One-Pot Total Synthesis of Cannabinol via Iodine-Mediated Deconstructive Annulation

Article abstract of DOI:10.1021/acs.orglett.9b02258

The thermal degradation of cannabichromene (CBC, 3) is dominated by cationic reactions and not by the pericyclic rearrangements observed in model compounds. The rationalization of these differences inspired the development of a process that coupled, in an aromatization-driven single operational step, the condensation of citral and alkylresorciniols to homoprenylchromenes and their in situ deconstructive annulation to benzo[c]chromenes. This process was applied to a total synthesis of cannabinol (CBN, 5) and to its molecular editing.

Full text of DOI:10.1021/acs.orglett.9b02258

Letter
Cite This: Org. Lett. XXXX, XXX, XXX−XXX
pubs.acs.org/OrgLett
One-Pot Total Synthesis of Cannabinol via Iodine-Mediated
Deconstructive Annulation
†
†
†
†
‡
Diego Caprioglio, Daiana Mattoteia, Alberto Minassi, Federica Pollastro, Annalisa Lopatriello,
§
,‡
,†
Eduardo Munoz, Orazio Taglialatela-Scafati,* and Giovanni Appendino*
̌
†
‡
§
Dipartimento di Scienze del Farmaco, Universita
Dipartimento di Farmacia, Universita di Napoli Federico II, Via Montesano 49, 80131 Napoli, Italy
Maimonides Biomedical Research Institute of Cordoba, University of Cordoba, Avda Menendez Pidal s/n, 14004 Cor
S Supporting Information
̀
del Piemonte Orientale, Largo Donegani 2, 28100 Novara, Italy
̀
́
́
́
́
doba, Spain
*
ABSTRACT: The thermal degradation of cannabichromene
CBC, 3) is dominated by cationic reactions and not by the
(
pericyclic rearrangements observed in model compounds. The
rationalization of these differences inspired the development of a
process that coupled, in an aromatization-driven single opera-
tional step, the condensation of citral and alkylresorciniols to
homoprenylchromenes and their in situ deconstructive annula-
tion to benzo[c]chromenes. This process was applied to a total
synthesis of cannabinol (CBN, 5) and to its molecular editing.
espite the availability of phytocannabinoid-based main-
D
stream drugs and significant advances in our knowledge
1
on their targets and mode of action, medicinal marijuana, that
is, the inhaling of vapors of crude Cannabis products
(
flowerheads, cannabis oil, hashish), is still receiving
considerable attention by both the media and the biomedical
2
community. Within the various arguments raised to justify
research in medicinal marijuana, a popular one is our limited
knowledge on the complex phytocannabinoid mixtures that
Figure 1. Thermal rearrangement of the cannabichromene (CBC)
model compound 1 to afford a mixture of cis/trans Δ -THC
analogues (2).
3
originate when crude Cannabis products are vaporized. By
9
4
decarboxylating acidic cannabinoids to their neutral forms,
heating dramatically changes the profile of Cannabis products,
and more complex reactions that generate new phytocanna-
binoids and/or interconvert existing ones are also possible
because of the tendency of phytocannabinoids to undergo
complex rearrangements. In this context, based on the
reactivity of the dihydrochromenone model 1 (Figure 1) and
6
so high (77%) as claimed for compound 1. The model
reaction, while mimicking with silica gel the acidic pyrolytic
milieu of Cannabis, was, nevertheless, mechanistically ques-
tionable in terms of substrate selection. Thus, because of
regulatory considerations associated with the generation of
THC, a Schedule 1 compound in USA, a nonaromatic version
5
on calculations, it was postulated that cannabichromene (CBC,
6
3
), a non-narcotic major constituent of hemp leaves, could
generate a mixture of narcotic tetrahydrocannabinols (THCs,
) by electroreversion to a quinonemethide and intramolecular
cycloaddition to the terminal isoprenyl double bond (Figure
of CBC (1) was used, overlooking the relevance of
aromaticity loss for the electrocyclic opening that triggers the
deconstructive annulation. Surprisingly, this study had no
follow-up, and the possibility that THC could be generated
from CBC under pyrolytic conditions was not further
investigated.
These considerations, and our interest in the rearrangements
of cannabinoids, provided a rationale to study the thermal
4
6
1
). If confirmed, this claim would have the significant
implication that also the leaves of fiber hemp could be
considered potentially narcotic. The claim was, however,
surprising since the chemistry of cannabichromene has been
investigated since the outset of modern studies on
8
7
cannabinoids, and such an important reaction could hardly
Received: July 2, 2019
have escaped detection, especially if the conversion yield was
©
XXXX American Chemical Society
A
DOI: 10.1021/acs.orglett.9b02258
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
degradation of CBC (3) under the conditions of the
conversion of 1 to 2. While disproving this conversion, our
study led to the development of a practical one-pot synthesis of
cannabinol (CBN, 5), the first phytocannabinoid to be
9
10
purified and structurally elucidated.
When CBC was treated with silica gel under the conditions
used for the model compound 1 (Figure 1), a mixture of three
6
Figure 3. Possible rationalization for the formation of cannabicyclol
(
CBL, 8) from the acidic thermolysis of cannabichromene (CBC, 3)
compounds (6−8) was obtained (Figure 2). Any THC isomer,
+
+
1
(E = H or a Lewis acid, R = n-C H11). The reaction is a mechanistic
5
if formed at all, was below the detection threshold of the H
pericyclic “false friend”, just like the Staudinger β-lactam synthesis.
NMR spectrum (700 MHz) of the reaction mixture.
Regeneration of aromaticity by decomplexation and tautome-
rization eventually terminates the reaction.
Compared to the thermal degradation of the model
compound 1, the one of CBC (3) is characterized by
replacement of the electrocyclic opening of the chromene
ring by a series of cationic processes, a predictable behavior
since in the natural product electroreversion involves loss of
15
aromaticity. Dihydroresorcinols are difficult to aromatize,
and the deconstructive annulation exemplified by the
conversion of 1 to 2 has therefore remained relegated into
the limbo of the generation of “model compounds” for natural
chromenes. Spurred by the potential of the reaction to
generate natural product diversity, we wondered if aromatiza-
tion of the carbocyclic moiety could provide an alternative
driving force to steer the reaction toward the electrocyclization,
overcoming, or at least balancing, the unfavorable initial
dearomatization step. Since iodine can efficiently aromatize p-
Figure 2. Pyrolysis of cannabichromene (150 °C, microwave 300 W,
2
10 min) in the presence of silica gel.
11
9
Cannabicitran (citrylidene-cannabis, bicyclo-CBC, 6) and
menthenes, including Δ -THC, via a tandem iodine addition−
8
12
21,22
Δ -iso-cis-THC (cyclo-CBC, 7) are derived from the
protonation of the chromene olefin bond and formation of a
benzyl cation, next trapped by the terminal isoprenyl double
hydrogen iodide elimination mechanism,
we wondered if
iodine, rather than acids, could materialize the chemistry
postulated for the model compound, eventually turning CBC
8
9
bond, with termination either by proton loss to Δ -iso-cis-THC
7) or by intramolecular oxygen trapping to cannabicitran (6).
into cannabinol (CBN, 5), the aromatized version of Δ -THC.
(
CBN is a major constituent of traditionally produced
10
The structure of the third compound, cannabicyclol (CBL, 8),
was confirmed by comparison with an authentic sample
prepared by intramolecular [2 + 2] photocycloaddition of
hashish, but mostly for its limited availability, it has largely
remained terra incognita in terms of bioactivity studies. This is
9
surprising since the conversion of Δ -THC into CBN is
1
3
CBC (45% yield) and CBC treatment with FeCl (65%
relatively fast (5% per month at room temperature in hashish
3
1
4
23
yield). The three compounds formed in the pyrolytic study
had already been described from the treatment of CBC and
other cannabinoids with acids and/or by heating or
and even faster in extracts), suggesting that the early clinical
studies on Cannabis were carried out on a material containing
significant amounts of CBN, possibly even higher than those of
5
9
24
irradiation. The formal [2π + 2π] thermal cyclization of
its precursor Δ -THC.
CBC (3) to CBL (8) was, nevertheless, mechanistically
In the event, we were pleasingly surprised to observe that the
intriguing, and a similar transformation is documented also for
treatment of CBC (3, general formula 13 with R = n-C H )
5
11
15−17
other homoisoprenylchromenes.
considered a Gassman-type cationic [2π + 2π] cyclization,
The reaction has been
under the conditions we had developed for the aromatization
18
9
of Δ -THC (refluxing in toluene with two equivalents of
22
but the anti-Markovnikow regiochemistry is at odds with the
considerations of carbocation stability that governs this type of
iodine) led, in a spot-to-spot fashion, to its conversion to
CBN in 82% yield (5, general formula 14 with R = n-C H ,
5
11
1
8,19
cycloaddition.
An alternative, and more plausible mecha-
Figure 4).
nism, is presented in Figure 3. The reaction might involve a
cyclopropane intermediate (9), the result of a concerted
process triggered by electrophilic (acidic) activation of the
carbonyl tautomer of the resorcinyl moiety, and terminated by
electrophilic Markovnikov addition to the electron-rich
homoisoprenyl terminal double bond (Figure 3). This process,
somewhat reminiscent of the santonine−desmotroposantonine
Other sources of electrophilic iodine species gave a
decreased yield (ICl) or, in the case of N-iodosuccinimide
(NIS), afforded a mixture of iodinated products derived from
attack to the aromatic resorcinyl moiety (15, 20.6% and 16,
5.6%) or from an iodine-induced citran-type cyclization (17,
21.0%) (Figure 2). Taken together, these observations
highlight the delicate role of the iodine source to trigger the
2
0
rearrangement, generates a tertiary cation that is turned into
an oxonium ion by opening of the cyclopropane ring.
deconstructive annulation. The hidden Bronsted acidity of
̈
iodine, that is, its capacity to generate HI by interaction with
B
DOI: 10.1021/acs.orglett.9b02258
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
The one-pot total synthesis of CBN from two commodities
citral and olivetol) solves a long-standing problem in
(
cannabinoid chemistry and natural product synthesis, since
aryl coupling of a resorcinyl benzoate, the most obvious
retrosynthetic strategy to cannabinol, gives the wrong
29
regioisomer (unnatural CBN). Thus, in his classic studies
on cannabinoids, Adams managed to synthesize via aryl
coupling of resorcinyl benzoates all possible regioisomers of
10
CBN except the natural one. This regiochemical issue was
only recently solved by Hertweck using a different tether and a
different aryl coupling strategy (photosplicing of a benzylsul-
30
fonamidic precursor).
Our one-pot synthesis of dibenzochromenes is general for 5-
alkylresorcinols, an important and densely populated class of
31
plant and microbial secondary metabolites, and citral (Figure
). In this way, it was possible to edit the structure of
Figure 4. Total synthesis of cannabinol (5) and analogues. In the one-
pot procedure for cannabinol (5), Dowex 50 WX8 was added after the
first step, and after filtration over Celite, iodine was added. CBN was
obtained in 55% yield.
4
cannabinol, expeditely generating natural (14a,b) and
5
synthetic analogues from the α,α-dimethylheptyl (14c) and
5
the phenethyl (14d) series useful to explore the biological
space associated with its chemotype (Figure 4).
hydroxyl groups from the substrate or from traces of protic
However, the deconstructive annulation of homoprenyl-
chromenes also has some limitations. The reaction was
investigated with higher homologues of citral (C-10), namely,
farnesal (C-15) and geranylgeranial (C-20), but in both cases,
complex mixtures were formed. In addition, nonsymmetric
resorcinols like 4-hexylresorcinols afford isomeric chromenes
in the condensation step, and even when the single isomers
were reacted with iodine, parasite reactions plagued the
deconstructive annulation, affording mixtures. Conjugating
groups, like double bonds or a carbonyl, on the resorcinol
moiety are not tolerated, and regrettably, the reaction could
not be applied to resorcinol flavonoids and stilbenoids, two
very common classes of natural products. Attempts to expand
the application of the reaction to heteroaromatic analogues of
alkylresorcinols also met with limited success. Thus, the
2
5
solvents, was ruled out as a possible mechanism since
treatment of CBC with Bronsted acids affords compounds
̈
resulting from the formation of a benzyl cation and not by
7
cycloreversion. A soft polarizability/polarization of the I−X
bond seems, conversely, critical to avoid electrophilic attack to
the electron-rich aromatic ring and the dihydropyrane double
bond. Electroreversion could then be promoted by halogen
25
bonding to the chromene oxygen and/or to its aromatic ring,
while the final aromatization eventually funnels the various
equilibria toward the generation of dibenzochromenes. It is
remarkable that, despite the plethora of reaction pathways
available, only the deconstructive annulation was observed, at
1
least within the limit of H NMR sensitivity, by analysis of the
crude reaction mixture.
32
pyranocoumarin ferprenin (18) gave the corresponding
dibenzochromene (19) in a mixture with two “benzylic”
cyclization products (20a,b) that became the exclusive reaction
products with the pyranopyrone 21 (Figure 5).
In a further development of the reaction, we wondered if
iodine-promoted aromatization could redirect the condensa-
tion of citral (10) and resorcinols (11) from the generation of
chromenes (13) to the one of benzochromenes (14) since it
generates the same quinonmethide intermediate 12 postulated
for the deconstructive annulation of chromenes to benzochro-
menes (Figure 4), and iodine has been reported to promote
the condensation of cyclic β-dicarbonyls and unsaturated
26
aldehydes to dihydrochromenes. However, the direct treat-
ment of citral (10) and olivetol (11, R = n-C H ) with iodine
5
11
only produced a complex mixture, confirming that cyclo-
hexandiones are not good models of resorcinols, and amine
catalysis is required for the initial condensation step. Amines
Figure 5. Attempts to extend the chromene deconstructive annulation
to isosteric cannabichromene analogues.
2
7
(
pKBI2 ca. 3.8) show a stronger affinity for iodine than an
oxygen ether or an aromatic ring (pKBI2 ca. 0 in both cases),
27
and complexation with an amine evidently completely
neutralizes the σ-hole electrophilicity of iodine. To overcome
25
Even with these limitations, the simplicity and the easy
availability of its starting material make the iodine-mediated
deconstructive annulation of homoisoprenylchromenes to
benzo[c]chromenes a remarkable example of expedite
generation of biologically relevant chemical diversity from
simple building blocks, encouraging the biological evaluation
of compounds of limited availability by isolation.
this impasse, we scavenged the amine required for the
28
chromenylation (n-butylamine) with an acidic ion resin,
adding the iodine to the filtered solution and continuing
refluxing. In this way, after aqueous workup of the reaction,
CBN could be obtained in a rewarding 55% yield directly from
olivetol and citral (Figure 4).
C
DOI: 10.1021/acs.orglett.9b02258
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
(
13) Crombie, L.; Ponsford, R.; Shani, A.; Yagnitinsky, B.;
ASSOCIATED CONTENT
Supporting Information
■
Mechoulam, R. Tetrahedron Lett. 1968, 9, 5771.
*
S
(14) Li, X.; Lee, Y. R. Org. Biomol. Chem. 2014, 12, 1250.
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.or-
glett.9b02258.
(15) Riveira, M. J.; La-Venia, A.; Mischne, M. P. J. Org. Chem. 2016,
81, 7977−7983.
(16) Yeom, H.-S.; Li, H.; Tang, Y.; Hsung, R. P. Org. Lett. 2013, 15,
3
(
6
130.
17) Kurdyumov, A. V.; Hsung, R. P. J. Am. Chem. Soc. 2006, 128,
272.
Full experimental details and the characterization of
compounds (PDF)
(
(
18) Gassman, P. G.; Lottes, A. C. Tetrahedron Lett. 1992, 33, 157.
19) Roberts, R. M.; Landolt, R. G.; Greene, R. N.; Heyer, E. W. J.
AUTHOR INFORMATION
Corresponding Authors
■
Am. Chem. Soc. 1967, 1404.
(20) Andreocci, A. Gazz. Chim. Ital. 1893, 23, 469.
*
Prof. O. Taglialatela-Scafati. E-mail: scatagli@unina.it.
(21) Domingo, V.; Prieto, C.; Silva, L.; Rodilla, J. M. L.; Quílez del
Moral, J. F.; Barrero, A. F. J. Nat. Prod. 2016, 79, 831.
*Prof. G. Appendino. E-mail: giovanni.appendino@uniupo.it.
(22) Pollastro, F.; Caprioglio, D.; Marotta, P.; Schiano Moriello, A.;
ORCID
De Petrocellis, L.; Taglialatela-Scafati, O.; Appendino, G. J. Nat. Prod.
Alberto Minassi: 0000-0002-3640-2912
Orazio Taglialatela-Scafati: 0000-0001-8010-0180
Giovanni Appendino: 0000-0002-4170-9919
Author Contributions
2018, 81, 630.
(
(
23) Lerner, P. Bull. Narcotics 1969, 21, 39.
24) For an account of the early studies on hashish and cannabinol,
see: Mills, J. H. Cannabis Britannica: Empire, Trade, and Prohibition
800−1928; Oxford University Press: Oxford, 2005.
(25) Breugst, M.; von der Heiden, D. Chem. - Eur. J. 2018, 24, 9187.
26) Jung, E. J.; Lee, Y. R.; Lee, H.-J. Bull. Korean Chem. Soc. 2009,
0, 2833.
27) Laurence, C.; Graton, J.; Berthelot, M.; El Ghomari, M. J.
Chem. - Eur. J. 2011, 17, 10431.
28) Elsohly, E. A.; Turner, C. E.; Clark, A. M.; Elsohly, M. A. J.
Pharm. Sci. 1982, 71, 1319.
29) Adams, R.; Baker, B. B.; Wearn, R. B. J. Am. Chem. Soc. 1940,
2, 2204.
30) Kloss, F.; Neuwirth, T.; Haensch, V. G.; Hertweck, C. Angew.
1
All authors have given approval to the final version of the
manuscript.
Notes
(
3
(
The authors declare no competing financial interest.
(
ACKNOWLEDGMENTS
■
(
6
(
We are grateful to Emerald Biotechnology Spain and MIUR,
Italy (PRIN2017, Project 2017WN73PL, bioactivity-directed
exploration of the phytocannabinoid chemical space), for
financial support. We also thank Vittorio Farina (Farinachem
Consulting GmbH) for discussions and suggestions on the
mechanistic aspects of the formation of cannabicyclol from
cannabichromene.
Chem., Int. Ed. 2018, 57, 14476.
(31) Kozubek, A.; Tyman, J. H. P. Chem. Rev. 1999, 99, 1.
(32) Appendino, G.; Tagliapietra, S.; Nano, G. M.; Picci, V.
Phytochemistry 1988, 27, 944.
REFERENCES
1) Di Marzo, V. Nat. Rev. Drug Discovery 2018, 17, 623.
2) Mouhamed, Y.; Vishnyakov, A.; Qorri, B.; Sambi, M.; Frank, S.
■
(
(
S.; Nowierski, C.; Lamba, A.; Bhatti, U.; Szewczuk, M. R. Drug,
Healthcare Patient Saf. 2018, 10, 45−66.
(
3) Hazekamp, A.; Ruhaak, R.; Zuurman, L.; van Gerven, J.;
Verpoorte, R. J. Pharm. Sci. 2006, 95, 1308.
4) Wang, M.; Wang, Y.-H.; Avula, B.; Radwan, M. M.; Wanas, A. S.;
(
van Antwerp, J.; Parcher, J. F.; ElSohly, M. A.; Khan, I. A. Cannabis
Cannabinoid Res. 2016, 1, 262.
(
5) Hanus,
Appendino, G. Nat. Prod. Rep. 2016, 33, 1357.
6) Garcia, A.; Borchardt, D.; Chang, C.-D. A.; Marsella, M. J. J. Am.
̌
L. O.; Meyer, S. M.; Munoz, E.; Taglialatela-Scafati, O.;
̃
(
Chem. Soc. 2009, 131, 16640. For a study on the biological profile of
cis-THC, see: Chicca, A.; Schafroth, M. A.; Reynoso-Moreno, I.; Erni,
R.; Petrucci, V.; Carreira, E. M.; Gertsch, J. Sci. Adv. 2018, 4, 2166.
(
7) Pollastro, F.; Caprioglio, D.; Del Prete, D.; Rogati, F.; Minassi,
A.; Taglialatela-Scafati, O.; Munoz, E.; Appendino, G. Nat. Prod.
Commun. 2018, 13, 1189.
8) Lopatriello, A.; Caprioglio, D.; Minassi, A.; Schiano Moriello, A.;
̃
(
Formisano, C.; De Petrocellis, L.; Appendino, G.; Taglialatela-Scafati,
O. Bioorg. Med. Chem. 2018, 26, 4532.
(
9) Wood, T. B.; Spivey, W. T. S.; Easterfield, T. H. J. Chem. Soc.,
Trans. 1896, 69, 539.
10) Adams, R.; Hunt, M.; Clark, J. H. J. Am. Chem. Soc. 1940, 62,
96.
11) Crombie, L.; Redshaw, S. D.; Whiting, D. A. J. Chem. Soc.,
Chem. Commun. 1979, 630.
12) Yagen, B.; Mechoulam, R. Tetrahedron Lett. 1969, 10, 5353.
(
1
(
(
Modern numbering would locate the double bond at C-11, and the
name iso-THC is, nevertheless, confusing.
D
DOI: 10.1021/acs.orglett.9b02258
Org. Lett. XXXX, XXX, XXX−XXX