Short and Divergent Total Synthesis of (+)-Machaeriol B, (+)-Machaeriol D, (+)-Δ8-THC, and Analogues

Article abstract of DOI:10.1002/anie.201502595

Short and highly efficient stereoselective syntheses provide machaeriols and cannabinoids in a divergent approach starting from a common precursor, commercially available (S)-perillic acid. Key features of the novel strategy are a stereospecific palladium-catalyzed decarboxylative arylation and a one-pot sequence comprising a stereoselective hydroboration followed by oxidation or reduction of the corresponding intermediary boranes. The divergent approach is convincingly demonstrated by the five-step syntheses of (+)-machaeriol B, (+)-machaeriol D, and related analogues, and the four-step synthesis of (+)-Δ⁸-THC and an analogue.

Full text of DOI:10.1002/anie.201502595

Angewandte
Chemie
International Edition: DOI: 10.1002/anie.201502595
German Edition: DOI: 10.1002/ange.201502595
Natural Product Synthesis
Short and Divergent Total Synthesis of (+)-Machaeriol B,
(+)-Machaeriol D, (+)-D8-THC, and Analogues**
Felix Klotter and Armido Studer*
Abstract: Short and highly efficient stereoselective syntheses
provide machaeriols and cannabinoids in a divergent approach
starting from a common precursor, commercially available
(S)-perillic acid. Key features of the novel strategy are
a stereospecific palladium-catalyzed decarboxylative arylation
and a one-pot sequence comprising a stereoselective hydro-
boration followed by oxidation or reduction of the corre-
sponding intermediary boranes. The divergent approach is
convincingly demonstrated by the five-step syntheses of
(+)-machaeriol B, (+)-machaeriol D, and related analogues,
and the four-step synthesis of (+)-D8-THC and an analogue.
some of them have found applications as psychotomimetic or
other drugs.[5] Owing to their great pharmacological potential,
considerable efforts have been made towards the synthesis of
naturally and structurally related, non-natural cannabinoids
(e.g., by the groups of Trost,[6] Zhou,[7] and Bräse[8]). Recently,
Carreira et al. presented an elegant approach to the four
stereoisomers of D9-THC by using stereodivergent dual
catalysis in the key step.[9] Efficient synthetic procedures for
members of this important natural product class open the
door for further investigations of the pharmacology of natural
and—thanks to the synthesis approach—also unnatural
isomers.
S
ince the discovery of the biological and medical impor-
As a novel class of cannabinoid-related natural products,
machaeriols, such as (+)-machaeriol B (3) and (+)-machaer-
iol D (4), were isolated from the stem bark of Machaerium
multiflorum between 2001 and 2003 (Figure 1a).[10] First
investigations revealed that these compounds have potent
in vitro antimicrobial activity against methicillin-resistant
S. aureus and in vitro antimalarial activity against Plasmo-
dium falciparum W2 clones. Despite their important
biological activities and their promising pharmacological
potential, only very few total syntheses of these natural
products have been reported thus far.[11] In 2007, Pan
et al. accomplished the first total synthesis of
(+)-machaeriol D (4) using a regio- and stereoselective
SN2’ reaction as the key step (longest linear sequence:
18 steps).[11b] Very recently, Dethe and co-workers could
reduce the amount of synthetic operations towards 4 to
six steps by applying an elegant strategy that also
provides access to the unnatural enantiomer.[11f] How-
ever, a modular approach allowing for simple and short
syntheses of (+)-machaeriol B (3) and (+)-machaeriol D
(4) starting from a common building block has not been
disclosed to date.
tance of (À)-D9-trans-tetrahydrocannabinol ((À)-D9-THC, 1)
and its isomer (À)-D8-THC (2),[1] the cannabinoid family,
which includes phytocannabinoids[2] and endogenous[3] and
synthetic cannabinoids,[4] has attracted the attention of
scientists worldwide (Figure 1a). Thus far, more than 60 can-
nabinoids and related analogues have been identified, and
Herein, we report a divergent strategy for the short
synthesis of (+)-machaeriol B (3), (+)-machaeriol D (4;
currently the shortest), and related analogues using
a
stereospecific palladium-catalyzed decarboxylative
arylation and a one-pot stereocontrolled hydroboration,
followed by oxidation or reduction of the intermediary
borane, as key steps. Moreover, we will show that this
modular approach can also be applied to the synthesis of
cannabinoid derivatives, such as (+)-D8-THC (ent-2), starting
from the same, readily available (S)-perillic acid.
The antimalarial agents (+)-machaeriol B (3) and
(+)-machaeriol D (4) are structurally related to natural
THC derivatives, such as 1 and 2. Although their absolute
configurations differ, these compounds feature a tricyclic
dibenzopyran motif (ABC ring system) with trans connectiv-
ity between the pyran skeleton (ring B) and the cyclohexane
Figure 1. a) Natural cannabinoids and analogues. b) Retrosynthetic analysis.
[*] F. Klotter, Prof. Dr. A. Studer
Organisch-Chemisches Institut
Westfälische Wilhelms-Universität Münster
Corrensstrasse 40, 48149 Münster (Germany)
E-mail: studer@uni-muenster.de
[**] We thank the Fonds der Chemischen Industrie (stipend for F.K.) and
the Deutsche Forschungsgemeinschaft (DFG) for financial support.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201502595.
Angew. Chem. Int. Ed. 2015, 54, 8547 –8550
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
8547

.
Angewandte
Communications
subunit (ring C) at positions C6a and C10a as a common
structural unit. However, in contrast to THC derivatives,
machaeriols 3 and 4 contain a benzofuran moiety at the
C3 position (instead of a pentyl chain) and have a fully
saturated cyclohexane core bearing an equatorial methyl
group at the C9 position. Compared to the THCs,
(+)-machaeriol D (4) contains an additional hydroxy-bearing
stereogenic center at the C8 position anti to the methyl group,
rendering its synthesis certainly more challenging.
The structural similarity of all of these compounds
motivated us to develop a divergent and short approach
that will not only provide access to the natural machaeriols 3
and 4 or related analogues but can also be used for the
synthesis of cannabinoids, such as (+)-D8-THC (ent-2). As
one of our key steps, we planned to use a stereospecific
palladium-catalyzed decarboxylative g-arylation[12] for the
introduction of various trans-configured aryl groups (in
ring C) starting from chiral (1S,4S)-disubstituted cyclohex-
ene-1-carboxylic acid 5 (Figure 1b). The D9-tetrahydrobi-
phenyls A that are thus obtained should be readily trans-
formed into intermediates of type B by a short sequence that
includes oxycyclization to install the trans pyran core (ring B)
and subsequent double-bond isomerization. D8-Tetrahydro-
dibenzopyrans B represent our key intermediates as three
divergent transformations of these compounds should lead to
all of the targeted compounds. In detail, (+)-machaeriol B
(3), (+)-machaeriol D (4), and related analogues should be
accessible from common intermediate B (R = 2-benzofuryl)
in a one-pot key sequence that is based on a stereocontrolled
hydroboration followed by oxidation (for 4) or reduction (for
3). On the other hand, simple deprotection of B (R = pentyl)
will lead to cannabinoids, such as (+)-D8-THC (ent-2).
Our synthetic studies commenced with the preparation of
chiral (1S,4S)-disubstituted cyclohexene-1-carboxylic acid 5
from commercially available (S)-perillic acid (6; Scheme 1).
Importantly, (R)-perillic acid (ent-6) is also accessible and can
be prepared on large scale, which opens the door to the
synthesis of both enantiomers of all compounds presented
herein.[13] We assumed that the isopropenyl substituent in 6
could be utilized to induce stereoselectivity in the a-methyl-
ation. An additional challenge was the control of the a/g-
regioselectivity in the alkylation step. After intensive exper-
imentation, we found that the dienolate generated from 6
upon treatment with LDA in a THF/DMPU[14] mixture
reacted with dimethyl sulfate to yield carboxylic acid 5 with
excellent yield (90%, gram scale), complete a-selectivity, and
moderate anti diastereoselectivity favoring the desired isomer
(1.7:1; Scheme 1). The proper choice of base, alkylation
reagent, solvent mixture, and reaction temperature turned out
to be crucial for successful methylation. All of our attempts to
increase the diastereoselectivity by using chiral bases or chiral
ligands in combination with sterically demanding organo-
lithium bases were not fruitful. However, as the targeted
isomer was formed as the major product, the alkylation was
the first step in the synthesis, and we assumed that separation
of isomers might not be necessary at this stage, we began to
investigate the key step for our synthesis—the stereoselective
introduction of aryl groups trans to the isopropenyl group.
Recently, we developed a Pd-catalyzed stereospecific
decarboxylative coupling of substituted 1,4-cyclohexadiene-
Scheme 1. Total synthesis of machaeriols and cannabinoid analogues. Reagents and conditions: a) LDA (2.5 equiv), À788C to RT, 2 h; then DMPU
(10 equiv), 15 min, RT; then DMS (1.3 equiv), À788C to RT, 1 h; b) [Pd(dba)2] (10 mol%), Cs2CO3 (1.1 equiv), 7a–c (1.0 equiv), toluene (0.3m), 1108C,
26 h; c) TMSCl (1.5 equiv), NaI (2.0 equiv), 708C, 24 h; then NaI (0.2 equiv), 708C, 24 h; d) NaSEt (10 equiv), DMF, 1408C, 12 h; e) for hydroboration/
reduction: Sia2BH (4 equiv), 608C, 2 d; then 11 (10 equiv), 808C, open atmosphere, 4 h; for hydroboration/oxidation: Sia2BH (2 equiv), 608C, 2 d; then
H2O, NaOH, H2O2, 2 h, RT.
8548
www.angewandte.org
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2015, 54, 8547 –8550

Angewandte
Chemie
3-carboxylic acids with aryl iodides leading to 5-arylated 1,3-
cyclohexadiene derivatives.[12a,b] Decarboxylative g-arylation
was also successfully applied to the total synthesis of
resveratrol-based natural products.[12c] Notably, the carboxylic
imum ratio of 3:1 in favor of the desired stereoisomer,
whereas homogenous (non-chiral and chiral) metal catalysts
provided the undesired isomer as the major product in all
cases. To overcome this stereochemical problem, we tested
a method developed by Renaud and co-workers, which is
3
2
À
acids that were used in these successful C(sp ) C(sp )
couplings result, upon decarboxylation, in allylic Pd inter-
mediates that are further stabilized by conjugation either with
an additional double bond or an arene. The stability of an
anion formed by decarboxylation is known to be an important
issue in Pd-catalyzed decarboxylation reactions for the
generation of Pd allyl complexes. However, an additional
p-stabilizing entity is missing in 5, and therefore, we expected
5 to be a difficult substrate for decarboxylative g-arylation.
Pleasingly, when we applied our standard conditions of
the stereospecific decarboxylative g-arylation[12a] to anti-5
using [Pd(dba)2] (10 mol%), Cs2CO3 (1.1 equiv), and iodide
7a, product 8a was obtained in 65% yield with complete
stereospecificity. Importantly, upon using the 1.7:1 anti/syn
mixture of 5, 8a was obtained in an even slightly better yield
(73%, calculated based on anti-5), showing that 1) syn-5 does
not undergo g-arylation and, more importantly, that 2) sep-
aration of the anti/syn isomers of 5 is not necessary. This is
highly helpful from a practical point of view, in particular for
larger-scale preparations of 8a. Unreacted acid syn-5 can be
reisolated after work-up. Therefore, all following experiments
were conducted with the isomeric mixture of 5. With aryl
iodides 7b and 7c (1.0 equiv), the (3S,4S)-trans-D9-tetrahy-
drobiphenyls 8b and 8c were obtained in good yields as single
diastereoisomers (74% and 81% based on anti-5).
based on a one-pot hydroboration and subsequent radical
[19]
À
C B reduction with 4-tert-butylcatechol (11). This elegant
strategy was used by these authors as the key step for an
efficient synthesis of all-cis-trimethyldecalin derivatives.[19c]
Pleasingly, hydroboration of 9a and 9c with disiamylborane
(Sia2BH) and subsequent radical reduction with 11 provided
hexahydrodibenzopyran derivatives 12a and 12c in good
yields and very good diastereoselectivities (d.r. 17:1 for 12a
and 19:1 for 12c). Products 12a and 12c were used in the final
deprotection step without further purification.[18] Removal of
the methyl group afforded (++)-machaeriol B (3) and its
analogue 13a in good yield (39%, d.r. 22:1 for 3 and 43%, d.r.
19:1 for 13a over 3 steps). With Sia2BH as a selective
hydroboration agent in hand, we focused on the completion
of the synthesis of (+)-machaeriol D (4). Hydroboration of
9a and 9c and oxidative work-up (H2O2, NaOH) provided the
secondary alcohols 14a and 14c as single diastereoisomers
(d.r. > 99:1) in good yields.[20,21] Final deprotection of 14a and
14c using NaSEt in DMFat reflux afforded (+)-machaeriol D
(4) and the related analogue 15a (d.r. > 99:1) in very good
yields (82% for 4 and 83% for 15a; 115 mg of 4 prepared in
one sequence).
In summary, we have reported a novel, short, and
divergent approach for the stereoselective synthesis of
machaeriols and cannabinoid-related compounds from
a common precursor. Various aryl groups could be readily
installed through stereospecific palladium-catalyzed decar-
boxylative couplings. Compounds of type B were identified as
key structural intermediates that provide access to a variety of
machaeriols, cannabinoids, and related compounds through
efficient follow-up transformations. Therefore, the antima-
larial agents (+)-machaeriol B (3) and (+)-machaeriol D (4)
were synthesized in just five steps (thus far the shortest route
towards 4) in 18% and 19% overall yield (0.12 g of 4
prepared). Furthermore, (+)-D8-THC was obtained in four
steps and 27% overall yield (0.18 g prepared). Importantly,
our modular strategy will enable the large-scale synthesis of
further machaeriol and cannabinoid analogues.
We next studied the construction of the trans-configured
pyran core (ring B) starting from 8a–c through selective
deprotection of one methyl ether using trimethylsilyl iodide
(TMSI) and subsequent oxycyclization in a one-pot fashion,
according to a method of Childers and Pinnick.[15] Unfortu-
nately, upon treatment of 8a with TMSI, a complex mixture
containing various iodinated species was obtained. The
quality of the TMSI sample used turned out to be important.
As the handling of TMSI is problematic, we envisioned to
generate this reagent in situ from trimethylsilyl chloride
(TMSCl) and sodium iodide (NaI).[16] After a thorough
screen, we found conditions that allowed for the reproducible
generation of the trans-configured pyran skeleton on larger
scale (see Scheme 1). Moreover, under these conditions, D9 to
D8 double-bond isomerization also occurred, likely for
thermodynamic reasons.[17] The D8-tetrahydrodibenzopyran
derivatives 9a–c thus obtained were used for the next steps
without any further purification.[18]
With this highly efficient three-step synthesis to inter-
mediates containing the requisite 6,6,6-tricyclic system of
machaeriols and cannabinoids in hand, we then investigated
the final divergent transformations. Whereas the deprotection
of the remaining methoxy group in 9a and 9b to form the
synthetic cannabinoids (+)-D8-THC (ent-2, 64% over 2 steps)
and 10a (58% over 2 steps) was readily achieved with NaSEt
in DMF at elevated temperatures, first attempts towards the
synthesis of (+)-machaeriol B (3) through a stereocontrolled
reduction of the D8-configured double bond in 9a and 9c were
not rewarding. In particular, hydrogenation of 9a with
heterogeneous catalysts such as Pd/C proceeded with a max-
Keywords: cannabinoids · decarboxylative couplings ·
divergent synthesis · natural products · total synthesis
How to cite: Angew. Chem. Int. Ed. 2015, 54, 8547–8550
Angew. Chem. 2015, 127, 8667–8670
[1] a) Y. Gaoni, R. Mechoulam, J. Am. Chem. Soc. 1964, 86, 1646 –
1647; b) E. C. Taylor, K. Lenard, Y. Shvo, J. Am. Chem. Soc.
1966, 88, 367 – 369; c) R. S. Wilson, E. L. May, J. Med. Chem.
1975, 18, 700 – 703; d) S. Valiveti, D. C. Hammell, D. C. Earles,
A. L. Stinchcomb, J. Pharm. Sci. 2004, 93, 1154 – 1164.
[2] a) A. T. El-Alfy, K. Ivey, K. Robinson, S. Ahmed, M. Radwan, D.
Slade, I. Khan, M. ElSohly, S. Ross, Pharmacol. Biochem. Behav.
2010, 95, 434 – 442; b) R. G. dos Santos, J. E. C. Hallak, J. P.
Leite, A. W. Zuardi, J. A. S. Crippa, J. Clin. Pharm. Ther. 2014,
Angew. Chem. Int. Ed. 2015, 54, 8547 –8550
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 8549

.
Angewandte
Communications
40, 135 – 143; c) R. Mechoulam, L. O. Hanus, R. Pertwee, A. C.
Howlett, Nat. Rev. Neurosci. 2014, 15, 757 – 764.
Koch, A. Studer, Angew. Chem. Int. Ed. 2013, 52, 4933 – 4936;
Angew. Chem. 2013, 125, 5033 – 5036; c) F. Klotter, A. Studer,
Angew. Chem. Int. Ed. 2014, 53, 2473 – 2476; Angew. Chem.
2014, 126, 2505 – 2509.
[3] a) R. I. Wilson, R. A. Nicoll, Nature 2001, 410, 588 – 592; b) D.
Cota, G. Marsicano, B. Lutz, V. Vicennati, G. K. Stalla, R.
Pasquali, U. Pagotto, Int. J. Obes. Relat. Metab. Disord. 2003, 27,
289 – 301; c) B. M. Fonseca, M. A. Costa, M. Almada, G.
Correia-da-Silva, N. A. Teixeira, Prostaglandins & Other Lipid
Mediators 2013, 102 – 103, 13 – 30; Other Lipid Mediators 2013,
102 – 103, 13 – 30.
[4] a) V. D. Marzo, L. D. Petrocellis, Annu. Rev. Med. 2006, 57, 553 –
574; b) N. Uchiyama, R. Kikura-Hanajiri, J. Ogata, Y. Goda,
Forensic Sci. Int. 2010, 198, 31 – 38; c) E. E. Bonar, L. Ashra-
fioun, M. A. Ilgen, Drug Alcohol Depend. 2014, 143, 268 – 271.
[5] a) M. Ben Amar, J. Ethnopharmacol. 2006, 105, 1 – 25; b) M. A.
Huestis, Chem. Biodiversity 2007, 4, 1770 – 1804; c) E. B. Russo,
Ther. Clin. Risk Manage. 2008, 4, 245 – 259.
[6] B. M. Trost, K. Dogra, Org. Lett. 2007, 9, 861 – 863.
[7] L.-J. Cheng, J.-H. Xie, Y. Chen, L.-X. Wang, Q.-L. Zhou, Org.
Lett. 2013, 15, 764 – 767.
[8] F. Gläser, M. C. Brçhmer, T. Hurrle, M. Nieger, S. Bräse, Eur. J.
Org. Chem. 2015, 1516 – 1524.
[9] M. A. Schafroth, G. Zuccarello, S. Krautwald, D. Sarlah, E. M.
Carreira, Angew. Chem. Int. Ed. 2014, 53, 13898 – 13901; Angew.
Chem. 2014, 126, 14118 – 14121.
[10] a) I. Muhammad, X.-C. Li, D. C. Dunbar, M. A. ElSohly, I. A.
Khan, J. Nat. Prod. 2001, 64, 1322 – 1325; b) I. Muhammad, X.-C.
Li, M. R. Jacob, B. L. Tekwani, D. C. Dunbar, D. Ferreira, J. Nat.
Prod. 2003, 66, 804 – 809.
[13] D. J. Kerr, M. Miletic, N. Manchala, J. M. White, B. L. Flynn,
Org. Lett. 2013, 15, 4118 – 4121.
[14] T. Mukhopadhyay, D. Seebach, Helv. Chim. Acta 1982, 65, 385 –
391.
[15] W. E. Childers, H. W. Pinnick, J. Org. Chem. 1984, 49, 5276 –
5277.
[16] G. A. Olah, S. C. Narang, B. G. B. Gupta, R. Malhotra, J. Org.
Chem. 1979, 44, 1247 – 1251.
[17] We assume that the isomerization is thermodynamically favored.
Notably, double-bond isomerization occurs before cyclization
takes place, as determined by GC-MS analysis.
[18] After purification by column chromatography, target com-
pounds contained tiny amounts of impurities. These side
products could be removed during the work-up of the next step.
[19] a) G. Povie, G. Villa, L. Ford, D. Pozzi, C. H. Schiesser, P.
Renaud, Chem. Commun. 2010, 46, 803 – 805; b) G. Villa, G.
Povie, P. Renaud, J. Am. Chem. Soc. 2011, 133, 5913 – 5920; c) G.
Villa, B. Bradshaw, C. Bürki, J. Bonjoch, P. Renaud, Tetrahedron
Lett. 2014, 55, 4608 – 4611.
[20] Although the hydroboration was not completely selective
(> 17:1), we could not identify the minor alcohol isomer after
oxidation. Presumably, the minor alkylborane diastereomer is
either very slowly or not oxidized under the reaction conditions.
[21] Notably,
a
similar hydroboration/oxidation sequence of
[11] a) A. G. Chittiboyina, C. R. Reddy, E. B. Watkins, M. A. Avery,
Tetrahedron Lett. 2004, 45, 1689 – 1691; b) Q. Wang, Q. Huang,
B. Chen, J. Lu, H. Wang, X. She, X. Pan, Angew. Chem. Int. Ed.
2006, 45, 3651 – 3653; Angew. Chem. 2006, 118, 3733 – 3735; c) Q.
Huang, Q. Wang, J. Zheng, J. Zhang, X. Pan, X. She, Tetrahedron
2007, 63, 1014 – 1021; d) L. Xia, Y. R. Lee, Synlett 2008, 1643 –
1646; e) H. J. Lee, Y. R. Lee, S. H. Kim, Helv. Chim. Acta 2009,
92, 1404 – 1412; f) D. H. Dethe, R. D. Erande, S. Mahapatra, S.
Das, V. Kumar B. , Chem. Commun. 2015, 51, 2871 – 2873.
[12] a) C.-M. Chou, I. Chatterjee, A. Studer, Angew. Chem. Int. Ed.
2011, 50, 8614 – 8617; Angew. Chem. 2011, 123, 8773 – 8776; b) E.
methoxy-protected (À)-D8-THC (2) using BH3 was conducted
by Huffman et al. However, the corresponding alcohol was
obtained non-selectively as a nearly 1:1 mixture of diastereoiso-
mers; see: J. W. Huffman, W. Kenneth Banner, G. K. Zoorob, H.
Howard Joyner, P. H. Reggio, B. R. Martin, D. R. Compton,
Tetrahedron 1995, 51, 1017 – 1032.
Received: March 20, 2015
Published online: June 16, 2015
8550
www.angewandte.org
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2015, 54, 8547 –8550