Collective Syntheses of Icetexane Natural Products Based on Biogenetic Hypotheses

Article abstract of DOI:10.1002/chem.201603932

A divergent synthesis of 10 icetexane natural products based on a proposed biogenetic cationic ring expansion of a reduced carnosic acid derivative is described. Of these icetexanes, (+)-salvicanol, (?)-cyclocoulterone, (?)-coulterone, (?)-obtusinone D, and (?)-obtusinone E have been synthesized for the first time. In addition, the hypothesis for the non-enzymatic biogenesis of benzo[1,3]dioxole natural products has been experimentally investigated. Additional experimental evidence for the abiotic formation of the methylenedioxy unit is provided, as photolysis of the quinone (+)-komaroviquinone resulted in the formation of the [1,3]dioxole-containing natural product (?)-cyclocoulterone and (+)-komarovispirone.

Full text of DOI:10.1002/chem.201603932

DOI: 10.1002/chem.201603932
Full Paper
&
Natural Products
Collective Syntheses of Icetexane Natural Products Based on
Biogenetic Hypotheses
Christophe Thommen,[a, b] Markus Neuburger,[b] and Karl Gademann*[a, b]
Abstract: A divergent synthesis of 10 icetexane natural
products based on a proposed biogenetic cationic ring ex-
pansion of a reduced carnosic acid derivative is described.
Of these icetexanes, (+)-salvicanol, (ꢀ)-cyclocoulterone, (ꢀ)-
coulterone, (ꢀ)-obtusinone D, and (ꢀ)-obtusinone E have
been synthesized for the first time. In addition, the hypothe-
sis for the non-enzymatic biogenesis of benzo[1,3]dioxole
natural products has been experimentally investigated. Addi-
tional experimental evidence for the abiotic formation of the
methylenedioxy unit is provided, as photolysis of the qui-
none (+)-komaroviquinone resulted in the formation of the
[1,3]dioxole-containing natural product (ꢀ)-cyclocoulterone
and (+)-komarovispirone.
Introduction
Terpenes display an extraordinary wealth of fascinating chemi-
cal structures, which has provided inspiration for chemists over
decades.[1] The understanding and development of cyclization
reaction cascades,[2] cationic rearrangements,[3] and the mecha-
nism of CꢀH oxidation[4] have greatly influenced science, from
reaction mechanisms to biosynthesis. The methylenedioxy-
bridged phenol unit, benzo[1,3]dioxole, is found in many com-
pounds,[5] including commercially important compounds such
as the fragrances piperonal and helionalꢀ and drugs such as
ecteinascidin 743 or etoposide (Figure 1). Consequently, the
biosynthesis of this important moiety has been investigated
for decades by chemical, enzymatic, and genetic approaches.[6]
In a landmark study published in 1962, Sir Derek Barton re-
ported experimental evidence that the origin of the ben-
zo[1,3]dioxole unit found in the alkaloid haemanthamine re-
sides in a monomethyl-catechol unit, and he postulated that
this transformation proceeds by an oxidative process.[7] This hy-
pothesis has been substantiated by many experiments, and in
recent years the involvement of cytochrome P450 enzymes in
the oxidative cyclization has been unequivocally demonstrated
for several natural products[6] (Scheme 1, top). In the context of
taiwaniaquinol A, we have previously identified an alternative
pathway to benzo[1,3]dioxole formation that is 1) redox-neu-
tral, 2) light-mediated, and, most importantly, 3) without the in-
volvement of any enzymes (Scheme 1, bottom).[8]
Figure 1. Terpenoids featuring the benzo[1,3]dioxole unit, including those
with commercial interest.
[a] Dr. C. Thommen, Prof. Dr. K. Gademann
Department of Chemistry, University of Zurich
Winterthurerstrasse 190, 8057 Zꢀrich (Switzerland)
E-mail: karl.gademann@uzh.ch
[b] Dr. C. Thommen, Dr. M. Neuburger, Prof. Dr. K. Gademann
Department of Chemistry, University of Basel
St-Johanns-Ring 19, 4056 Basel (Switzerland)
Supporting information and ORCID number(s) for the author(s) for this arti-
cle are available under http://dx.doi.org/10.1002/chem.201603932.
Scheme 1. Biogenesis of benzo[1,3]dioxole natural products.
Chem. Eur. J. 2016, 22, 1 – 9
1
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
In this report we present evidence further corroborating this
abiotic hypothesis through the preparation of two benzo[1,3]-
dioxole natural products in the chemical laboratory and de-
scribe a model study aimed at elucidating the mechanistic de-
tails of their synthesis. In addition, we investigated potential
routes for the biogenesis of icetexane terpenes and provide
the first experimental evidence for the abiotic formation of the
tertiary OH group during the ring expansion from the abietane
to the icetexane family. Finally, the abiotic formation of the
complex C40 dimers obtusinones D and E is reported and their
configurations are reassigned. In summary, the first preparation
in synthetic form of five natural products is reported, that is,
(+)-salvicanol, (ꢀ)-cyclocoulterone, (ꢀ)-coulterone, (ꢀ)-obtusin-
one D, and (ꢀ)-obtusinone E, along with alternative ap-
proaches for the preparation of komaroviquinone, komarovis-
pirone, brussonol, przewalskin E, and demethylsalvicanol qui-
none.
yield.[10] However, these authors detected an unknown product
in low yield, which, however, was not characterized at the
time. We expected that the use of the conditions developed
for the photolysis of taiwaniaquinone F would switch the
course of the photolysis of komaroviquinone (2) in favor of the
generation of cyclocoulterone (1). If successful, this abiotic,
non-enzymatic transformation of komaroviquinone (2) into cy-
clocoulterone (1) would further substantiate the hypothesized
abiotic formation of benzo[1,3]dioxoles. To test this hypothesis,
the icetexane precursor 2 first had to be prepared in synthetic
form.
Results and Discussion
Investigation of the ring expansion in icetexane biogenesis
Icetexanes are a variety of diterpenoids isolated from diverse
plant sources, in particular, from the Labiatae family.[11] These
compounds exhibit a unique carbon skeleton consisting of a 6-
7-6 tricyclic core structure (Figure 2). The systematic name
9(10!20)-abeo-abietane given to icetexanes indicates that an
abietane precursor such as 4 could be involved in the biosyn-
thetic origin of icetexanes (Scheme 3). To the best of our
knowledge, Watson and Dominguez and co-workers were the
first to propose such a connection between abietanes and ice-
texanes,[12] but several other research laboratories have pro-
posed similar mechanisms to connect abietanes to their rear-
ranged congeners.[13] Hydride abstraction from the angular
methyl group (X=H) or solvolysis of the primary alcohol (X=
OH) of abietane 4 generates a primary cation that can immedi-
We have previously reported the photolysis of taiwaniaqui-
none F to taiwaniaquinol A (Scheme 2A).[8] We hypothesized
that such
a 1,2,4-hydroxylated aryl moiety incorporating
a methylenedioxy bridge can be biosynthetically generated
through a light-promoted remote CꢀH activation. Regarding
the biosynthesis of this structural unit, this postulate would be
complementary to the well-accepted and established cytochro-
me P450 mechanism[6] for producing the methylenedioxy
bridge, especially in the plant kingdom in which the use of
light as a source of energy is omnipresent. Therefore, the first
goal of this study consisted of demonstrating that the previ-
ously discovered photolysis of taiwaniaquinone F can also be
applied to additional substrates in order to emphasize the gen-
erality of this transformation in Nature. Therefore, we proposed
that the complex benzo[1,3]dioxole derivative cyclocoulterone
(1;[9] Scheme 2B) could be generated from komaroviquinone
(2)[9] by an abiotic ring-forming reaction such as that postulat-
ed above. In support of the hypothesized biogenetic basis of
this transformation, both cyclocoulterone (1) and komaroviqui-
none (2) were isolated from the same plant (Dracocephalum
komarovi), thereby corroborating a possible biogenetic connec-
tion between the two natural products.[9] Majetich and Jianhua
previously described the photolysis of komaroviquinone (2) in
cyclohexane, which furnished komarovispirone (3) in high
Figure 2. Examples of icetexane natural products.[11]
Scheme 2. Remote CꢀH functionalization of taiwaniaquinone F and komaro-
viquinone (2).
Scheme 3. Hypothetical biogenetic abietane ring expansion to icetexanes.
X=H or OH.
&
&
Chem. Eur. J. 2016, 22, 1 – 9
www.chemeurj.org
2
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
ately rearrange to the more stable ring-expanded tertiary car-
bocation 5 thereby forming the icetexane core structure. Ring
expansions leading to alkenes 6 by a deprotonation of carbo-
cation 5 are experimentally supported.[14] Indeed, such rear-
rangements were performed by treating primary hydroxyabie-
tanes such as 4 with either thionyl chloride in dry benzene or
p-TsCl in dry pyridine to yield alkene icetexanes such as 6.[14]
Although trapping of carbocation 5 by a water molecule has
frequently been suggested as a route to access icetexanes
bearing a 10-hydroxy group, such as in alcohol 7, no experi-
mental evidence of such a mechanism has been reported.[13] In
this report, we provide the first experimental evidence in sup-
port of this transformation. In particular, we demonstrate that
no enzymes are necessary for the attack on the b face.
Scheme 5. Preparation of salvicanol (12) based on a biogenetic ring-expan-
sion hypothesis. Reagents and conditions: a) LAH, Et2O, 458C, 6 h, 89%;
b) MsCl, Et3N, THF, ꢀ108C, 20 min; H2O, 408C, 12 h, 94%; c) l-selectride solu-
tion in THF, 808C, 2 h, 80%; d) (i) m-CPBA, CH2Cl2, 08C, 2 h; (ii) l-selectride
(excess) solution in THF, 808C, 3 h, 42% (2 steps); e) Ac2O, pyridine, RT, 1 h,
90%.
The synthesis started with carnosic acid, a commercially
available and naturally abundant abietane.[15] Carnosic acid is
found in significant amount in Lamiaceae and particularly in
Rosmarinus officinalis and Salvia officinalis. The content of car-
nosic acid in the dried leaves varies in proportion from 1 to
10% (by weight), depending on the species, the plant age,
and the environmental stress.[16] The extraction, purification,
and isolation of carnosic acid was achieved by HPLC of crushed
and dried leaves of rosemary (Rosmarinus officinalis, origin:
Portugal) as reported by Albu et al.[17] However, the triple
methylation of pure carnosic acid by using the modified condi-
tions of Theoduloz et al. resulted only in a moderate yield of 8
(66%).[18] Encouraged by literature reports of related constitu-
ents of rosemary, we changed our strategy to obtain higher
quantities of ester 8.[17,19] In this regard, the dried ethanolic ex-
tract of rosemary was dissolved in Et2O and extracted with
NaOH (1m) to form the aqueous soluble sodium salts of differ-
ent phenols (Scheme 4). Acidification (1m aq. HCl) and ethereal
in Scheme 3. To this end, the primary alcohol 9 was treated
with different acids in aqueous solution. Only strong acids
were able to promote a ring expansion leading to an insepara-
ble mixture of alkenes 10 in low yields. Furthermore, such
harsh conditions prevented the formation of alcohol 11, proba-
bly due to direct elimination. Finally, basic treatment of 9
(MsCl, Et3N in THF followed by the addition of water) resulted
in the tertiary alcohol 11 bearing the icetexane skeleton in
a yield of 32% as a mixture with isomeric olefins 10 in an ex-
cellent overall yield of 94%. This remarkable transformation
warrants a number of mechanistic comments. First, it is likely
that the mesylate undergoes E1 elimination to generate the
primary cation. As can be deduced from the product, a cation
similar to 5 would have to be generated through a 1,2-alkyl
shift. Alternatively, a concerted process in which these two
steps merge could be operative. The next step now constitutes
the key scientific problem and has been debated in the litera-
ture for decades.[12,13] The attack of water on such a carbocation
was suggested to be disfavored from the b face according to
Dreiding models, therefore supporting an enzymatically
guided attack.[13b] However, our exclusively observed b selectivi-
ty refutes this enzymatic hypothesis and thus establishes
a non-enzymatic mechanism.[13b] This ring expansion is the first
experimental study invoking a water-mediated trapping of
a tertiary carbocation such as 5. In addition, this is the first ex-
perimental validation of the hypothesized biosynthetic forma-
tion of icetexane bearing a C10 tertiary alcohol, shown earlier
(Scheme 3).
Scheme 4. Isolation and derivatization of carnosic acid.
extraction of the aqueous phase yielded a phenolic abietane-
enriched extract, which was subjected to exhaustive methyla-
tion (NaH, Me2SO4, THF, then MeOH) to yield a mixture of the
desired trimethylated carnosic acid 8 and byproducts. The mix-
ture was further refined by dissolving the constituents in
MeOH followed by filtration and subsequent hydrogenation
using Pd(OH)2/C, which resulted in pure dimethoxycarnosate
methyl ester 8 (1.3 wt% from dried and crushed rosemary). A
smooth reduction of the angular methoxy ester group of 8
was achieved by a modified literature protocol using lithium
aluminium hydride (LAH) in hot diethyl ether, which resulted in
alcohol 9 in excellent yield (89%, Scheme 5).[14b]
A selective hydride-mediated demethylation (l-selectride in
THF, 808C, 2 h) originally developed by Majetich et al. was suc-
cessfully applied to the tertiary alcohol 11.[20] Indeed, the steri-
cally less congested 11-methoxy group was selectively deme-
thylated with complete regioselectivity in high yield (80%) to
result in the first synthesis of salvicanol (12). All the spectro-
scopic data gave good agreement with those reported for the
natural sample.[21] The alkenes 10, obtained in the course of
the ring expansion of 9, were converted into salvicanol (12) in
two consecutive steps consisting of an epoxidation carried out
by using m-chloroperoxybenzoic acid (m-CPBA)[22] and a well-
Having the desired alcohol 9 in hand, we wished to experi-
mentally investigate the ring-expansion hypothesis as outlined
Chem. Eur. J. 2016, 22, 1 – 9
www.chemeurj.org
3
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
established one-pot hydride (l-selectride)-mediated epoxide
opening/demethylation in a yield of 42% over the two
steps.[23] Salvicanol (12) was acetylated in high yield (90%) to
afford salvicanol acetate (13), which was used later on for the
subsequent oxidation reaction.[21]
The successful formation of 17 in the presence of TBHP en-
couraged us to investigate different combinations of transi-
tion-metal catalysts and TBHP (entries 5–12). Indeed, the for-
mation of the desired product 17 was observed by utilizing
several transition metals (entries 6–9, 11, and 12). The optimal
conditions were identified when treating acetate 13 under
a modified Hirao protocol (RuCl3·xH2O, TBHP, entry 12).[26] The
benzylic oxidation product 17 was obtained in moderate yield
(31%) along with the starting material (13, 19%). NMR analysis
of this compound was difficult due to keto–hemiketal tauto-
merism between the open (17) and closed form (18), as al-
ready observed for a similar hydroxy-ketone by Suto et al.[27]
Reaction conditions involving transition metals without TBHP
were also investigated (entries 13–16).[28] However, compound
19 (and 20, entry 16) or 17 (and 18, entries 13 and 15) was
produced only in trace amounts. These experiments revealed
RuCl3/TBHP to be the optimal oxidant, and was used in the
synthesis (Scheme 6).
Experimental evidence for abiotic benzo[1,3]dioxole forma-
tion
Having the methyl and acetyl derivatives of salvicanol (11 and
13, respectively) in hand, a C7 benzylic oxidation and an aro-
matic oxidation were required to accomplish the synthesis of
komaroviquinone (2). Majetich et al. reported a difficult C7
benzylic oxidation of a substrate very similar to 12 and 13.[24]
They concluded that the 10-OH group would not tolerate oxi-
dative conditions and result in a C10–C20 oxidative cleavage.
These results could be reproduced with substrate 12 and 13
under similar conditions (CrO3, AcOH/H2O, Table 1, entries 1
and 2) to give the unstable aldehydes 14 and 15, but only
trace amounts were obtained. Further attempts involving the
possible formation of benzylic radicals, such as the use of KBr
in the presence of Oxoneꢀ (entry 3), resulted in the formation
of the side-product 16, presumably by bromination with a Br+
species.[25] The transition-metal-free oxidation of 13 using tert-
butyl hydroperoxide (TBHP) under microwave conditions
(1808C, CH3CN, entry 4) resulted in trace amounts of the de-
1
sired ketone 17 (detected by H NMR spectroscopy).
Table 1. Benzylic oxidation of the salvicanol derivatives 11 and 13.
Scheme 6. Completion of the synthesis of komaroviquinone (2). Reagents
and conditions: a) RuCl3·xH2O, CH2Cl2/Pyr, tBuOOH, 508C, 12 h, 31%, 50%
brsm; b) KOH, MeOH, RT, 12 h; c) salcomine, O2 (5 bar), MeCN, 12 h, 56% (2
steps); d) sunlight, Et2O, 48C (winter), 45 min, 63%; e) Na2S2O4 (aq.), Et2O,
1 min.
To complete the synthesis of komaroviquinone (2,
Scheme 6), the ketone 17 was first deacetylated by using KOH
in methanol at room temperature for 12 h. The crude mixture
containing phenol 21 was subsequently submitted to aromatic
oxidation using salcomine in an O2 atmosphere with the strict
exclusion of light to produce synthetic komaroviquinone (2) in
a yield of 56% over two steps. This modest yield could be ex-
plained by the presence of the C7 keto group, which lowers
the electron density of the aromatic ring and therefore de-
creases the reactivity towards oxidation. Furthermore, komaro-
viquinone (2) was obtained as the sole tautomer, which could
be explained by the presence of the newly formed hydrogen
bond at the 14-keto group and by the increased electrophilici-
ty of the C7 vinylogous ketone compared with the benzylic
ketone in structure 17.
Entry
Conditions
Substrate
Observations
1
2
3
4
5
6
7
8
CrO3
CrO3
11
13
11
13
11
13
13
13
13
13
13
13
13
13
13
11
14[a]
15[a]
KBr, Oxone, DCM/H2O
TBHP, microwave
[Cr(CO)6], TBHP
16[a]
17[a]
no conversion
17: 10%
17[a]
17:10%, 13: 40%
17[a]
no conversion
17: 17%, 13: 13%
17: 31%, 13: 19%
17[a]
[Cr(CO)6], TBHP
[Rh2(cap)4], TBHP
[Ru(Cl)2(PPh3)2], TBHP
[ReOCl3(OPPh3)(SMe)2], TBHP
[VCp2Cl2], TBHP
Bi0, picolinic acid, TBHP
RuCl3·xH2O, TBHP
Fe(OTf)2, ligands, H2O2
NaIO4, RuCl3·xH2O
KMnO4/CuSO4·5H2O
KMnO4/CuSO4·5H2O
9
10
11
12
13
14
15
16
complex mixture
17[a]
19[a]
Next, komaroviquinone (2) was subjected to photolysis, ap-
plying similar conditions to those used for the synthesis of tai-
waniaquinol A (Et2O, sunlight, RT).[8] We were delighted to ob-
serve the formation of the cyclocoulterone (1) as well as the
[a] Observed in trace amounts.
&
&
Chem. Eur. J. 2016, 22, 1 – 9
www.chemeurj.org
4
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
known komarovispirone (3) in a nearly equimolar ratio in an
overall yield of 63%. These findings further supported our pre-
viously non-enzymatic biogenetic hypothesis for the biosyn-
thesis of such a methylenedioxy-bridged natural product. In
addition, this is the first time that synthetic cyclocoulterone (1)
has been produced.
Additionally, komaroviquinone (2) was reduced in an aque-
ous ethereal solution of Na2S2O4 in a separatory funnel at room
temperature for 1 min to afford coulterone (22) in high yield
(82%). This natural product was isolated over 20 years ago,[29]
but had never been synthesized.
The previous photolysis of komaroviquinone (2), achieved
with a low-pressure mercury lamp by Majetich and Jianhua in
cyclohexane, led to the formation of komarovispirone (3) as
the sole product.[10] However, under irradiation with sunlight,
the formation of an additional unknown compound in CDCl3
solution was reported. This byproduct could correspond to cy-
clocoulterone (1), but no analytical data was reported[10] to
verify this hypothesis.
Scheme 8. Synthesis of model substrate 23 and its photolysis. Reagents and
conditions: a) see ref. [XXX], 45% over 2 steps; b) TiCl4, CH2Cl2, Cl2CHOMe,
ꢀ108C to RT, 2 h, 96%; c) Pd(OH)2/C, MeOH, H2 (80 bar), 24 h, 77%; d) TiCl4,
CH2Cl2, Cl2CHOMe, ꢀ108C to RT, 12 h, 98%; e) Pd(OH)2/C, MeOH, H2 (80 bar),
18 h, 86%; f) AgO, HNO3 (6m), dioxane, RT, 80 min, 62%; g) medium-pressure
Hg lamp, tBuOH, 45 min, RT, 47%.
thoxy-3-isopropylbenzene (24), which was obtained from the
commercially available 25 in two steps in an overall yield of
45%.[XXX] Rieche formylation of 24 yielded aldehyde 26 in an
excellent yield of 96%. The C=O group of 26 was reduced by
using Pd(OH)2/C in methanol under H2 pressure (80 bar) and
the electron-rich methylphenyl 27 was obtained in good yield
(77%). The same formylation and reduction sequence then fur-
nished the hexasubstituted benzene 28, via aldehyde 29, in
a good overall yield of 84%. Silver(II) oxide demethylative oxi-
dation of catechol 28 succeeded to provide the quinone 23 in
good yield (62%). Having this model system in hand, remote
CꢀH functionalization was explored with quinone 23. Applying
the conditions developed earlier (direct sunlight, Et2O, RT) led
to full conversion of the starting material (observed by the de-
coloration of the reaction mixture and confirmed by TLC analy-
sis), however, without the formation of the desired methylene-
dioxy bridge. Different solvents (benzene, tBuOH, CHCl3) and
light sources (sunlight, medium-pressure mercury lamp) were
evaluated next. The photolysis of 23 in tBuOH was effective
and resulted in the formation of the desired compound 30
featuring the methylenedioxy bridge in a yield of 47% by
using a medium-pressure mercury lamp at room temperature.
The feasibility of this reaction in tBuOH rather than in Et2O
could be explained by the higher ability of tBuOH to form hy-
drogen bonds with a phenoxyl radical and stabilize a putative
biradical species, which would favor the formation of the
methylenedioxy-bridged compound 30.
A possible mechanism to rationalize the formation of cyclo-
coulterone (1), in accordance with the suggested proposal for
the formation of komarovispirone (3),[10] is depicted in
Scheme 7. Light excitation of the n!p* transition of the qui-
none 2 could generate the quinol diradical A, which could un-
dergo a 1,5 hydrogen atom transfer (HAT) between the alkoxyl
radical at C14 and the hydroxy group at C7 to give B. At this
point, compound B could fragment either at the C6ꢀC7 bond
to form the biradical C, which recombines to komarovispirone
(3), or at the C7ꢀO(C10) ether bond, which after successive 1,5
and 1,6 HAT recombines to yield cyclocoulterone (1).
Model substrate synthesis and photolysis
The generality of this reaction was investigated by using
a non-natural model substrate. The fully substituted quinone
23 (Scheme 8) was therefore synthesized from 1,2,4-trime-
Synthesis of members of the barbatusol class natural prod-
ucts
Having salvicanol (12) in hand, the synthesis of several barba-
tusol natural products[11] was investigated (Scheme 9). Deme-
thylative oxidation of salvicanol (12) by using 2,3-dichloro-5,6-
dicyano-1,4-benzoquinone (DDQ) in aqueous acetone generat-
ed the o-quinone 31 in excellent yield (96%) after aqueous
workup.[22,30] However, purification by silica was difficult, as 31
reacted further. Therefore, we investigated the outcome of
prolonged exposure of o-quinone 31 to silica gel upon adsorp-
Scheme 7. Proposed mechanism for the co-formation of cyclocoulterone (1)
and komarovispirone (3) through the photolysis of komaroviquinone (2).
Chem. Eur. J. 2016, 22, 1 – 9
www.chemeurj.org
5
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
31 adsorbed on magnesium silicates (Florisilꢀ) in a microwave
oven (1508C, 5 min, 37%).[30b] However, none of these methods
was suitable for producing przewalskin E (32) directly from 31.
Interestingly, very few reports of catechol oxidation to o-qui-
none mediated on silica gel under aerobic conditions have
been published.[22] Furthermore, the crucial role of silica gel
was further substantiated experimentally when an acetonitrile
or acetone solution of brussonol (33) was stirred under oxygen
atmosphere and only trace amounts of przewalskin E (32) were
obtained after several days.
Synthesis of the proposed structure 32 of przewalskin E and
a proposal for revision of its constitution
The formation of such a reactive quinone methide from o-qui-
none 31 and aerobic catechol oxidation under comparable re-
action conditions has already been reported by Burnell et al.[32]
Brussonol (33) was first isolated from Salvia brussonetii and
shows insecticidal and anti-P388 leukemia cell activity. The
spectral data of 33 gave good agreement with those reported
for the natural and synthetic products.[13c,30b] The spectral data
of quinone 32, first synthetically prepared from natural deme-
thylsalvicanol[30b] and subsequently isolated from Salvia prze-
walskii MAXIM,[33] correspond perfectly with those of the syn-
thetic material, but significantly deviate from the natural
sample. Although the 13C NMR chemical shifts (0.1–0.4 ppm)
and the optical activity of the isolated sample reasonably
Scheme 9. Synthesis of przewalskin E (32), brussonol (33), and first synthesis
of obtusinones D (34) and E (35): Reagents and conditions: a) DDQ, acetone
(aq.), RT, 10 min, 96%; b) SiO2, 48 h, 39–50% (2 steps); c) Na2S2O4 (aq.), Et2O,
1 min, 94%; d) DDQ, dioxane, 80%; e) neat, 1008C, 6 h, 69%.
tion. The quinone 31 was dissolved in diethyl ether, adsorbed
on silica, and subsequently the solvent was removed. After
48 h at room temperature, this impregnated silica was purified
by flash chromatography on silica gel to afford two main prod-
ucts identified as przewalskin E (32) and brussonol (33). Inter-
estingly, when the adsorbed o-quinone 31 was left in an open
flask with frequent mixing of the impregnated silica powder
(every 4 h), a higher yield of oxidized przewalskin E (32, 50%)
was obtained without any detectable formation of brussonol
(33).
The outcome of this reaction could be rationalized by
a silica gel acid-catalyzed tautomerization (Scheme 10), with
the formation of a reactive quinone methide intermediate,
which, upon nucleophilic attack of the angular 10-OH group,
could result in brussonol (33). In aerobic conditions, 33 can be
further converted into its natural congener przewalskin E (32).
Majetich and Zou have already reported such a synthesis of
brussonol (33) from 31 through the heating of a concentrated
solution of 31 in Et2O (608C, 40 h, 70%).[22] Similarly, Takeya
and co-workers described the same transformation by heating
1
match those of our synthetic substance, the H NMR chemical
shifts (0.01–0.08 ppm) and FTIR absorption bands (the main
signal at 1660 cmꢀ1 was not reported and the signals at 1722
and 1680 cmꢀ1 were found with much lower intensity) signifi-
cantly diverge from the synthetic samples. Moreover, the natu-
ral compound was isolated as a white powder and our syn-
thetic sample as a red amorphous solid, the typical color of o-
quinones. Unfortunately, the authors did not compare their
sample with the synthetic compound of Takeya and co-work-
ers.[30b] However, the structure of 32 could be verified by con-
verting it into brussonol (33) in excellent yield (94%) by reduc-
tion with aqueous Na2S2O4. The resulting 33 was then oxidized
back to 32 following the Takeya protocol (DDQ, dioxane).[30b]
As the preparation of the proposed structure of 32 failed to
produce a compound that could be identified as the natural
product, as judged by different data, we propose that a struc-
tural reassignment of przewalskin E is necessary.
Synthesis of the proposed structure of dimeric C40 terpenes
obtusinones D and E—reassignment of the configuration of
obtusinone D
Having strong experimental evidence for the structure of syn-
thetic 32 (by reduction to known brussonol (33) and subse-
quent re-oxidation), the unusual dimeric icetexanes obtusin-
ones D (34) and E (35) were next synthesized (Scheme 9). The
first reported example of such an icetexane dimer was grandi-
one, which was isolated by Riccio and co-workers in 1999 from
Toreya grandis[34] and synthesized by Takeya and co-workers in
2005 (and the structure was reassigned)[35] by solid-state
Scheme 10. Mechanistic proposal for the formation of brussonol (33) und
przewalskin E (32).
&
&
Chem. Eur. J. 2016, 22, 1 – 9
www.chemeurj.org
6
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
hetero-Diels–Alder dimerization of o-quinone 31. Obtusino-
nes D (34) and E (35), which were recently isolated by Salae
and Boonnak, were suggested to originate biogenetically from
a similar Diels–Alder dimerization of przewalskin E (32).[36]
Having 32 in hand, modified Takeya conditions (neat, 1008C,
6 h) were applied to achieve the first synthesis of both obtusi-
nones D (34) and E (35) (1.7:1.0) in high overall yield (69%).
The separation of the two synthesized natural products proved
to be difficult by either flash chromatography or preparative
TLC on silica gel utilizing the eluent reported in the isolation
study.[36] We expected that p stacking of the two natural prod-
ucts could be responsible for the identical retention times.
Consequently, we decided to change our eluent system by
adding toluene to our solvent mixture to break the p-stacking
interaction between 34 and 35. Gratifyingly, the synthetic com-
pounds were then successfully separated by preparative TLC
after optimization of the eluent (pentane/toluene/Et2O, 5:5:1).
The spectral data of the two synthesized natural products
were in perfect agreement with the natural samples. However,
an X-ray analysis of a single crystal of obtusinone D (34,
Figure 3) was performed and the structure shows opposite
Figure 4. A) Originally proposed, and B) revised structures of obtusinones D
and E.
37, 1.7:1.0) could be explained by the higher temperature
needed to achieve the Diels–Alder reaction in high yield and
reasonable reaction time in the chemical laboratory.
Conclusions
A divergent, elegant, and short synthesis of 10 members of the
icetexane family of natural products has been achieved in four
to nine steps from abundant (+)-carnosic acid, isolated from
Rosmarinus officinalis. Of these icetexanes, (+)-salvicanol (12),
(ꢀ)-cyclocoulterone (1), (ꢀ)-coulterone (22), (ꢀ)-obtusinone D
(36), and (ꢀ)-obtusinone E (37) have been synthesized for the
first time. A straightforward, inexpensive, and efficient se-
quence, consisting of the methylation and hydrogenation of
rosemary ethanol extracts enriched with carnosic acid and re-
lated congeners, has been developed to obtain the desired tri-
methyl carnosic acid 8 in high yield. The salient features of the
synthesis of salvicanol (12) include a single-step biomimetic
ring expansion of the abietane 9 to the icetexane 11 bearing
an angular 10-hydroxy group. This is the first experimental evi-
dence for the trapping of a tertiary carbocation, such as 5, by
a water molecule and a selective aromatic demethylation.
(+)-Salvicanol (12) was used as a common intermediate to syn-
thesize 1) (ꢀ)-brussonol (33) and the reported structure of (ꢀ)-
przewalskin E (32) in two steps by an aromatic oxidation to
form the o-quinone 31, followed by an aerobic silica gel based
tautomerization and oxidation, 2) (ꢀ)-obtusinone D (36), which
was structurally reassigned, and (ꢀ)-obtusinone E (37) through
hetero-Diels–Alder reaction of (ꢀ)-przewalskin E (32), 3) (+)-ko-
maroviquinone (2) by a selective benzylic oxidation and an ar-
omatic oxidation, 4) (ꢀ)-cyclocoulterone (1) and (+)-komarovis-
pirone (3) by photolysis of (+)-komaroviquinone (2) in sun-
light, which supports our previous hypothesis for the non-en-
zymatic biogenetic formation of methylenedioxy catechol nat-
Figure 3. X-ray crystallographic structure of 34. Red=oxygen, grey=carbon,
white=hydrogen.
configurations at C13 and C14 to those reported by Salae and
Boonnak.[36] Indeed, they assigned the configurations at C13
and C14 of 34 on the basis of a NOESY correlation between H7
and H14 and between H14 and Me16. The close proximity of
H7 and H14 revealed by the X-ray crystallographic analysis of
34 (Figure 3) could explain the NOESY correlation measured by
Salae and Boonnak[36] between these two centers. This NOESY
correlation was later observed by us, therefore further substan-
tiating the identity of our synthetic obtusinone D as the natu-
ral product.
The X-ray crystallographic analysis of synthetic obtusinone D
in combination with the matching spectral data of both the
natural and synthetic samples therefore unambiguously re-
quires the structural reassignment of the natural product,
which should, as a consequence, be represented by the revised
structure 36 (Figure 4B). Consequently, the configurations of
C13 and C14 in obtusinone E are suggested to be identical to
those of obtusinone D, as shown for the revised structure 37
(Figure 4B).
The different ratio of the isolated natural products (36/37,
4:1) detected compared with that obtained by synthesis (36/
Chem. Eur. J. 2016, 22, 1 – 9
www.chemeurj.org
7
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
[14] a) M. Yatagai, T. Takahashi, Phytochemistry 1980, 19, 1149–1151; b) A.
Kelecom, Tetrahedron 1983, 39, 3603–3608; c) S. Hasegawa, Y. Hirose,
M. Yatagai, T. Takahashi, Chem. Lett. 1984, 13, 1837–1838; d) T. Kameta-
ni, H. Kondoh, M. Tsubuki, T. Honda, J. Chem. Soc. Trans. 1990, 5–10.
ural products, and finally, 5) (ꢀ)-coulterone (22) through the
hydrogenation of (+)-komaroviquinone (2).
´
[15] For a review on carnosic acid, see: S. Birtic, P. Dussort, F.-X. Pierre, A. C.
Experimental Section
Bily, M. Roller, Phytochemistry 2015, 115, 9–15.
[16] S. Munnꢅ-Bosch, L. Alegre, Planta 2000, 210, 925–931.
[17] S. Albu, E. Joyce, L. Paniwnyk, J. P. Lorimer, T. J. Mason, Ultrason. Sono-
chem. 2004, 11, 261–265.
For the details of all experiments, see the Supporting Information.
[18] C. Theoduloz, M. W. Pertino, J. A. Rodrꢄguez, G. Schmeda-Hirschmann,
Planta Med. 2011, 77, 882–887.
Acknowledgements
[19] a) N. Bai, K. He, M. Roller, C.-S. Lai, X. Shao, M.-H. Pan, C.-T. Ho, J. Agric.
Food Chem. 2010, 58, 5363–5367; b) E. H. A. Doolaege, K. Raes, K. Smet,
M. Andjelkovic, C. Van Poucke, S. De Smet, R. J. Verhꢅ, J. Agric. Food
Chem. 2007, 55, 7283–7287.
[20] G. Majetich, Y. Zhang, X. Tian, J. E. Briton, Y. Li, R. Phillips, Tetrahedron
2011, 67, 10129–10146.
We thank Prof. Dr. Christof Sparr for helpful discussions during
the preparation of the manuscript and Dr. D. Hꢂussinger and
Dr. H. Gsellinger (all University of Basel) for NMR analysis sup-
port. Part of this work was supported by the 2010 Novartis
Early Career Award (to K.G.).
[21] B. M. Fraga, A. G. Gonzꢃlez, J. R. Herrera, J. G. Luis, A. G. Ravelo, Phyto-
chemistry 1985, 25, 269–271.
[22] G. Majetich, G. Zou, Org. Lett. 2008, 10, 81–83.
Keywords: CꢀH activation · Diels–Alder reactions · natural
products · reaction mechanisms · ring expansion
[23] G. Majetich, Y. Zhang, K. Wheless, Tetrahedron Lett. 1994, 35, 8727–
8730.
[24] G. Majetich, Y. Li, G. Zou, Heterocycles 2007, 73, 217–225.
[25] K. Moriyama, M. Takemura, H. Togo, Org. Lett. 2012, 14, 2414–2417.
[26] T. Amaya, M. Hifumi, M. Okada, Y. Shimizu, T. Moriushi, K. Segawa, Y.
Ando, T. Hirao, J. Org. Chem. 2011, 76, 8049–8052.
[1] E. Breitmaier, Terpenes: Importance, General Structure, and Biosynthesis,
in Terpenes: Flavors, Fragrances, Pharmaca, Pheromones, Wiley-VCH,
Weinheim, 2006, pp. 1–9.
[2] For a review of cascade reactions, see: K. C. Nicolaou, D. J. Edmonds,
P. G. Bulger, Angew. Chem. Int. Ed. 2008, 47, 4217–4220; Angew. Chem.
2008, 120, 4285–4288.
[27] Y. Suto, K. Kaneko, N. Yamagiwa, G. Iwasaki, Tetrahedron Lett. 2010, 51,
6329–6330.
[28] a) G. Olivo, G. Arancio, L. Mandolini, O. Lanzalunga, S. Di Stephano,
Catal. Sci. Technol. 2014, 4, 2900–2903; b) M. T. Bhatt, P. T. Perumal, Tet-
rahedron Lett. 1981, 22, 2605–2608; c) S. Ghosh, U. R. Ghatak, Tetrahe-
dron 1992, 48, 7289–7296; d) N. A. Noureldin, D. Zhao, D. G. Lee, J. Org.
Chem. 1997, 62, 8767–8772.
[29] B. Frontana, J. Cꢃrdenas, L. Rodrꢄguez-Hahn, Phytochemistry 1994, 36,
739–741.
[30] a) Z.-Y. Jiang, Y.-J. Yu, C.-G. Huang, X.-Z. Huang, Q.-F. Hu, G.-Y. Yang, H.-B.
Wang, X.-Y. Zhang, G.-P. Li, Planta Med. 2015, 81, 241–246; b) Y. Aoyagi,
Y. Takahashi, H. Fukaya, K. Takeya, R. Aiyama, T. Matsuzaki, S. Hashimoto,
T. Kurihara, Chem. Pharm. Bull. 2006, 54, 1602–1604.
[31] a) A. Takuwa, K. Ryoji, Bull. Chem. Soc. Jpn. 1990, 63, 623–625; b) V. B.
Vol’eva, I. S. Belostotskaya, A. Y. Bundel’, N. L. Komissarova, V. V. Ershov,
Russ. Chem. Bull. 1997, 46, 385–386; c) V. B. Vol’eva, A. I. Prokof’ev, A.;
Karmilov, N. L. Komissarova, I. S. Belostotskaya, T. I. Prokof’eva, V. V.
Ershov, Russ. Chem. Bull. 1998, 47, 1920–1923; Karmilov, N. L. Komissar-
ova, I. S. Belostotskaya, T. I. Prokof’eva, V. V. Ershov, Russ. Chem. Bull.
1998, 47, 1920–1923.
[32] R. H. Burnell, M. Jean, S. Marceau, Can. J. Chem. 1987, 65, 227–230.
[33] G. Xu, L.-Y. Peng, L. Tu, X.-L. Li, Y. Zhao, P.-T. Zhang, Q. S. Zhao, Helv.
Chim. Acta 2009, 92, 409–413.
[34] B. Galli, F. Gasparrini, V. Lanzotti, D. Misiti, R. Riccio, C. Villani, H. Guan-
Fu, M. Zhong-Wu, Y. Wan-Fen, Tetrahedron 1999, 55, 11385–11394.
[35] Y. Aoyagi, Y. Takahashi, Y. Satake, H. Fukaya, K. Takeya, R. Aiyama, T. Mat-
suzaki, S. Hashimoto, T. Shiina, T. Kurihara, Tetrahedron Lett. 2005, 46,
7885–7887.
[3] For
a review of cationic rearrangements, see: R. R. Naredla, D. A.
Klumpp, Chem. Rev. 2013, 113, 6905–6948.
[4] For a review of CꢀH oxidations, see: T. Newhouse, P. S. Baran, Angew.
Chem. Int. Ed. 2011, 50, 3362–3374; Angew. Chem. 2011, 123, 3422–
3435.
[5] Review: R. G. Kerr, N. F. Miranda, J. Nat. Prod. 1995, 58, 1618–1621.
[6] a) W. Bauer, M. H. Zhenk, Phytochemistry 1991, 30, 2953–2961; b) E.
Ono, M. Nakai, Y. Fukui, N. Tomomori, M. Fukuchi-Mizutami, M. Saito, H.
Satake, T. Tanaka, M. Katzuka, T. Umezawa, Y. Tanaka, Proc. Natl. Acad.
Sci. USA 2006, 103, 10116–10121; c) N. Ikezawa, K. Iwasa, F. Sato, FEBS J.
2007, 274, 1019–1035.
[7] a) D. H. R. Barton, G. W. Kirby, J. B. Taylor, Proc. Chem. Soc. 1962, 341–
342; b) D. H. R. Barton, G. W. Kirby, J. B. Taylor, G. M. Thomas, J. Chem.
Soc. 1963, 4545–4558.
[8] C. Thommen, C. K. Jana, M. Neuburger, K. Gademann, Org. Lett. 2013,
15, 1390–1393.
[9] N. Uchiyama, F. Kiuchi, M. Ito, G. Honda, Y. Takeda, O. K. Khodzhimatov,
O. A. Ashurmetov, J. Nat. Prod. 2003, 66, 128–131.
[10] G. Majetich, Y. Jianhua, Org. Lett. 2008, 10, 89–91.
[11] Review: E. M. Simmons, R. Sarpong, Nat. Prod. Rep. 2009, 26, 1195–
1217.
[12] W. H. Watson, Z. Taira, X. A. Dominguez, H. Gonzales, M. Guiterrez, R.
Aragon, Tetrahedron Lett. 1976, 17, 2501–2502.
[13] a) C. Sꢃnchez, J. Cꢃrdenas, L. Rodrꢄguez-Hahn, T. P. Ramamoorthy, Phyto-
chemistry 1989, 28, 1681–1684; b) A. G. Gonzꢃlez, L. S. Andrꢅs, J. G.
Luis, I. Brito, M. L. Rodrꢄguez, Phytochemistry 1991, 30, 4067–4070;
c) B. M. Fraga, C. E. Dꢄaz, A. GuadaÇo, A. Gonzꢃlez-Coloma, J. Agric. Food
Chem. 2005, 53, 5200–5206; d) E. Simmons, J. R. Yen, R. Sarpong, Org.
Lett. 2007, 9, 2705–2707; e) Z.-Y. Ziang, C.-G. Huang, H.-B. Xiong, K.
Tian, W.-X- Liu, Q.-F. Hu, H.-B. Wang, G.-Y. Yang, X.-Z. Huang, Tetrahedron
Lett. 2013, 54, 3886–3888.
[36] A.-W. Salae, N. Boonnak, Tetrahedron Lett. 2013, 54, 1356–1359.
Received: August 18, 2016
Published online on && &&, 0000
&
&
Chem. Eur. J. 2016, 22, 1 – 9
www.chemeurj.org
8
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
FULL PAPER
Let the sun shine in! A collective syn-
thesis of ten icetexane natural products
has been completed based on several
biogenetic hypotheses. Of these icetex-
anes, (ꢀ)-cyclocoulterone was synthe-
sized from (+)-komariviquinone by
a redox-neutral CꢀH activation triggered
by sunlight (see figure).
&
Natural Products
C. Thommen, M. Neuburger,
K. Gademann*
&& – &&
Collective Syntheses of Icetexane
Natural Products Based on Biogenetic
Hypotheses
Chem. Eur. J. 2016, 22, 1 – 9
www.chemeurj.org
9
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ