Levonantradol: Asymmetric synthesis and structural analysis

Article abstract of DOI:10.1039/c3cc41388h

The first asymmetric synthesis of a synthetic cannabinoid levonantradol was accomplished, and the 3-D solution structure of its core architecture was confirmed by NMR and computational methods. The Royal Society of Chemistry 2013.

Full text of DOI:10.1039/c3cc41388h

ChemComm
View Article Online
View Journal | View Issue
COMMUNICATION
Levonantradol: asymmetric synthesis and structural
analysis†
Cite this: Chem. Commun., 2013,
49, 3685
Andrey E. Sheshenev, Ekaterina V. Boltukhina and King Kuok (Mimi) Hii*
Received 22nd February 2013,
Accepted 15th March 2013
DOI: 10.1039/c3cc41388h
www.rsc.org/chemcomm
The first asymmetric synthesis of a synthetic cannabinoid levonan-
The analgesic potency of the molecule was found to be highly
tradol was accomplished, and the 3-D solution structure of its core stereospecific, being dependent upon the trans-(6a, 10a) stereo-
architecture was confirmed by NMR and computational methods. chemistry along with b-orientation of substituents at C-6 and C-9.6
Optically pure form of the compound was found to be twice as
The use of cannabis for therapeutic and recreational purposes dates potent as its racemic mixture, while the epimeric form, containing
back several millennia, but its role in the modern society remains a the opposite stereochemistry at the C-3 side chain, is ten times less
topic of great debate and controversy.1 Clinically, cannabinoids have efficacious. Although the clinical use of levonantradol was curtailed
been approved for use as anti-emetics and analgesics for cancer, by its unacceptable toxicity,7 it remains an important tool for the
glaucoma and multiple sclerosis sufferers; potential therapeutic study of the pharmacology of cannabinoid compounds.8
uses have also been suggested for the treatment of epilepsy,2a
The synthesis of levonantradol was reported more than 30 years
Tourette’s syndrome,2b Huntington disease2c and cancer.2d,e How- ago,4c where the optically active form of the compound was
ever, long-term use of cannabis has also been associated with obtained by performing four kinetic resolutions over the course
undesirable psychiatric side effects such as schizophrenia, bipolar of the synthesis consisting of 10 linear synthetic steps (overall
disorder and major depression.2f Consequently, there is a great deal yield unspecified). Despite plethora of works devoted to the
of interest in the development of synthetic variants that can activate investigation of the biological activities of levonantradol, an
CB1 and CB2 cannabinoid receptors without inducing psychosis.3
improved synthesis has never been reported; moreover, the
Delta-9-tetrahydrocannabinol (D9-THC) is the principal characterisation data for levonantradol are surprisingly scant.
psychoactive ingredient in cannabis, from which levonantradol Most notably, the stereochemistry of levonantradol was assigned
(CP-50,556-01) was derived in the 1980s by Pfizer scientists based on an unpublished X-ray structure of 30-epilevonantradol;4a
(Fig. 1).4 As
levonantradol is 30 times more potent than THC in activating
a
structurally novel synthetic cannabinoid, independent verification of its structure has never been reported.
Herein, we will report the first enantioselective synthesis of
the CB1 and CB2 receptors and possesses analgesic and antiemetic levonantradol, followed by its full characterisation and verifica-
properties with 9–14 times greater potency than morphine.5
tion of its conformational and stereochemical structures by a
combination of computational and NMR techniques.
It was envisaged that the core structure of the target mole-
cule could be constructed enantioselectively in three key steps
(Fig. 2). The key idea was to use the stereocenter created at C-3
in the aza-Michael adduct (corresponds to C-6 in 2) to direct the
selectivity at C-6a, 9 and 10a in a chiral relay process.
The chirality on the C-3 side chain required the preparation
of (S)-5-phenyl-2-pentanol, (S)-1. This could be achieved by
asymmetric synthesis, however, such approaches are either
time-consuming or too expensive.9 Conversely, kinetic resolution
of the precursor can be achieved by using a lipase,10a or by the
resolution of the hemiphthalate derivative using (+)-brucine.10b
In this work, an effective procedure was devised using the less
toxic and inexpensive a-methylbenzylamine as the resolving
agent (Scheme 1). This approach afforded (S)-(+)-1 in 97% optical
purity, from which corresponding mesylate 6 can be derived.
Fig. 1 Structures of THC and levonantradol.
Department of Chemistry, Imperial College London, Exhibition Road,
South Kensington, London SW7 2AZ, UK. E-mail: mimi.hii@imperial.ac.uk;
Fax: +44 (0)2075945804
† Electronic supplementary information (ESI) available: Experimental details. See
DOI: 10.1039/c3cc41388h
c
This journal is The Royal Society of Chemistry 2013
Chem. Commun., 2013, 49, 3685--3687 3685

View Article Online
Communication
ChemComm
Scheme 2 Asymmetric Pd-catalysed aza-Michael reaction. (a) [(R-BINAP)-
Pd(m-OH)]2[OTf]2 (2 mol%), THF, ꢀ10 1C, 48 h, then 0 1C, 24 h; (b) single
recrystallization from toluene.
methyl vinyl ketone. Decarbonylation of the Michael adduct
occurred under very mild conditions to give compound 10 as a
3 : 2 mixture of epimers. This material was subjected to cyclization
accompanied by concomitant removal of the N-formyl group
under basic conditions, furnishing the target tricycle 11. The
endo-cyclic CQC bond was then reduced under dissolved Li-metal
conditions to afford a 2 : 1 mixture of trans–cis-isomers across the
fused six-membered rings (at C-6a and C-10a). Following acylation
of the phenol group, the requisite trans-isomer 12 was isolated in
55% yield. Finally, reduction of the ketone at C-9 with sodium
borohydride at ꢀ60 1C ensued with excellent stereoselectivity to
give a single diastereoisomer of levonantradol with enantiomeric
purity of 99% which was confirmed by chiral HPLC (see ESI†).
The prepared levonantradol was converted into the hydrochloride
salt and its melting point and optical rotation corresponded well
with the values reported by Pfizer in 1981,13 however, further com-
parisons with a genuine sample were not possible.14 Nevertheless,
Fig. 2 Retrosynthesis of levonantradol.
Asymmetric conjugate addition of 3,5-dimethoxyaniline to a
chelating Michael acceptor 3 was achieved by using a Pd-catalyzed
process developed in our laboratory.11 Guided by our previous
work,11c the (R)-isomer of the catalyst was employed to induce the
requisite (S)-stereochemistry at C-3 in 7 (Scheme 2). Due to the
nucleophilicity of the aniline substrate, triflate salt was employed in
conjunction with the dimeric [(R-BINAP)Pd(m-OH)]2[OTf]2, relying on
the Brønsted basicity of the catalyst to control the release of the
nucleophile during the reaction.12 Competitive uncatalyzed process
was further suppressed by conducting the reaction at low tempera-
ture. Under these conditions, the Michael adduct 7 can be obtained
with excellent yield (93%) and enantioselectivity (97.5% ee). Optically
purity can be achieved after a single recrystallization from toluene.
With the chiral precursors in hand, the final assembly of the
target molecule commenced with the hydrolysis of the Michael
adduct 7 to the N-aryl amino acid 8. Treatment of 8 with
HBr–acetic acid mixture triggered an electrophilic cyclisation,
accompanied by the spontaneous deprotection of the phenolic
groups, to afford the quinolone 2 (Scheme 3).
Coupling between 2 (100% ee) and 6 (97% ee) was conducted in
the presence of a weak inorganic base. The single diastereoisomer
9 was obtained with complete inversion of the S-stereochemistry at
C-20 centre of C-3 chain with 97% ee (verified by chiral HPLC; see
ESI†). Further deprotonation of 9 with sodium hydride followed by
the treatment with methyl formate resulted in the introduction of
N-formyl and a-formyl ketone groups; the latter had the requi-
site reactivity to undergo conjugate addition reaction with
1
the H and 13C NMR spectra of the synthesised molecule contain
distinct resonance signals, which can be fully assigned by COSY,
HMQC and HMBC experiments (Table S2, ESI†). This allowed the
relative stereochemistry of levonantradol to be established based on
the NOE experiments. Relative positions of all the a-Hs in the
carbocyclic ring C can be established by the selective irradiation of
the H-9 resonance signal, while the irradiation of H-6 and H-10a
confirmed their positions relative to the S-configuration at C-6 (ESI†).
Consequently, 3-dimensional structure of levonantradol can
be simulated computationally by applying a Hartree–Fock energy-
minimisation using a 3-21G basis set. The optimised structure
was subjected to DFT calculations by employing GAIO method
with the B3LYP/6-311+(2d,p) basis set, to predict NMR properties.
Pleasingly, the predicted d and J values (Tables S2 and S3
respectively; ESI†) fitted well with that observed experimentally.
Simultaneously, the model also provided inter-proton distances
that showed excellent correlation with the observed NOE values
(Table S4, ESI†). There is very good evidence that the fused rings
B and C of levonantradol are locked in a rigid, flattened boat–
chair conformation in solution (Fig. 3). Most notably, the
carbocyclic ring C in levonantradol (a boat conformation) is
significantly different from the equivalent cyclohexene ring in
the O-tosyl derivative of D9-THC (a distorted screw boat).15
In summary, asymmetric synthesis of levonantradol is
achieved, for the first time, in 11 linear steps with 16.2% overall
yield. This was accomplished by devising a new kinetic resolu-
tion process for the preparation of the (S)-1, combined with an
Scheme 1 Kinetic resolution of alcohol 1 and preparation of optically pure
precursor 6. (a) Phthalic anhydride, Py, 94 1C; (b) (S)-a-methylbenzylamine, EtOAc,
60 1C; (c) recrystallization (Table S1, ESI†); (d) 1.5 M aq. HCl; (e) 3.75 M aq. NaOH;
(f) MsCl, NEt3, THF, 0 1C to RT.
c
3686 Chem. Commun., 2013, 49, 3685--3687
This journal is The Royal Society of Chemistry 2013

View Article Online
ChemComm
Communication
Scheme 3 Final assembly of levonantradol. (a) 1.0 M KOH–MeOH, 3 h; (b) AcOH–HBr (1 : 1, v/v), reflux, 1.5 h; (c) 6, K2CO3, DMF, 80 1C, 2.5 h; (d) NaH, toluene,
HCO2Me, 20 h; (e) MVK, NEt3, CH2Cl2, 20 h; (f) K2CO3, MeOH, 0 1C, 4 h (76% over steps d–f); (g) MeONa, MeOH, reflux, 48 h; (h) Li–NH3, THF, ꢀ70 1C, 20 min; (i) Ac2O,
NEt3, 4-DMAP, CH2Cl2, 0 1C, 1 h; (j) NaBH4, EtOH–THF (1 : 1, v/v), ꢀ60 1C, 30 min.
2009, CD006565; (c) O. Sagredo, M. R. Pazos, S. Valdeolivas and
J. Fernandez-Ruiz, Recent Pat. CNS Drug Discovery, 2012, 7, 41;
(d) G. Velasco, C. Sanchez and M. Guzman, Nat. Rev. Cancer, 2012,
12, 436; (e) G. Calvaruso, O. Pellerito, A. Notaro and M. Giuliano, Int.
J. Oncol., 2012, 41, 407; ( f ) K. A. Seely, P. L. Prather, L. P. James and
J. H. Moran, Mol. Interventions, 2011, 11, 36.
3 (a) L. O. Hanus and R. Mechoulam, Curr. Med. Chem., 2010,
17, 1341; (b) S. Tambaro and M. Bortolato, Recent Pat. CNS Drug
Discovery, 2012, 7, 25.
4 (a) G. M. Milne and M. R. Johnson, J. Clin. Pharmacol., 1981, 21, S367;
(b) H. M. McIlhenny, R. W. Mast, M. R. Johnson and G. M. Milne,
J. Pharmacol. Exp. Ther., 1981, 219, 363; (c) M. R. Johnson and
G. M. Milne, J. Heterocycl. Chem., 1980, 17, 1817; (d) M. R. Johnson
Fig. 3 Solution structure of levonantradol’s fused rings established by compu-
and G. M. Milne, U.S. Patent, US4228169, 1980.
5 (a) R. Stuartharris, C. Mooney and I. E. Smith, Br. J. Cancer, 1983, 47, 568;
(b) R. C. Stuartharris, C. A. Mooney and I. E. Smith, J. Clin. Oncol., 1983,
9, 143; (c) M. L. Citron, T. S. Herman, F. Vreeland, S. H. Krasnow,
B. E. Fossieck, S. Harwood, R. Franklin and M. H. Cohen, Cancer Treat.
Rep., 1985, 69, 109; (d) M. L. Citron, T. S. Herman, B. E. Fossieck,
S. H. Krasnow, F. Vreeland, S. Harwood, L. G. Ortega and M. H. Cohen,
Proc. Am. Assoc. Cancer Res., 1983, 24, 165.
6 The original numbering system is adopted in this work.
7 R. A. Joss, R. L. Galeazzi, A. Bischoff, D. D. Do, A. Goldhirsch and
K. W. Brunner, Cancer Chemother. Pharmacol., 1982, 9, 61.
8 (a) F. C. M. Rocha, S. C. Stefano, R. D. C. Haiek, L. M. Q. R. Oliveira
and D. X. Da Silveira, Eur. J. Cancer Care, 2008, 17, 431;
(b) J. C. Ashton and E. D. Milligan, Curr. Opin. Invest. Drugs, 2008,
9, 65; (c) M. B. Amar, J. Ethnopharmacol., 2006, 105, 1; (d) J. R. Ineck
and T. L. Burns, Ann. Pharmacother., 2006, 40, 251.
9 (a) B. W. Cue Jr. and B. S. Moore, US Patent, US4412958, 1983;
(b) R. Kuwano, M. Sawamura, J. Shirai, M. Takahashi and Y. Ito, Bull.
Chem. Soc. Jpn., 2000, 73, 485; (c) R. Kuwano, T. Uemura, M. Saitoh
and Y. Ito, Tetrahedron: Asymmetry, 2004, 15, 2263; (d) Y. Uozumi,
K. Kitayama, T. Hayashi, K. Yanagi and E. Fukuyo, Bull. Chem. Soc.
Jpn., 1995, 68, 713.
tational chemistry and NMR (C-3 side chain was omitted).
enantioselective aza-Michael reaction for the preparation of
quinolone intermediate 2. The two precursors were then used
to construct the target molecule in a chiral relay process, thus
eliminating the need for costly separation processes during the
synthesis. It is envisaged that structural modifications can be
easily introduced for future medicinal chemistry programmes.
Finally, the solution structure of levonantradol was derived
successfully from
a combination of NMR experiments
with DFT calculations, providing confirmation of its stereo-
chemistry and conformation. Interestingly, the 3-D structure of
levonantradol was found to be significantly different from
D9-THC from which it was derived. This may have important
implications in the understanding of its biological activity, and
the subsequent development of non-toxic, but equally or more
potent analogues.
10 (a) K. Burgess and L. D. Jennings, J. Am. Chem. Soc., 1991, 113, 6129;
(b) B. S. Moore, U.S. Patent, US4386205, 1983.
11 (a) K. L. Li, P. N. Horton, M. B. Hursthouse and K. K. Hii,
J. Organomet. Chem., 2003, 665, 250; (b) K. L. Li, X. H. Cheng and
K. K. Hii, Eur. J. Org. Chem., 2004, 959; (c) P. H. Phua, S. P. Mathew,
A. J. P. White, J. G. de Vries, D. G. Blackmond and K. K. Hii,
Chem.–Eur. J., 2007, 13, 4602; (d) M. Guino, P. H. Phua, J. C. Caille
and K. K. Hii, J. Org. Chem., 2007, 72, 6290; (e) L. L. Taylor,
F. W. Goldberg and K. K. Hii, Org. Biomol. Chem., 2012, 10, 4424.
12 Y. Hamashima, H. Somei, Y. Shimura, T. Tamura and M. Sodeoka,
Org. Lett., 2004, 6, 1861.
AES is supported by a Marie Curie Intra-European Fellow-
ship within the 7th EC Framework Programme (252247). We
thank Johnson Matthey plc for the loan of Pd salts, and Prof.
Henry Rzepa and the EPSRC UK national Service for Computa-
tional Software (NSCCS) for help with calculations.
Notes and references
13 M. R. Johnson, U.S. Patent, US4260764, 1981.
1 M. Booth, Cannibis: a history, Bantam Books, 2003.
2 (a) D. Gloss and B. Vickrey, Cannabinoids for epilepsy, Cochrane Database
14 Pfizer no longer supplies the molecule (personal communication with
Dr Ruth McKernan, CSO Neusentis, Senior Vice-President, Pfizer).
Syst. Rev., 2012, CD009270; (b) A. Curtis, C. E. Clarke and H. E. Rickards, 15 W. Gul, P. Carvalho, D. W. Berberich, M. A. Avery and M. A. ElSohly,
Cannabinoids for Tourette’s Syndrome, Cochrane Database Syst. Rev.,
Acta Crystallogr., Sect. E: Struct. Rep. Online, 2008, 64, O1686.
c
This journal is The Royal Society of Chemistry 2013
Chem. Commun., 2013, 49, 3685--3687 3687