Biomimetic Cannabinoid Synthesis Revisited: Batch and Flow All-Catalytic Synthesis of (±)-ortho-Tetrahydrocannabinols and Analogues from Natural Feedstocks

Article abstract of DOI:10.1002/ejoc.201800064

Using a combination of Au nanoparticle-catalyzed oxidation under an O₂ atmosphere and a Ti-doped montmorillonite (Ti-MMT)-catalyzed tandem arylation/double cyclization, we developed an original and highly selective method for the synthesis of ortho-tetrahydrocannabinol derivatives from simple substrates. The reaction sequence could be performed in two steps in batch mode or in a single operation in continuous-flow reactors. The abnormal regioselectivity was proposed to be the result of the non-innocent role of the MMT support, Ti^IV cation coordination, and a Lewis acid-assisted Br?nsted acid (LBA) mechanism.

Full text of DOI:10.1002/ejoc.201800064

DOI: 10.1002/ejoc.201800064
Communication
Abnormal Cannabinoids
Biomimetic Cannabinoid Synthesis Revisited: Batch and Flow
All-Catalytic Synthesis of ( )-ortho-Tetrahydrocannabinols and
Analogues from Natural Feedstocks
Pascal D. Giorgi,[a] Virginie Liautard,[b] Mathieu Pucheault,[b] and Sylvain Antoniotti*[a]
Abstract: Using a combination of Au nanoparticle-catalyzed in two steps in batch mode or in a single operation in continu-
oxidation under an O2 atmosphere and a Ti-doped mont- ous-flow reactors. The abnormal regioselectivity was proposed
morillonite (Ti-MMT)-catalyzed tandem arylation/double cycli- to be the result of the non-innocent role of the MMT support,
zation, we developed an original and highly selective method TiIV cation coordination, and a Lewis acid-assisted Brønsted acid
for the synthesis of ortho-tetrahydrocannabinol derivatives from (LBA) mechanism.
simple substrates. The reaction sequence could be performed
Introduction
oxidized and cyclized by tetrahydrocannabinolic acid (THCA)
synthase, an oxidoreductase, to form tetrahydrocannabinolic
In the most recent studies of natural cannabinoid biosynthesis acid, which upon decarboxylation leads to the formation of tet-
in Cannabis sativa, olivetolic acid, arising from a polyketide rahydrocannabinol (THC) (Scheme 1).[4]
pathway by the action of an olivetolic acid cyclase (OAC),[1] is
The most common natural product with this structure
combined with geranyl diphosphate (from mevalonate or meth- features a double bond connecting carbon atoms 8 and 9 or
ylerythritol)[2] to yield cannabigerolic acid.[3] The latter is then carbon atoms 9 and 10 (following the dibenzopyran numbering
Scheme 1. Biosynthesis of tetrahydrocannabinol. GOT: geranyl-pyrophosphate-olivetolic acid geranyltransferase; THCAS: tetrahydrocannabinolic acid synthase.
[a] Université Cote d'Azur, Institut de Chimie de Nice, CNRS,
system) and two vicinal asymmetric carbon atoms in either a
trans or cis relative configuration. From a historical perspective,
the first reported synthesis of THCs followed a biomimetic
route, using olivetol and citral (mixture of neral, Z, and geranial,
E, isomers) as substrates, in which citral offered the required
oxidation state. A series of papers described the use of BF3·OEt2,
Parc Valrose, 06108 Nice cedex 2, France
E-mail: sylvain.antoniotti@unice.fr
www.univ-cotedazur.fr
[b] Institut des Sciences Moléculaires, UMR 5255 CNRS - Université de Bordeaux,
351 cours de la libération, 33405 Talence cedex, France
Supporting information and ORCID(s) from the author(s) for this article are
available on the WWW under https://doi.org/10.1002/ejoc.201800064.
Eur. J. Org. Chem. 2018, 1307–1311
1307
© 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
Scheme 2. Synthetic plan for a direct multicatalytic approach to THC derivatives.
mineral acid,[5] and Mo complexes[6] to assemble these two sub- ing on the second step. The acids screened included Lewis acids
strates or the use of pyridine to shift the reactivity towards the such as ZnBr2, Ti(OiPr)4, and TiCl4; solid catalysts such as Amber-
formation of isomeric cannabichromene (upon hydroalkoxyla- lyst and zeolite; and montmorillonite-doped with metal cations
tion of the C2–C3 double bond of the terpenyl side chain).[7] (M-MMT), a type of solid catalyst that we recently used as a
This low-yielding and moderately selective route was aban- Lewis acid in C–O and C–C bond-forming processes.[15] The
doned in favor of a multistep synthesis involving terpenoids most interesting results are summarized in Table 1 (see Sup-
such as (–)-verbenol, (+)-chrysanthenol, para-mentha-2,8-dien- porting Information for additional examples). Indeed, conven-
1-ol, and (+)-carene oxide, which could yield enantioenriched tional Lewis acids delivered mixtures containing Δ9- and
products.[8] Recently, iterative strategies were devised to con- Δ8-1a, generally in low yields (Table 1, entries 1–4), and para-
struct the target molecule by using Diels–Alder cycloaddition cymene upon intramolecular carbonyl-ene reaction of citral
as the key step[9] or in an enantioselective fashion by stereodi- followed by dehydration. However, the use of M-MMT enabled
vergent allylation of aliphatic aldehyde[10] or stereospecific de- the relatively selective formation of ortho-1a, which is usually
carboxylative arylation.[10] Interestingly, regioisomers featuring referred to as an “abnormal” derivative in the chemistry of natu-
inversion on the benzene ring between the alkyl chain and the ral cannabinoids.
hydroxy group, that is, ortho-THCs, were sometimes described
and were considered to be minor secondary products of syn-
thetic reactions.[11] To our best knowledge, there are no reports
in the literature describing selective access to these ortho prod-
ucts. At the same time, growing interest in natural and synthetic
cannabinoids has been observed owing to the various promis-
ing biological activities and the need for molecules that are
selective for cannabinoid receptor 2 (CB2) over CB1.[12]
Under the optimized conditions with the use of Ti-MMT as
the catalyst (10 mol-%), ortho-1a was obtained in up to 98 %
yield as a 83:17 mixture of the Δ9 and Δ8 isomers (Table 1,
entry 13). The cis/trans ratio was found to be 2:8 in most cases.
This result was rather unexpected, as these compounds were
typically observed as side products in various syntheses of nat-
ural cannabinoids. Under the same reaction conditions but with
the use of olivetol instead of orcinol, Δ9-ortho-1b (Δ9-ortho-
We recently became interested in developing multicatalytic THC) was obtained in 77 % yield (Scheme 3).
reactions triggered by the catalytic oxidation of activated alco-
hols by Au nanoparticles (NPs).[13] In particular, we devised a
bicatalytic one-pot synthesis of substituted chromenes by the
assembly of nerol and salicylaldehyde in the presence of K2CO3
following an oxidation/oxa-Michael addition/intramolecular
aldolization/crotonization sequence.[14] We reasoned that by
Scheme 3. Formation of the ortho analogue of pharmacologically relevant
Δ9-THC.
switching to acid catalysis in the second step instead of base
catalysis we could catalyze an intramolecular carbonyl-ene reac-
tion, which, after protonation/dehydration, could potentially
lead to an allylic cation that acts as an electrophile and could
be further combined with 5-alkylresorcinols such as olivetol to
yield THCs (Scheme 2).
We next focused on the cascade of reactions leading to the
formation of the Δ9- and Δ8-ortho-products. Given that the
formation of small amounts of para-cymene was sometimes
observed during catalyst screening, we reasoned that the se-
quence could start with the monomolecular carbonyl-ene reac-
tion of citral leading to isopiperitenol isomers. The latter could
lead to para-cymene by dehydration and isomerization of the
Results and Discussion
We screened a series of Lewis and Brønsted acids in a test reac- isoprenyl double bond to form a more stable aromatic struc-
tion with citral and orcinol (5-methylresorcinol), thereby focus- ture. However, several attempts to observe (by GC–MS) or to
Eur. J. Org. Chem. 2018, 1307–1311
www.eurjoc.org
1308
© 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
Table 1. Screening of acid catalysts for the reaction of citral (E/Z = 1:1) with orcinol.[a]
Entry Catalyst[b]
Time [h]
CV[c] [%]
Yield[c] [%]
Δ9-ortho-1a Δ9-ortho-2a Δ8-ortho-1a 3a/ortho-3a
Δ9-1a
Δ9-2a
Δ8-1a
1
2
3
4
5
6
ZnBr2 (1 equiv.)
8
8
2
2
8
2
0
0
0
0
0
0
0
0
0
0
38
36
0
0
0
0
0
13
0
0
0
0
21
9
0
Ti(OiPr)4 (15 mol-%)
BF3·OEt2 (10 mol-%)
Bi(OTf)3 (10 mol-%)
TiO2 (10 mol-%)[e]
TiCl4 (10 mol-%)
86[d]
100
10
38
18
17
0
0
0
0
5
0
31:10
12:0
0
32:5
0
100
100[d]
100[d]
0
0
Na-MMT (100 mg)
zeolite (75 mg)
Li-MMT (100 mg)
TiCl4 (10 mol-%)
7
8
9
10
8
24
8
5
11
0
0
0
0
0
0
0
0
0
0
0
0
0
5
15
8
0
2
0
0
3
0
6
19
2:0
2:0
11:3
0
100[d,f]
100[d]
TiCl4 (10 mol-%)
2
Li-MMT (100 mg)
Ge-MMT (10 mol-%)
Sn-MMT (10 mol-%)
Ti-MMT (10 mol-%)
11
12
13
2
72
2
100[d]
58
100
0
0
0
0
0
0
0
0
0
6
0
10
n.d.
34
19
17
0
5:5
0
19
81[g]
[a] Reaction conditions: citral (1 mmol), orcinol (1 mmol), toluene (2 mL), 80 °C, air, 180 min. [b] For metal-based catalysts, catalyst ratio is related to the metal.
[c] Conversion (CV) and selectivity were determined by 1H NMR spectroscopy analysis of the crude product isolated after workup; both cis and trans
diastereomers were formed for 1a and 2a, typically in a 2:8 ratio; n.d.: not determined. [d] Paracymene was formed upon the intramolecular carbonyl-ene
reaction of citral and dehydration in the reaction medium. [e] Peptized solution. [f] Degradation occurred. [g] Mixture of trans and cis diastereomers in a
78:22 ratio (see the Supporting Information for full NMR spectroscopy experiments, including the NOE spectrum).
trap the isopiperitenol intermediate were unsuccessful. Consid-
ering that the conversion of isopiperitenol into para-cymene
could be too fast, we turned our attention to (R)-citronellal
in lieu of citral to prevent conversion into stable benzene
structures by isomerization. Indeed, in the presence of Ti-MMT,
citronellal was converted into a carbonyl-ene product of citron-
ellal, isopulegol, as the major product. Upon performing the
reaction with citronellal and orcinol, under our optimized con-
ditions with Ti-MMT, the reaction followed the expected path
and delivered ortho-hexahydrocannabiorcol ortho-dihydro-1a
in 41 % yield together with a small amount of corresponding
alkylated orcinol ortho-dihydro-2a from the uncomplete reac-
tion (Scheme 4, top). However, with isopulegol as the starting
material and in the presence of orcinol, a different outcome
was observed after 2 h with quantitative formation of product
4a, a class of products not previously observed during our in-
vestigations (Scheme 4, middle). This product would be formed
Scheme 4. Mechanistic investigations with (R)-citronellal and isopulegol.
by a retro-carbonyl-ene process followed by the formation of
the 4-hydroxytetrahydropyrane core by reaction with iso-
pulegol.
This set of results suggested that an initial carbonyl-ene reac-
tion was not the most critical event in the formation of ortho-
submitted to the reaction with Ti(OiPr)4 in THF at 60 °C. In this
case, styrenyl derivative 5a was obtained in 51 % yield as the
sole product after 2 h (Scheme 4, bottom).
THC derivatives. To investigate this idea further and to substi-
This result suggested an alternative reaction pathway, by
tute Ti-MMT with a less-efficient catalytic system, citronellal was which the starting event would be the arylation of the aldehyde
Eur. J. Org. Chem. 2018, 1307–1311
www.eurjoc.org
1309
© 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
Scheme 5. Putative template effect of Ti-MMT, LBA mode of activation, and subsequent regioselectivity of nucleophilic attack of the arene towards citral in
an aldol-Friedel–Crafts-type reaction.
Figure 1. Continuous-flow chemistry setup for the synthesis of ortho-THCs. TBHP: tert-butyl hydroperoxide, BPR: back-pressure regulator.
function by orcinol, followed by dehydration. Starting from ci- Lewis acid assisted Brønsted acid process allowed by Ti species.
tral, this key intermediate could explain the reactivity observed Coordination of the OH groups to TiIV would indeed result in
during screening experiments, including the formation of the acidification of the bound hydrogen atom, which would be fur-
corresponding quinone methide intermediate, which is some- ther transferred to the carbonyl group of citral. Nucleophilic
times proposed as an intermediate in the synthesis of THC and attack of the arene would thus proceed through the most ac-
its analogues by 6π electrocyclization.[16]
cessible position, that is, the ortho position relative to the alkyl
substituent (Scheme 5).
Several attempts to trap a putative quinone methide inter-
mediate by conjugate addition of nucleophiles (nPrSH,
BnCH2SH, diethyl malonate, allyltrimethylsilane) or by blocking
the phenolic group (CH3COCl, TMSCl, vinyl acetate) were unsuc-
cessful. Attempts to prepare deuterated orcinol in D2O and in
MeOD resulted in the formation of isotopomer mixtures as a
result of a high degree of scrambling (mostly on the OH and
CH3 groups).
A series of experiments were then undertaken to gain a bet-
ter understanding of the regioselectivity of the reaction. First,
the role of the hydroxy groups of orcinol was evaluated. Citral
was thus treated with modified alkyl resorcinols in the presence
of Ti-MMT (10 mol-%) under our optimized conditions. Interest-
ingly, blockage of both hydroxy groups by methylation or acet-
ylation resulted in complete inhibition of the arylation reaction
and recovery of unchanged citral. Surprisingly, blockage of a
single hydroxy group also completely inhibited the reaction.
Given our interest in the design of multicatalytic chemical
processes,[14] we performed a preliminary series of tests to com-
bine the Au NP-catalyzed oxidation of allylic alcohols, such as
nerol and geraniol,[13] with the Ti-MMT-catalyzed cyclization de-
scribed herein in the same reactor. Unfortunately, the catalysts
were not compatible under our conditions, and the oxidation
step was quenched in the presence of Ti-MMT. We thus moved
towards continuous-flow chemistry reactors.
The first stage of the batch→flow transposition was optimi-
zation of the Ti-MMT-catalyzed step in flow. At 0.05
M and on a
4 mmol scale, a citral conversion of 100 % and 98 % yield of
the cyclized products (Δ9-ortho-1a, 17 %; Δ8-ortho-1a, 72 %;
and Δ9-ortho-2a, 9 %) were obtained with a column containing
400 mg of Ti-MMT within a 5 min residence time. With the
implementation of a second catalytic column filled with Au
NPs/Al2O3 (1 g) and using the tube-in-tube technology for effi-
On the basis of these observations and taking the regiose- cient O2 supply, the flow synthesis of ortho-THC and analogues
lectivity in favor of the ortho isomers of THC into consideration, was possible from nerol (2.5 mmol scale) and enabled total con-
we reasoned that Ti-MMT could influence the selectivity version and the formation of ortho-1a (Δ8/Δ9 = 1:3.5) and or-
through a template effect, by which both hydroxy groups tho-1b (Δ8/Δ9 = 1:6) in yields of 81 and 72 % from orcinol and
would coordinate to the interlamellar surface, combined with a olivetol, respectively (Figure 1).
Eur. J. Org. Chem. 2018, 1307–1311
www.eurjoc.org
1310
© 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
[4] Y. Shoyama, A. Takeuchi, F. Taura, T. Tamada, M. Adachi, R. Kuroki, Y.
Shoyama, S. Morimoto, Acta Crystallogr., Sect. F 2005, 61, 799–801.
[5] E. C. Taylor, K. Lenard, Y. Shvo, J. Am. Chem. Soc. 1966, 88, 367–369.
[6] A. V. Malkov, P. Kočovský, Collect. Czech. Chem. Commun. 2001, 66, 1257–
1268.
[7] L. Crombie, R. Ponsford, Tetrahedron Lett. 1968, 9, 4557–4560.
[8] a) R. K. Razdan in The Total Synthesis of Natural Products (Ed.: J. ApSimon),
John Wiley & Sons, Inc., Hoboken, NJ, 2007, vol. 4, pp. 185–262; b) T.
Petrzilka, W. Haefliger, C. Sikemeier, G. Ohloff, A. Eschenmoser, Helv.
Chim. Acta 1967, 50, 719–723; c) R. Mechoulam, P. Braun, Y. Gaoni, J. Am.
Chem. Soc. 1967, 89, 4552–4554.
[9] a) D. C. Kapeller, S. Bräse, Synlett 2011, 2011, 161–164; b) F. Gläser, M. C.
Bröhmer, T. Hurrle, M. Nieger, S. Bräse, Eur. J. Org. Chem. 2015, 1516–
1524.
[10] F. Klotter, A. Studer, Angew. Chem. Int. Ed. 2015, 54, 8547–8550; Angew.
Chem. 2015, 127, 8667–8670.
The change in Δ8/Δ9 selectivity was attributed to a chromat-
ographic effect, as Ti-MMT was charged with only 250 mg in
this case. The Δ9/Δ8 isomerization, which is known to be a
thermal phenomenon, was observed upon using an 8:2 mixture
of Δ9-ortho-1a/Δ8-ortho-1a under our reaction conditions over
2 h, leading to a 6:4 mixture.
Conclusions
In conclusion, we found an efficient, selective, and straightfor-
ward method to form ortho-THC derivatives by combining two
catalytic steps. In the future, these structures could serve in the
design of novel bioactive molecules for cannabinoid research.
Current pharmacological interest in this field focuses on drug
candidates with a weak affinity for CB1 and a high affinity for
CB2. The former is indeed believed to be involved in the psy-
choactive characteristics of THC derivatives, whereas the latter
is associated with analgesic properties. The phenol hydroxy
group was long ago identified as a mandatory structural re-
quirement for the biological activity of THC.[17] Recent struc-
ture–activity studies and crystallographic analysis of Δ9-1a
showed the importance of the phenol group in hydrogen bond-
ing with a serine residue of the CB1 pocket.[18] However, addi-
tional studies on CB2 also showed that the absence of this
phenol hydroxy group improved CB2/CB1 selectivity.[19] With
the efficient production of ortho-THC derivatives presented
herein, a new subclass of drug candidates with high CB2 selec-
tivity could be envisaged. Moreover, given the current renewed
interest in natural cannabinoids, this method for the production
of original cannabinoid structures will be useful in the increas-
ing number of pharmacological studies for a broad range of
diseases.[12c,12d,12f]
[11] a) L. Crombie, W. M. L. Crombie, Phytochemistry 1975, 14, 213–220; b)
D. B. Uliss, G. R. Handrick, H. C. Dalzell, R. K. Razdan, Tetrahedron 1978,
34, 1885–1888.
[12] a) F. Astruc-Diaz, S. W. McDaniel, J. J. Xu, S. Parola, D. L. Brown, M. Naguib,
P. Diaz, J. Pharm. Sci. 2013, 102, 352–364; b) R. R. Petrov, L. Knight, S.-R.
Chen, J. Wager-Miller, S. W. McDaniel, F. Diaz, F. Barth, H.-L. Pan, K.
Mackie, C. N. Cavasotto, P. Diaz, Eur. J. Med. Chem. 2013, 69, 881–907; c)
C. Bladen, S. W. McDaniel, V. M. Gadotti, R. R. Petrov, N. D. Berger, P.
Diaz, G. W. Zamponi, ACS Chem. Neurosci. 2015, 6, 277–287; d) E. Hebert-
Chatelain, T. Desprez, R. Serrat, L. Bellocchio, E. Soria-Gomez, A. Bus-
quets-Garcia, A. C. Pagano Zottola, A. Delamarre, A. Cannich, P. Vincent,
M. Varilh, L. M. Robin, G. Terral, M. D. García-Fernández, M. Colavita, W.
Mazier, F. Drago, N. Puente, L. Reguero, I. Elezgarai, J.-W. Dupuy, D. Cota,
M.-L. Lopez-Rodriguez, G. Barreda-Gómez, F. Massa, P. Grandes, G. Bén-
ard, G. Marsicano, Nature 2016, 539, 555–559; e) A. Bilkei-Gorzo, O. Al-
bayram, A. Draffehn, K. Michel, A. Piyanova, H. Oppenheimer, M. Dvir-
Ginzberg, I. Racz, T. Ulas, S. Imbeault, I. Bab, J. L. Schultze, A. Zimmer,
Nat. Med. 2017, 23, 782–787; f) J. S. Mounessa, J. A. Siegel, C. A. Dunnick,
R. P. Dellavalle, J. Am. Acad. Dermatol. 2017, 77, 188–190; g) M.
Soethoudt, U. Grether, J. Fingerle, T. W. Grim, F. Fezza, L. de Petrocellis,
C. Ullmer, B. Rothenhäusler, C. Perret, N. van Gils, D. Finlay, C. MacDonald,
A. Chicca, M. D. Gens, J. Stuart, H. de Vries, N. Mastrangelo, L. Xia, G.
Alachouzos, M. P. Baggelaar, A. Martella, E. D. Mock, H. Deng, L. H. Heit-
man, M. Connor, V. Di Marzo, J. Gertsch, A. H. Lichtman, M. Maccarrone,
P. Pacher, M. Glass, M. van der Stelt, Nat. Commun. 2017, 8, 13958.
[13] P. D. Giorgi, N. Elizarov, S. Antoniotti, ChemCatChem 2017, 9, 1830–1836.
[14] P. D. Giorgi, P. J. Miedziak, J. K. Edwards, G. J. Hutchings, S. Antoniotti,
ChemCatChem 2017, 9, 70–75.
Acknowledgments
[15] a) I. Notar Francesco, B. Cacciuttolo, O. Pascu, C. Aymonier, M. Pucheault,
S. Antoniotti, RSC Adv. 2016, 6, 19807–19818; b) N. Elizarov, M. Pucheault,
S. Antoniotti, ChemistrySelect 2016, 1, 3219–3222; c) I. Notar Francesco,
B. Cacciuttolo, M. Pucheault, S. Antoniotti, Green Chem. 2015, 17, 837–
841; d) K. Ben Hadj Hassen, K. Gaubert, M. Vaultier, M. Pucheault, S.
Antoniotti, Green Chem. Lett. Rev. 2014, 7, 243–249.
[16] a) A. Garcia, D. Borchardt, C.-E. A. Chang, M. J. Marsella, J. Am. Chem. Soc.
2009, 131, 16640–16641; b) E. J. Jung, B. H. Park, Y. R. Lee, Green Chem.
2010, 12, 2003–2011.
This work was supported by the University Nice Sophia Antipo-
lis and the Centre National de la Recherche Scientifique (CNRS).
The French Ministère de l'enseignement supérieur et de la re-
cherche (MESR) is acknowledged for a Ph.D. fellowship to
P. D. G. We are grateful to Dr. Abby Cuttriss for proofreading
this manuscript and to Mr. Lucas Jelensperger for his technical
assistance in the purification of Δ9-1b.
[17] D. B. Uliss, H. C. Dalzell, G. R. Handrick, J. F. Howes, R. K. Razdan, J. Med.
Chem. 1975, 18, 213–215.
Keywords: Cannabinoids · Supported catalysts · Flow
chemistry · Gold · Nanoparticles
[18] T. Hua, K. Vemuri, S. P. Nikas, R. B. Laprairie, Y. Wu, L. Qu, M. Pu, A. Korde,
S. Jiang, J.-H. Ho, G. W. Han, K. Ding, X. Li, H. Liu, M. A. Hanson, S. Zhao,
L. M. Bohn, A. Makriyannis, R. C. Stevens, Z.-J. Liu, Nature 2017, 547, 468–
471.
[19] a) A. Poso, J. W. Huffman, Brit. J. Pharmacol. 2008, 153, 335–346; b) Y.
Pei, R. W. Mercier, J. K. Anday, G. A. Thakur, A. M. Zvonok, D. Hurst, P. H.
Reggio, D. R. Janero, A. Makriyannis, Chem. Biol. 2008, 15, 1207–1219.
[1] J. M. Stout, Z. Boubakir, S. J. Ambrose, R. W. Purves, J. E. Page, Plant J.
2012, 71, 353–365.
[2] M. Rohmer, Nat. Prod. Rep. 1999, 16, 565–574.
[3] S. J. Gagne, J. M. Stout, E. Liu, Z. Boubakir, S. M. Clark, J. E. Page, Prog.
Natl. Acad. Sci. USA 2012, 109, 12811–12816.
Received: January 15, 2018
Eur. J. Org. Chem. 2018, 1307–1311
www.eurjoc.org
1311
© 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim