopwatch. The mice were treated with purified tramadol isolated
from N. latifolia (8, 16 or 32 mg/kg, p.o.), aspirin (150 mg/kg, p.o.), morphine (5 mg/kg, s.c.),
naloxone + HPLC fraction (1 mg/kg, i.p. + 32 mg/kg, p.o.) and 0.9% NaCl (p.o.). The reaction time of
each mouse was taken at intervals 15, 30 and 60 min after a latency period of 1 hr following the
administration of the decoction and drugs.

Glutamate-induced nociception
Animals were treated with the tramadol purified from N. latifolia (8, 16 or 32 mg/kg, p.o.), dipyrone
(60 mg/kg, p.o.) or 0.9% NaCl (p.o.) 1 h before the test. A volume of 20 µL of glutamate (30
µmol/paw) was injected intraplantarly in the ventral surface of the right hind paw. Animals were
observed individually for 15 min following glutamate injection. The amount of time they spent licking
the injected paw was recorded with a chronometer and was considered as indicative of nociception [7].

Chiral HPLC analysis

The HPLC system consistent of an Agilent 1290 system with a binary pump, degasser, autosampler,
thermostated (30°C) column compartment and an Agilent 1260 Infinity Diode Array Detector
(detection wavelength 270 mm). The column was a Chiralpak® AD-H (Daicel Chemical Industries, 4.6
mm × 250 mm, 5 µm). Analytes were eluted with an isocratic mobile phase (heptane/i-
PrOH/diethylamine = 97/3/0.1, v/v/v) at a flow rate of 1 mL/min [8].

UHPLC-PDA-ESI-TOF-MS analyses
Metabolite profiling of the constituents of the crude methanolic and the alkaloid extract of N. latifolia
was performed by UHPLC-PDA-TOF-MS. This also allowed the analysis of the isolated tramadol and
the commercial reference tramadol. The analytical platform consisted of a Water Acquity UHPLC
system coupled to a Waters Micromass LCT Premier Time-of-Flight mass spectrometer (Milford, MA,
USA), equipped with an electrospray interface (ESI) and a PDA detector. Compound separations were
performed on a C18 column (Waters Acquity UPLC BEH C18, 50 mm x 10 mm, 1.7 µm). The mobile
phase consisted of H2O (A) and acetonitrile (B), each containing 0.1% formic acid (v/v). For the high
resolution profiling, the following gradient was used: from 5-95% B in 30 min at a flow rate of 460
µL. The auto sampler and the column oven were set at 10 and 40°C, respectively. UV analyses were
performed between 210 and 450 nm, and 1 µL of each sample was injected. Solution concentration
was 1 mg/mL in MeOH for the extract and 1 µg/mL for the commercial tramadol. Injection volume

S-4
was 0.5 and 1 µL, respectively. Detection was performed in both positive and negative ionisation
modes in the 100-1000 Da range with acquisition times of 0.3 sec in centroid mode. The ESI
conditions were set as follows: capillary voltage 2800 V, cone voltage 40 V, source temperature
120°C, desolvatation temperature 330°C, cone gas flow 20 L/h, desolvatation gas flow 600 L/h, and
MCP (micro channel plate) detector voltage 2400 V. Data were processed using MassLynx software,
version 4.1 SCN#639 (Waters Corporation, Milford, MA, USA). For LC-peak annotation, all possible
molecular formulae were extracted (elements C; H, N, O, tolerance of 15 ppm) with the Elemental
Composition tool of Mass Lynx. Molecular formulae were further validated by heuristic filtering. To
display all ions recorded, a 2D-map was performed on MZ mine software (version 2.9.1, Free software
Foundation, Boston, MA 021110-1301, USA).

Isotope ratio measurement by mass spectrometry

Sample was dissolved in MeOH and an aliquot containing approximately 0.6 mg of compound was
pipetted into each of two tin capsules (solids “light” 5x9 mm, Thermo Fisher Scientific). After
complete evaporation of solvent with warm N2 gas, the precise mass was measured.
13
The C/12C and 15
N/14N ratios were obtained by isotope ratio measurement by mass spectrometry
(irm-MS) using a Sigma2 spectrometer (Sercon Instruments, Crewe, UK) linked to a Sercon elemental
analyser. Isotope ratios (δ13C and δ15N (‰)) were expressed relative to the international references
using the following equation:
æ R ö
d (‰) = çç - 1÷÷ ´ 1000
è Rstd ø
where for δ13C the reference (Rstd) is Vienna-Pee Dee Belemnite (V-PDB) and for δ15N it is
atmospheric N2. The calibrated international reference materials NBS-22, SUCROSE-C6, and PEF-1
(IAEA, Vienna, Austria) were used for δ13C calibrations, via a laboratory standard of glutamic acid.
The calibrated international reference materials IAEA-N1 or IAEA-N2 (IAEA, Vienna, Austria) were
used for δ15N calibrations, via a laboratory standard of glutamic acid.

NMR analysis of the crude extract and quantification of the natural tramadol
For the quantitative studies, the PULCON method[9] was used to measure the concentration of
tramadol in the crude extract without using an internal reference. This quantitative NMR method
correlates the absolute intensity of NMR signals in different samples by the measurement of 360° r.f.
pulse. A solution with a known concentration of tramadol (263.4 mM) was used as external reference.
A recycle delay of 20 sec and an acquisition time of 2 sec were chosen for each sample in order to
assume a full relaxation of protons. The optimal pulse width at 90° (precisely measured for each
sample at 360°) was used to record each spectrum (i.d. 9.025 msec for the crude extract and 8.625 msec
for the tramadol solution). FIDs were Fourier transformed without apodization function, the resulting
spectra were manually phased, baseline corrected using a 3rd order Bernstein polynomial function and

S-5
calibrated to the residual methanol peak at d 3.31 using MestReNova (version 8.0.2, Mestrelab
Research S.L.). The signals were integrated using the peak picking algorithm of MestReNova which
take into account only the integrals of the GSD (Global Spectral Deconvolution) spectrum.

Crystal X-Ray diffraction and structure analysis

Extracted tramadol was converted to its chlorhydrate salt and crystalized in acetonitrile. These were
used to obtain a single-crystal X-ray diffraction pattern.

Results

Supporting Figure 1: Extraction of an active anti-pain component from N. latifolia root bark. a)
HPLC-UV profile of the crude extract at 210 nm. Major
peaks are numbered and fractions (F25 to F29) investigated
for the characterization of the anti-pain compound are
shown by the red circle. Flow rate 10 mL/min, fraction size
500 µL, and fraction sampling was 20 fractions/min. b)
Effects of F27 of N. latifolia on the writhing induced by
acetic acid. Results are expressed as mean ± S.E.M (n=6).
Statistical analyses were performed on absolute data. The
fractions began manifesting their assuaging effects on the
writhing reflex 45 min following administration. *, P <
0.05, ***, P < 0.001 significantly different compared to the
control groups. Data were analysed by two-way Anova
followed by Tukey’s (HSD) multi-comparison test. c)
Influence of various concentrations of N. latifolia F27 on
acetic acid-induced writhing and absence of counter effect
of naloxone. Results are expressed as mean ± S.E.M (n=6).
Statistical analyses were performed on data obtained 45 min
following administration of F27. **, P < 0.01; ***, P <
0.001 significantly different compared to the control groups.
Data were analysed by two-way Anova followed by
Tukey’s (HSD) multi-comparison test. Abbreviations are
ASA, aspirine; Morph, morphine; Nalox, naloxone.

Supporting Table 1: Influence of Nauclea latifolia fraction 27 on acetic acid-induced writhing

and hotplate-induced pain in mice.

S-6
The extract, at all doses used, began manifesting its assuaging effect on the writhing reflex 1 hr
following administration. *, P < 0.05; **, P < 0.01; ***
, P < 0.001, significantly different compared to
the control group. Data were analysed by two-way Anova followed by the Tukey’s (HSD) multi-
comparison test.

Acetic acid- Hotplate-induced pain in mice

induced writhing
Treatments Doses (mg/kg) Inhibition (%) Protection against thermal nociception (%)
0 hr 0.5 hr 1 hr 2 hrs 3 hrs 4 hrs 5 hrs 6 hrs
NaCl _ _ -6.6 -2.2 3.9 -0.6 10.0 3.3 9.6
_ * **
N. latifolia 8 19.3 -2.2 22.9 20.2 50.6 4.6 -0.2 -0.5
_ * * **
N. latifolia 16 43.7 1.9 27.0 21.3 52.4 6.5 -2.0 4.6
_ ** ** *** * * *
N. latifolia 32 56.8 1.9 45.2 50.9 71.9 25.2 21.3 27.8
_ * * *
Aspirin 150 52.9 1.7 35.5 2.2 39.5 8.5 10.6 3.5
_ ** ** ** ** *
Morphine 5 59.2 7.4 44.4 46.8 40.4 43.6 24.6 17.1
_ * * ** ** * **
N. latifolia + Naloxone 32 + 1 49.0 -2.1 33.8 30.6 58.9 42.4 23.7 27.7

Supporting Figure 2: Anti-nociceptive activity of N. latifolia fraction F27 in various animal pain
models. a) Influence of N. latifolia fraction F27 on
formalin-induced pain. Results are expressed as mean ±
S.E.M. (n=6). The amount of time spent licking and biting
the injected paw was indicative of pain and was recorded
over 0-5 min (first phase) and 15-30 min (second phase). *,
P < 0.05; **, P < 0.01; ***, P < 0.001 are significantly
different compared to the control group. Data were analysed
by two-way Anova followed by Tukey’s (HSD) multi-
comparison test. b) Influence of N. latifolia fraction 27 on
hotplate-induced pain in mice. Results are expressed as
mean ± S.E.M. c) Influence of N. latifolia fraction F27 on
tail flick response in mice after immersion in a 55°C water
bath. Results are expressed as mean ± S.E.M.

Comments: Formalin test

F27 from N. latifolia had analgesic effects on both the first
(0–5 min) and second phases (15–30 min) as shown in
Supporting Figure 2a. These phases correspond to
neurogenic and inflammatory pains, respectively. The
neurogenic-induced pain blockade is greatest at 32 mg/kg
(64.07%, [F(6, 27) = 95.17; p< 0.001] while it initiates at 8
mg/kg. Similarly 32 mg/kg of F27 significantly blocks pain
emanating from inflammation (31.74%,) [F(6, 27) = 95.17;
P<0.05].

S-7
Comments: Hot plate-test
Supporting Figure 2b shows that the purified fraction isolated from N. latifolia lead to a marked
increase in the latency response in the hot plate algesiometer model of nociception, with the maximal
effect was observed for the higher dose administered (32 mg/kg) and in the later times after oral
administration (1 - 3 h). At 60 min after oral administration, a significant increase could be observed
[F(6, 43) = 127.05; p<0.001] in baseline that reached its maximal level at 3 h. Naloxone antagonized
the anti-nociceptive effect of the purified tramadol isolated from N. latifolia in the hot plate assay
procedures.

Comments: Tail-flick test

After a latency period of 30 min following oral administration of fraction F27 at a dose of 32 mg/kg
(63.05%), [F(6, 57) = 97.24; p<0.001], there was a significant dose-dependent reduction of the
sensation of pain due to tail immersion in warm water (Supporting Figure 2c). The inhibitory effects
of fraction F27 became pronounced between 30 and 60 min post-dosing and reached a maximum of
63.5% [F(6, 57) = 97.24; p<0.001] with the dose of 800 mg/kg. The analgesic activity of the extract
was blocked by naloxone.

Supporting Table 2: Influence of N. latifolia fraction F27 on the tail flick response in mice.
Doses (mg/kg) Percentage of protection
0 min 15 min 30 min 60 min

NaCl - - -18.8 -11.6 -16.1

* *
N. latifolia 8 - 7.7 14.4 21.4
** ***
N. latifolia 16 - 5.6 34.3 42.9
*** ***
N. latifolia 32 - 18.7 63.0 63.5

Aspirin 150 - 6.4 9.5 7.1

*** ***
Morphine 5 - 27.3 62.1 77.8
*** ***
N. latifolia + Naloxone 32 + 1 - 19.5 58.3 62.2

Results are expressed as percentage of protection against thermal nociception. Statistical analyses were
performed on absolute data, n = 6. *, P < 0.05; **
, P < 0.01; ***
, P < 0.001 significantly different
compared to the control group. Data were analyzed by two-way Anova followed by Tukey’s (HSD)
multi-comparison test.

Supporting Figure 3: Influence of N. latifolia fraction F27 on glutamate-induced pain. Results are
expressed as mean ± S.E.M. (n=6). The amount of time spent licking and biting the injected paw was
indicative of pain. *, P < 0.05, **, P < 0.01, ***, P < 0.001 significantly different compared to the

S-8
 100
250 NaCl
8 mg/kg control group. Data were analysed by two-
16 mg/kg
32 mg/kg 80 way Anova followed by Tukey’s (HSD)
200
Licking time (sec)

Dipyrone ***
***

Inhibition (%)
*
**
multi-comparison test.
60
150

** *
40
100 ** Comments: Glutamate-induced
***

50 20
nociception
Interestingly, in the glutamate-induced
0 0
nociception in mice, fraction F27 caused
Treatments Treatments
marked and dose-related anti-nociception
(Supporting Figure 3). The calculated mean ID50 values, (and its 95% confidence limits), and the
maximal inhibition were 21.2 (6.5 – 47.2) and 66.6% [F(4, 62) = 105.12; p<0.001], respectively.
Given orally, dipyrone produced a significant inhibition of 70.3% [F(4, 62) = 105.12; p<0.001] of the
glutamate-induced nociception in mice (Supporting Figure 3).

Supporting Figure 4: MS and NMR spectra of tramadol isolated from N. latifolia. a) Positive
chemical ionization mass spectrometry (CIMS) spectrum of tramadol. b) 1H-NMR spectrum. c) 13C-
NMR spectrum. d) COSY spectrum. e) HMBC spectrum. f) DEPT spectrum.

Supporting Figure 5: UHPLC-TOF-MS analysis of the crude ethanolic extract of root bark of N.
latifolia. (ESI positive ion mode). a) Extracted ion chromatogram at m/z 306.16 +/- 0.10
corresponding to the [M+H]+ of nauclechine demethoxycarbonyl. b) Extracted ion chromatogram at
m/z 288.11 +/- 0.10 corresponding to the [M+H]+ of nauclefine. c) BPI trace of the crude ethanolic
extract showing the detection of vincosamide ([M+H]+: 499.21) and tramadol ([M+H]+: 264.19).

S-9
a 306.16

2500
Nauclechine demethoxycarbonyl

2000
TOF-MS

1500

1000

500

0
0 10 20 30 40

Time (min)
b 288.11

2500 Nauclefine

2000
TOF-MS

1500

1000

500 314.13
152.04

0
0 10 20 30 40

Time (min)
c 264.19
2.5e+5 Tramadol

2.0e+5

Vincosamide
TOF-MS

1.5e+5

1.0e+5
499.21
280.20

5.0e+4
348.18
338.35 226.02

0.0
0 10 20 30 40 50

Time (min)

S-10
Supporting Figure 6: Putative pathway for the biosynthesis of tramadol

The proposed biosynthetic pathway implies the formation of two key intermediates (acetophenone part
A and an enaminopentanal part B). The moiety A could readily originate either from acetyl-CoA or
from the phenylalanine pathway. The presence of a 3’-methoxyl group in the acetophenone part is
rather rare but a few reports underline the presence of 3’-methoxyacetophenone as a secondary
metabolite.
The moiety B of the molecule could readily originate from lysine via the classic amine pathway.
Decarboxylation of lysine gives the diamine, cadaverine (1,5-dimaniopentane), a well-described
precursor of several alkaloids, such as the quinolizidine alkaloids and minor Nicotiana alkaloids. In
the formation of the heterocyclic Nicotiana alkaloids, such as anabasine, it has been shown that
cadaverine undergoes oxidative deamination, providing iminopentanal. This can readily isomerize to
enaminopentanal. It should be noted that anabasine is an example of a racemic natural product. N-
Dimethylation probably occurs on the amine. The reaction between A and B could occur through a
cascade of reactions including aldolization/crotonization, addition, and reduction finally to provide
tramadol. We suggest that the racemate may be produced at the cyclisation step.

Supporting Information References

[1] L. C. Hendershot, J. Forsaith, J Pharmacol Exp Ther 1959, 125, 237.
[2] V. Sinniger, C. Porcher, P. Mouchet, A. Juhem, B. Bonaz, Pain 2004, 110, 738.
[3] S. Hunskaar, K. Hole, Pain 1987, 30, 103.
[4] N. B. Eddy, D. Leimbach, J Pharmacol Exp Ther 1953, 107, 385.
[5] M. C. Lanhers, J. Fleurentin, P. Dorfman, F. Mortier, J. M. Pelt, Planta Med 1991, 57, 225.
[6] F. E. D'Amour, Smith, D.L., J. Pharmacol. Exp. Ther. 1941, 72, 74.

S-11
[7] A. Beirith, A. R. Santos, J. B. Calixto, Brain Res 2002, 924, 219.
[8] B. Elsing, G. Blaschke, J Chromatogr 1993, 612, 223.
[9] N. Bohni, M. L. Cordero-Maldonado, J. Maes, D. Siverio-Mota, L. Marcourt, S. Munck, A. R.
Kamuhabwa, M. J. Moshi, C. V. Esguerra, P. A. de Witte, A. D. Crawford, J. L. Wolfender, PLoS One, 8,
e64006.

S-12