Acid-catalyzed cyclization of terpenes under homogeneous and heterogeneous conditions as probed through stereoisotopic studies: A concerted process with competing preorganized chair and boat transition states

Article abstract of DOI:10.1002/chem.200901563

Based on stereoisotopic studies and β-secondary isotope effects, we propose that the acid-catalyzed cyclization of geranyl acetate proceeds through a concerted mechanism. Under heterogeneous conditions (zeolite Y confinement), a preorganized chairlike transition state predominates, whereas under homogeneous conditions the boat-and chairlike transition states are almost isoenergetic. For the case of farnesyl acetate, we propose that under homogeneous conditions a concerted dicyclization occurs with a preorganized boat-chair transition state competing with the chair-chair transition state. Under zeolite confinement conditions, the chair-chairlike dicyclization transition state is highly favorable. The preference of chairlike transition states within the cavities of zeolite Y is attributed to a transition state shape selectivity effect.

Full text of DOI:10.1002/chem.200901563

DOI: 10.1002/chem.200901563
Acid-Catalyzed Cyclization of Terpenes Under Homogeneous and
Heterogeneous Conditions as Probed Through Stereoisotopic Studies:
A Concerted Process with Competing Preorganized
Chair and Boat Transition States
Christos Raptis, Ioannis N. Lykakis, Constantinos Tsangarakis, and Manolis Stratakis*[a]
Abstract: Based on stereoisotopic stud-
ies and b-secondary isotope effects, we
propose that the acid-catalyzed cycliza-
tion of geranyl acetate proceeds
states are almost isoenergetic. For the
case of farnesyl acetate, we propose
that under homogeneous conditions a
concerted dicyclization occurs with a
preorganized boat–chair transition
state competing with the chair–chair
transition state. Under zeolite confine-
ment conditions, the chair–chairlike di-
cyclization transition state is highly fa-
vorable. The preference of chairlike
transition states within the cavities of
zeolite Y is attributed to a transition
state shape selectivity effect.
through
a
concerted mechanism.
Under heterogeneous conditions (zeo-
lite Y confinement), a preorganized
chairlike transition state predominates,
whereas under homogeneous condi-
tions the boat- and chairlike transition
Keywords: heterogeneous catalysis ·
homogeneous catalysis
effects reaction mechanisms
terpene cyclization
· isotope
·
·
Introduction
The remarkable product variety and stereocontrol in the
enzyme-catalyzed cyclization of polyene terpenes has fasci-
nated chemists for over half a century.[1] An array of poly-
cyclic natural products is derived from protonation of a suit-
able double bond proximal to an acidic amino acid residue
(initiation) and termination of the (poly)cyclization occurs
by a suitable basic amino acid residue within the enzyme
cavity. An impressive example is the cyclization of squalene
by squalene–hopene cyclase, which yields the pentacyclic
hopene and diplopterol via a hopanyl cation. Within a single
step, nine stereocenters are generated in a totally diastereo-
selective manner (Scheme 1).[2] The enzyme cavity provides
ideal stereochemical restrictions for the polyene to adopt
Scheme 1. The enzymatic cyclization of squalene to hopene and diplop-
terol.
[a] C. Raptis, Dr. I. N. Lykakis, Dr. C. Tsangarakis, Prof. Dr. M. Stratakis
Department of Chemistry, University of Crete
71003 Voutes, Iraklion (Greece)
the all-pre-chair conformation required for the stereospecif-
ic formation of the pentacyclic hopanyl skeleton.
Fax : (+30)2810-545001
E-mail: stratakis@chemistry.uoc.gr
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200901563. It contains NMR
spectra for the intermediate compounds in the synthesis of labeled
compounds, cyclization reactions, NOE experiments, and the GC
chromatographs recorded during the measurement of the b-secondary
isotope effects.
Numerous attempts to replicate this fascinating one-step
polycyclization in squalene by chemical means (acid cataly-
sis) have been unsuccessful. For smaller acyclic terpenes
(geranyl, farnesyl, geranylgeranyl), some efficient cyclization
protocols are known. The most acceptable methodology
11918
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 11918 – 11927

FULL PAPER
makes use of strong Brønsted acids,[3] such as halosulfonic
acids. Although the reaction is formally considered as cata-
lytic, the acid is used in multimolar equivalents relative to
the reacting substrate. Furthermore, the temperature has to
be maintained at À788C, or even lower. Formation of a six-
membered carbocyclic ring by the intramolecular nucleo-
philic attack of a C=C double bond to a carbocation is
highly exothermic and requires a very low activation energy.
An acyclic terpene, which is conformationally mobile, may
cyclize upon protonation under several diastereoselective
modes of low activation energy as well. Therefore, a variety
of isomeric products may be formed, unless the reaction
temperature is kept very low. Upon increasing the reaction
temperature, the product selectivity drops substantially,[4] as
various diastereoselective or other isomeric cyclization path-
ways also compete. Apart from diastereoselection, another
significant issue in the non-enzyme-catalyzed cyclization of
terpene polyenes is enantiocontrol. Yamamoto and co-work-
ers have significantly contributed to this field by using a
combination of a chiral Brønsted acid and a Lewis acid as a
catalyst.[5] Ishihara and co-workers achieved high enantio-
control in polyene cyclization promoted by a “chiral” elec-
trophilic halogen.[6] Moreover, Loh and Zhao reported an
asymmetric polyene cyclization methodology using a combi-
nation of a chiral acetal and a Lewis acid as initiating elec-
trophile.[7]
Recently, we examined the efficiency of acidic porous ma-
terials (zeolites)[8] as hosts and catalysts for the cyclization
of polyene terpenoids. Zeolites are mixed aluminosilicates,
which catalyze reactions by confining the substrates within
“active site” cavities.[9] Our concept postulated that the zeo-
lite cavity would provide the necessary acidic sites for ini-
tiating the cationic cyclization, whereas substrate confine-
ment would decrease its conformational mobility and pro-
vide stereocontrol. We chose faujasites (zeolites HY and
NaY) as possible catalysts. They possess a three-dimensional
framework that consists of supercages, the interior diameter
of which is approximately 13 ꢁ. The supercages are inter-
connected by four “windows” that are tetrahedrally distrib-
uted around each cage and have a pore entrance diameter
of around 7.5 ꢁ. The dimensions of the cavities are ideal to
host relatively big organic molecules, even steroids.[10] Zeo-
lite NaY was our first choice because it is known to contain
Brønsted and Lewis acidic sites.[11] The catalytic activity of
zeolite Y has been attributed to synergistic Brønsted and
Lewis acidity.[12] We were pleased to find that small acyclic
terpenes undergo a clean cyclization reaction under essen-
tially mild and environmentally friendly conditions
(Scheme 2).[13] Furthermore, NaY promotes the selective
monocyclization of epoxy polyene terpenoids,[14] as a unique
methodology for the synthesis of a series of incompletely cy-
clized naturally occurring terpenols.[15] For example, geranyl
acetate (1) forms at the initial reaction stages primarily g-cy-
clogeranyl acetate (1a), which on prolonged reaction time
isomerizes to the thermodynamically more stable a-cyclo-
geranyl acetate (1b). A remarkable result among those stud-
ies was the cyclization of geranylacetone (2)[13] and its deriv-
Scheme 2. Zeolite NaY promoted cyclization of small terpenoids.
atives,[16] which provide a direct one-pot cascade route to
the synthesis of a-ambrinol (2b); a valuable compound in
the fragrance industry. A unique tandem dicyclization path-
way was also uncovered for the case of farnesal adsorbed
within NaY.[17] The ability of NaY to promote terpene cycli-
zation was also reported immediately after our publication
by Yu and co-workers.[18]
Although the mechanism of epoxy terpene cyclization
under Brønsted,[19] Lewis acid[20] or zeolite[15] catalysis is well
established to proceed through a concerted pathway, the
mechanism of the acid-catalyzed cyclization of polyene ter-
penes is still not fully understood.[1d] The abundant stereo-
chemical studies of the past often contradict each other,[1c]
rather than merging upon accepting a concerted mechanism.
Yet, the nature of the transition states is still only hypothe-
sized. From a theoretical point of view, Rajamani and Gao
applied a combined quantum mechanics/molecular mechan-
ics (QM/MM) approach and presented discrete formation of
incompletely cyclized carbocationic intermediates en route
from squalene to hopenes.[21] On the other hand, DFT calcu-
lations by Hess and Smentek on a terpenoid fragment of
squalene supported a concerted mechanism for the A–B
ring formation.[22] Later on, Matsuda and co-workers report-
ed systematic errors in the DFT method as a tool to calcu-
late the energetics and transition states of polyene cycliza-
tion.[23] It is important to emphasize herein that all theoreti-
cal calculations deal with the cyclization of a tertiary carbo-
cation (resulting from protonation of the terminal double
bond), which would not exist as an intermediate in a con-
certed polycyclization process.
Prompted by our studies on the zeolite Y promoted cycli-
zation of terpenoids, we undertook a mechanistic investiga-
tion of the cyclization of geranyl and farnesyl acetate under
heterogeneous (zeolite NaY or HY confinement) and homo-
geneous (ClSO3H, solvent) conditions. These studies are
based on stereoisotopic labeling of the reacting substrates
and kinetic b-secondary isotope effects, and support a con-
certed mechanism for the monocyclization (geranyl acetate)
or dicyclization (farnesyl acetate) of small acyclic polyene
terpenes.
Chem. Eur. J. 2009, 15, 11918 – 11927
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
11919

M. Stratakis et al.
Results and Discussion
Cyclization of geranyl acetate: So far, we have shown that
in the cyclization of labeled 8,8,8-[D3]-geranyl acetate ([D3]-
1) a significant difference was found regarding the stereo-
chemical disposition of the gem-dimethyl group in the cy-
clized a- or g-cyclogeranyl acetate on going from homoge-
neous (ClSO3H, 2-nitropropane, À788C) to heterogeneous
(zeolite NaY) conditions (Scheme 3).[24] Under homogene-
Scheme 4. Synthesis of [D]-1.
The cyclization of [D]-1 under zeolite confinement condi-
tions[13] afforded a-cyclogeranyl acetate as a mixture of the
two diastereomers 11 and 12 in a relative ratio of 86:14
(Scheme 5). The reaction was almost exclusively driven to
Scheme 3. Stereochemical disposition of the gem-dimethyl group in the
cyclization of [D3]-1.
ous conditions the diastereomeric products 3 and 4 are
formed in almost equimolar amounts, with 3 slightly prevail-
ing, whereas under heterogeneous conditions 3 substantially
predominates over 4. Although we postulated that within
the zeolite cavities an initially formed tertiary carbocation
possibly undergoes cyclization at a faster rate than the rota-
tion around the terminal Me2C+ C bond, we also consid-
À
ered the possibility of cyclization through competing chair
and boat transition states. In addition, a question was
raised: is the cyclization mechanism stepwise (formation of
a tertiary carbocation that undergoes carbocyclization in a
second step), or concerted (single step)?
Scheme 5. Stereochemistry in the cyclization of [D]-1.
Thus, for a deeper mechanistic analysis of the cyclization
of 1 we performed further stereoisotopic studies and addi-
tionally measured b-secondary isotope effects. As a first
target we examined the stereochemistry in the cyclization of
6-[D]-geranyl acetate ([D]-1). The synthesis of [D]-1 was ac-
complished with >97% deuterium incorporation on C-6 by
following the procedure shown in Scheme 4. The known al-
a-cyclogeranyl acetate (708C, 3 h) to avoid the interfering
absorptions in the aliphatic region of the initially formed
isomeric g-cyclogeranyl acetate. The stereochemistry of the
major and minor products was assigned based on an NOE
spectroscopic analysis of a-cyclogeranyl acetate.[25] In con-
trast to the zeolite-promoted reaction, the cyclization of
[D]-1 in a homogeneous environment (ClSO3H, 2-nitropro-
pane at À78 or À258C) afforded a mixture of the diastereo-
mers 11 and 12 in 82% isolated yield with a very low degree
of diastereoselection (diastereomeric ratio (dr)ꢀ6–8%),
without thus observing a temperature dependence on the
product ratio. Identical results were found by performing
the ClSO3H-catalyzed reaction in the less polar CH2Cl2
(À788C). The results presented in Scheme 5 resemble sub-
stantially the stereochemical outcome in the cyclization of
[D3]-1[24] (Scheme 3).
dehyde
5 produced by the oxidative cleavage of the
TBDPS-protected 6,7-epoxygeraniol[24] (TBDPS=tert-butyl-
diphenylsilyl) was oxidized with NaClO2 to the correspond-
ing carboxylic acid 6. The acid 6 was reduced with LiAlD4
to form the labeled [D2]-alcohol 7, which was oxidized with
pyridinium chlorochromate (PCC) to yield the deuterated
aldehyde 8. The aldehyde was coupled with isopropylidene-
triphenylphosphorane to produce the TBDPS-protected
alkene 9. Finally, alkene 9 was deprotected with tetrabutyl-
ammonium fluoride (TBAF) and the resulting alcohol 10 (6-
[D]-geraniol) was acetylated with acetic anhydride to form
[D]-1, in 14% overall isolated yield from geraniol.
A further mechanistic approach regarding the cyclization
of geranyl acetate was performed by measuring the intermo-
lecular kinetic b-secondary isotope effects upon competition
11920
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 11918 – 11927

Terpene Cyclization
FULL PAPER
of geranyl acetate ([D0]-1) with the labeled 8,8,8,9,9,9-[D6]-
geranyl acetate ([D6]-1).[26] An isotope effect of kH/kD =
1.30Æ0.02 was measured under homogeneous conditions
(ClSO3H, 2-nitropropane, À708C, 20–40 s), with a-cycloger-
anyl acetate being formed exclusively. Under heterogeneous
conditions (zeolite NaY, 208C, 3 min) a mixture of a-cyclo-
geranyl acetate (1a), g-cyclogeranyl acetate (1b), and a
minor amount of alcohol 1c were formed with kH/kD =
1.33Æ0.03 (Scheme 6). The isotope effects were measured
in a relative ratio of approximately 1:3. The competing cycli-
zation of [D0]-14 and [D6]-14 (1 min, ꢀ15–20% conversion)
provided an isotope effect of kH/kD =1.21Æ0.02 (Scheme 6).
Again, the isotope effect was measured by GC. The relative-
ly similar values of b-secondary isotope effects clearly indi-
cate a similarity between the transition states of terpene and
epoxy terpene carbocyclization.
Before analyzing and drawing conclusions from the results
of the above stereoisotopic studies, we emphasize that the
cyclization of [D3]-1 proceeds
either under zeolite confine-
ment[24] or under homogeneous
conditions without detectable
isomerization of the noncycl-
ized starting material. This im-
plies that the low diastereose-
lection under homogeneous
conditions (products 3 and 4),
or the higher diastereoselection
under heterogeneous conditions
(products 11 and 12) does not
result from any isomerization
on the terminal double bond,
C6=C7, of [D3]-1 prior to cycli-
zation. Based on the current re-
sults shown in Schemes 3, 5,
and 6, we propose that the
acid-catalyzed cyclization of
geranyl
through
acetate
concerted mecha-
proceeds
a
nism, regardless of the reaction
medium (homogeneous or het-
erogeneous in nature), via the
two competing chair- (TSchair
)
and boatlike (TSboat) transition
states shown in Scheme 7.
Essentially, for the reaction
to occur, the substrate requires
a conformational preorganiza-
Scheme 6. Intramolecular kinetic b-secondary isotope effects for the cyclization of geranyl acetate and 6,7-
epoxy geraniol.
by GC analysis of the reaction mixture at approximately 10–
40% conversion, as both competing [D0]-1 and [D6]-1, as
well as their cyclization products are clearly separated on a
60 m capillary column (see the Supporting Information).
This magnitude of isotope effect implies a partial positive
charge developing on the tertiary carbon atom C-7 in the
Scheme 7.
transition state of the rate-determining step of the cycliza-
tion. We decided to compare the value of kH/kD ꢀ1.3 to the
kinetic b-secondary isotope effect during the FeCl3-cata-
lyzed[27] cyclization of 6,7-epoxy geraniol ([D0]-14) and its la-
beled counterpart [D6]-14[26] (Scheme 6) because the acid-
catalyzed cyclization of epoxy polyene terpenes is a well-es-
tablished concerted process. In the presence of FeCl3
(1.0 equiv, CH2Cl2, 208C), 6,7-epoxygeraniol yields a mix-
ture of the carbocyclization product 15[14] and the bicyclic
ether 16,[14] through a common carbocationic intermediate,
tion,[28] so that the proximity of H+ to the C6=C7 double
bond and the suitable chair or boat conformation that
brings the two double bonds at a proximal orientation
allows the cyclization to take place in a concerted fashion.
This is clearly derived from the stereochemical outcome of
labeled [D3]-1 and [D]-1, for which the product distribution
is rationalized to arise through concerted chair- and boatlike
transition states (Scheme 8). Regarding the reaction in a ho-
mogeneous environment, we propose that the energy differ-
Chem. Eur. J. 2009, 15, 11918 – 11927
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
11921

M. Stratakis et al.
the three destabilizing non-
bonding steric interactions ap-
pearing in TSchair are absent,
with the only steric strain
caused by the interaction of the
CH2OAc group with the gem-
methyl group at C-8. The tor-
sional strain between the
À
eclipsed C H bonds within
TSboat is balanced by the signifi-
cant release of the steric strain
present in TSchair. Therefore, we
conclude that it is reasonable to
assume that the TSchair and
TSboat transition states are ex-
pected to be close in energy. In
addition, the magnitude of the
b-secondary isotope effect (kH/
kD ꢀ1.3, or kH/kD ꢀ1.04 per H/
D atom) is consistent with a
partial positive charge develop-
ing on C-7 in the transition
state of the reaction. A full pos-
itive charge, considering an in-
termediate acyclic tertiary car-
bocation on C-7, would be ex-
pected to provide a value of kH/
kD ꢀ1.8[31] (kH/kD ꢀ1.10 per H/
D atom). Our measured value
of kH/kD ꢀ1.3 is close to the b-
secondary isotope effect of kH/
kD =1.37[32] found during the
solvolysis of 7-chloro-3,7-dime-
thyloct-2-ene, in which the p
participation of the double
bond in a six-membered-ring
transition state analogous to the
proposed transition states of
Scheme 8. Proposed transition states for the acid-catalyzed cyclization of geranyl acetate through the stereoiso-
topic studies of [D3]-1, [D]-1 and the elucidation of the intramolecular b-secondary isotope effects.
geranyl
acetate
cyclization
ence between the concerted chair- (TSchair) and boatlike
(TSboat) transition states is very small, thus the product dia-
stereoselectivity (3 versus 4, or 11 versus 12), is very low.
On the other hand, for the acid-catalyzed intrazeolite cycli-
zation,[29] we postulate that the confined zeolite environment
can better accommodate within its cavities the chairlike
transition state than the boatlike one, thus 3 predominates
over 4, and 11 over 12. From the first point of view, a chair-
like transition state is expected to be significantly more
stable than a boatlike transition state because the chair con-
formation of cyclohexane is more stable relative to a boat
or twist-boat by 6 and 4.5 kcalmolÀ1, respectively. Yet, a
closer examination of the above shown transitions states re-
veals that the TSchair is destabilized by the steric strain be-
tween the methyl group at C-3 and gem-methyl of C-9. In
addition, the CH2OAc substituent develops steric strain with
both gem-methyl groups at C-8 and C-9, because it is in a
gauche conformation relative to them. In TSboat, two out of
shown in Scheme 8, had been postulated. Apart of the evi-
dence provided by the b-secondary isotope effects, a step-
wise mechanism with formation of a tertiary acyclic carbo-
cation as an intermediate can be undoubtedly ruled out: Al-
though the diastereoselective disposition of the gem-methyl
groups in [D3]-1 might be attributed to a slow rotation of
À
the C6 C7 bond in the tertiary carbocation prior to cycliza-
tion, for the case of [D]-1, no diastereoselection (11 versus
12) would be expected to result after protonation on C-6,
and subsequent cyclization. Based on these results, we em-
phasize that in considering terpene cyclization, a fully con-
certed carbocyclization process has to be taken into account
and not a tertiary acyclic intermediate carbocation as the
starting point of polycyclization. We consider the tempera-
ture independence of the stereochemical results for the cyc-
lization of [D]-1 and [D3]-1 as reasonable, because there is
not any obvious difference in the entropy of activation be-
tween a chair- and a boatlike transition state.
11922
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 11918 – 11927

Terpene Cyclization
FULL PAPER
The preference of the TSchair within the cavities of NaY is
attributed to a confinement effect. The absorbing efficiency
of porous materials, such as zeolites, is well known to
depend significantly on the steric demands of the guest mol-
ecules. For example, the relative ratio of the conformers of
cyclohexane and its derivatives in solution may significantly
change upon absorption in the constrained environment of
zeolites.[33] In contrast to silicalite-1, zeolite Y accommodates
cyclohexane exclusively in a chair conformation.[34] There-
fore, we postulate that TSchair is energetically more accom-
modated within the zeolite supercage than TSboat, due a
shape selective confinement effect (transition-state shape se-
lectivity[9a]).
porting Information). The cyclization of [D3]-17 was studied
under homo- and heterogeneous conditions. Under homoge-
neous conditions (ClSO3H (5 equiv), 2-nitropropane,
À608C, 5 min) mainly the labeled drimanediol monoacetate
was formed. The disposition of the gem-dimethyl group was
not stereoselective, with formation of the diastereomeric 20
and 21 in a relative ratio 20/21 of approximately 55:45
(Scheme 10). The stereochemical assignment of the products
Cyclization of farnesyl acetate: Continuing our studies, the
mechanism of bicyclization of farnesyl acetate (17) was tar-
geted. Under superacidic conditions,[35] farnesyl acetate
forms drimenyl acetate (18) and/or drimanediol monoace-
tate 19 depending on the amount of water present in the re-
action mixture (Scheme 9). In our hands, by using ClSO3H
Scheme 9. The acid-catalyzed cyclization of farnesyl acetate.
Scheme 10. Stereochemistry in the cyclization of farnesyl acetate [D3]-17
and the b-secondary isotope effects for the competition of [D0]-17 with
[D6]-17.
as the acidic catalyst (5 equiv), drimanediol monoacetate
was mainly formed. The product selectivity depends on the
reaction temperature. At À758C, 19 is primarily formed,
whereas by increasing the reaction temperature to À508C,
two additional bicyclic diastereomers are formed (GC–MS)
in up to 30% relative yield, in agreement to the tempera-
ture-dependent product selectivity observed in the ClSO3H-
catalyzed cyclization of geranyl derivatives.[4]
As indicated earlier, the theoretical calculations on the
mechanism of the dicyclization contradict each other. A
combined QM/MM approach[21] proposes discrete formation
of a monocyclized carbocationic intermediate that subse-
quently leads to the bicyclic product, whereas DFT calcula-
tions support[22] a concerted mechanism for A–B ring forma-
tion. Both studies, however, consider a protonated acyclic
tertiary carbocation as the reacting species. By analogy to
the stereoisotopic studies performed in geranyl acetate, we
studied the mechanism of farnesyl acetate cyclization by
preparing stereoselectively 12,12,12-[D3]-farnesyl acetate
([D3]-17),[20] as well as 12,12,12,13,13,13-[D6]-farnesyl acetate
([D6]-17), following similar synthetic approaches to the syn-
thesis of the corresponding geranyl analogues (see the Sup-
was achieved by an NOE spectroscopic analysis.[36] This
result remarkably resembles the stereochemical outcome in
the ClSO3H-promoted cyclization of the geranyl analogue
[D3]-1. The cyclization of [D3]-17 under zeolite confinement
conditions was achieved by using the highly dealuminated
zeolite HY (Si/Al=30:1) as a catalyst at ambient tempera-
ture. This relatively mildly acidic zeolite, in contrast to
NaY[13] and HY (Si/Al=2.7:1), promotes the cyclization of
farnesyl to drimenyl acetate in good yield, without forming
hydrocarbon byproducts (resulting from an acid-catalyzed
elimination of the acetate in the starting material).[37] In the
presence of dealuminated HY, [D3]-17 afforded a mixture of
22/23 in 72% isolated yield and in a relative ratio of approx-
imately 92:8 (Scheme 10). Thus, in contrast to the ClSO3H-
catalyzed cyclization, a remarkable stereoselectivity in the
disposition of the gem-methyl group was observed. Further-
more, the kinetic competition of farnesyl acetate ([D0]-17)
with its gem-dimethyl deuterated analogue, [D6]-17, was per-
formed. Under homogeneous conditions the relative reac-
tion rate [D0]-17/[D6]-17 (kH/kD) was 1.11Æ0.02, compared
Chem. Eur. J. 2009, 15, 11918 – 11927
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
11923

M. Stratakis et al.
with kH/kD =1.10Æ0.02 under heterogeneous conditions
(Scheme 10). Under homogeneous conditions (ClSO3H,
À508C), apart from drimanediol monoacetate ([D0]-19 and
[D6]-19), two other bicyclic epimers were formed (see the
Supporting Information) and the isotope effect was mea-
sured after 30 s of reaction when approximately a 30% con-
version had been achieved (the reaction was quenched by
adding an excess of Et3N to the reaction mixture). Under
HY treatment, approximately 30% conversion was achieved
after 30 s (the reaction was stopped by adding moistened
methanol into the heterogeneous slurry). Apart from dri-
menyl acetate (18), its exocyclic isomer albicanyl acetate
(24) was also detected, by comparison with an authentic
sample. It is notable that after a few minutes, the reaction
had gone to completion and albicanyl acetate had quantita-
tively isomerized to the thermodynamically more stable dri-
menyl acetate. Both isotope effects were measured by GC
(see the Supporting Information for the details).
To analyze the results of the stereoisotopic experiments,
we began with the intrazeolite experiments. The gem-di-
methyl group disposition in [D3]-17 and the intramolecular
isotope effect of [D0]-17/[D6]-17ꢀ1.10 support a concerted
mechanism for the formation of 22 via a chair–chair preor-
ganized transition state (Scheme 11). The very low magni-
tude of the b-secondary isotope effect indicates an extensive
delocalization of the positive charge along the transition-
state framework and is in agreement with the close to unity
b-secondary isotope effects obtained in the solvolysis of 2-
chloro-2,3-dihydrosqualene[31] and other polyene terpenyl-
type chlorides.[32] The minor 23 could derive from a boat–
chair preorganized concerted transition state. Regarding the
cyclization in solution, the low magnitude of the b-secon-
dary isotope effect supports a concerted bicyclization mech-
anism, as proposed under heterogeneous conditions, with
competition between the boat–chair and chair–chair transi-
tion states. A boat–chair transition state is not less thermo-
dynamically stable, for exactly the same reasons as put for-
ward earlier for the transition states of the cyclization of
geranyl acetate. The enzymatic cyclization of polyene ter-
penes through mixed chair- and boatlike transition states is
known.[38]
Finally, further support for the concerted nature of the bi-
cyclization of farnesyl acetate under zeolite catalysis was ob-
tained by studying the gem-dimethyl group disposition in
the zeolite NaY-promoted cyclization of 12,12,12-[D3]-far-
nesal ([D3]-25). Farnesal (25) forms primarily the bicyclic al-
dehyde 27 (major diastereomer shown in Scheme 12)
through a stepwise mechanism involving preferential forma-
tion of exomethylenic monocyclic 26a, which subsequently
leads to 27 through a Prins-type cyclization.[17] The stereose-
lectively labeled farnesal [D3]-25 was prepared in two steps
from 11,11,11-[D]-(E)-6,10-dimethylundeca-5,9-dien-2-one
(see the Supporting Information), which is readily available
from our previous studies as a mixture of E and Z isomers
at the enal moiety.[15] Treatment with NaY afforded the bicy-
clic aldehydes 27a and 27b in a relative ratio of 70:30
(Scheme 12). Due to the structural similarity of farnesyl ace-
tate and farnesal, we indicate that a hypothetical monocycli-
zation of farnesyl acetate followed by a second distinct di-
cyclization step should provide a ratio of 22/23 analogous to
that of 27a/27b (ꢀ70:30), and not the observed 22/23>
90:10. The relative ratio of 27a/27bꢀ70:30 is in close agree-
ment to the ratio of 3/4 obtained in the intrazeolite (mono)-
cyclization of [D3]-1 and is explained in terms of the chair-
and boatlike transition states shown in Scheme 12.
Conclusion
We have presented substantial evidence that the acid-cata-
lyzed cyclization of geranyl and farnesyl acetate proceeds
through a concerted mechanism. The cyclization occurs
through competing chair- and boatlike transition states and
a suitable conformation (preorganization) of the terpene is
necessary in order for the reaction to take place. The nature
of the transition states depends substantially on the reaction
medium, with the confinement conditions favoring the chair-
like transitions states. We believe that these studies will
challenge theoreticians to study the mechanism of terpene
cyclization considering a fully concerted process, and not an
elusive acyclic tertiary carbocation as the starting point of
(poly)cyclization.
Scheme 11. Mechanistic proposal for the acid-catalyzed cyclization of far-
nesyl acetate.
11924
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 11918 – 11927

Terpene Cyclization
FULL PAPER
13.5 mmol) in dry ether (20 mL) at
0 8C under an inert atmosphere. After
6 h the reaction mixture was quenched
by adding water (0.5 mL) and the re-
sulting slurry was left for an additional
30 min until precipitation of the inor-
ganic salts was complete. The superna-
tant solution was dried and the solvent
was removed to afford the crude alco-
hol, which was purified by column
chromatography (hexane/ethyl acetate
12:1) to produce
7 (3.04 g, 73%).
1H NMR (300 MHz, CDCl3): d=7.70
(d, J=7.0 Hz, 4H), 7.38–7.42 (m, 6H),
5.41 (brt, J=6.0 Hz, 1H), 4.22 (d, J=
6.0 Hz), 2.05 (t, J=7.0 Hz, 2H), 1.65
(t, J=7.0 Hz, 2H), 1.46 (s, 3H),
1.05 ppm (s, 9H); 13C NMR (75 MHz,
CDCl3): d=136.8, 135.6, 134.0, 129.5,
127.6, 124.4, 61.8 (quintet,
JACHTUNGTRENNUNG(C,D)=
22.0 Hz), 61.0, 35.7, 30.3, 26.8, 19.2,
16.2 ppm.
1-[D]-(E)-6-(tert-Butyldiphenylsily-
loxy)-4-methylhex-4-enal (8): The deu-
terated alcohol
7 (1.2 g, 3.18 mmol)
was oxidized with PCC (0.8 g,
3.82 mmol) in dry dichloromethane
(10 mL). The reaction was complete
after 4 h (monitored by TLC). The sol-
vent was concentrated under vacuum,
and the residue was purified by
column chromatography (hexane/ethyl
Scheme 12. Disposition of the gem-dimethyl group in the dicyclization of [D3]farnesal ([D3]-25) promoted by
zeolite NaY.
acetate 20:1) to afford
8 (0.89 g,
76%). 1H NMR (300 MHz, CDCl3):
d=7.67 (d, J=6.5 Hz, 4H), 7.38–7.44
(m, 6H), 5.38 (t, J=6.5 Hz, 1H), 4.22
(d, J=6.5 Hz, 2H), 2.50 (t, J=7.0 Hz,
2H), 2.30 (t, J=7.0 Hz, 2H), 1.45 (s,
Experimental Section
3H), 1.04 ppm (s, 9H); 13C NMR (75 MHz, CDCl3): d=135.6, 134.9,
133.9, 129.5, 127.6, 125.0, 124.8, 61.0, 41.7, 31.5, 26.8, 19.2, 16.4 ppm.
General: The reagents and solvents were purchased at the highest avail-
able commercial quality and used without purification. The reactions
were monitored by thin-layer chromatography (TLC) carried out on
silica gel plates (60F-254) ith UV light as the visualizing method and an
acidic mixture of phosphomolybdic acid/cerium(IV) sulfate accompanied
by heating of the plate as a developing system. Flash column chromatog-
raphy was carried out on SiO2 (silica gel 60, particle size 0.040–
0.063 mm) with the eluent specified below. NMR spectra were recorded
on a Bruker DPX-300 instrument. Electrospray ionisation mass spec-
trometry (ES-MS) experiments were performed on a GC-MS QP 5050
Shimadzu single-quadrupole mass spectrometer. GC analyses and kinet-
ics were performed using a Shimadzu GC-17A model equipped with a
60 m HP-5 capillary column.
6-[D]-(E)-tert-Butyl(3,7-dimethylocta-2,6-dienyloxy)diphenylsilane (9):
nBuLi (1.6m in hexane) was added to a slurry of the triphenylphosphoni-
um salt of 2-iodopropane (1.5 g, 3.47 mmol) in dry THF (10 mL) at 08C
under an inert atmosphere. After 40 min, aldehyde 8 (0.85 g, 2.3 mmol)
was added and the reaction was complete after 10 min. The solvent was
removed under vacuum and the resulting solids were washed with
hexane (4ꢂ10 mL). The combined solvents were concentrated under
vacuum and the residue was purified by column chromatography
(hexane/ethyl acetate 80:1) to afford
9
(0.65 g, 71%). 1H NMR
(300 MHz, CDCl3): d=7.71 (d, J=7.0 Hz, 4H), 7.37–7.42 (m, 6H), 5.39
(brt, J=6.0 Hz, 1H), 5.24 (d, J=6.0 Hz, 2H), 2.07 (t, J=7.5 Hz, 2H),
1.99 (t, J=7.5 Hz, 2H), 1.69 (s, 3H), 1.62 (s, 3H), 1.55 (s, 3H), 1.06 ppm
(s, 9H); 13C NMR (75 MHz, CDCl3): d=137.0, 135.7, 134.2, 131.3, 129.5,
127.6, 124.1, 61.2, 39.5, 26.9, 26.3, 25.7, 19.2, 17.7, 16.3 ppm.
(E)-6-(tert-Butyldiphenylsilyloxy)-4-methylhex-4-enoic acid (6): A solu-
tion of NaClO2 (13.6 g, 150.3 mmol) and NaH2PO4 (17.9 g, 114.6 mmol)
in water (90 mL) was added over 20 min to a stirred solution of 5[24]
(6.0 g, 16.4 mmol) and 2-methyl-2-butene (45 mL) in tert-butanol
(240 mL). The mixture was stirred at ambient temperature for 12 h and
then diluted with water (200 mL). The resulting solution was extracted
with hexane (2ꢂ150 mL) and the organic layer was dried. The crude acid
isolated after evaporation of the solvents was used in the next step with-
out chromatographic purification (4.2 g, 67%). 1H NMR (300 MHz,
CDCl3): d=7.70 (d, J=7.0 Hz, 4H), 7.37–7.43 (m, 6H), 5.42 (brt, J=
5.5 Hz, 1H), 4.23 (d, J=5.5 Hz, 2H), 2.44 (t, J=7.0 Hz, 2H), 2.30 (t, J=
7.0 Hz, 2H), 1.45 (s, 3H), 1.05 ppm (s, 9H); 13C NMR (75 MHz, CDCl3):
d=179.1, 135.6, 134.8, 134.0, 129.6, 127.6, 124.9, 61.0, 33.9, 32.4, 26.8,
19.2, 16.3 ppm.
6-[D]-Geraniol (10): A solution of TBAF (1m in THF, 1.7 mL, 1.7 mmol)
was added dropwise to a solution of 9 (0.62 g, 1.57 mmol) in dry THF
(15 mL). After 3 h the solution was diluted with diethyl ether, washed
with water, and the organic layer was dried. The residue was purified by
column chromatography (hexane/ethyl acetate 4:1) to afford 10 (0.24 g,
93%). 1H NMR (300 MHz, CDCl3): d=5.40 (t, J=6.5 Hz, 1H), 4.14 (d,
J=7.0 Hz, 2H), 2.09 (t, J=7.0 Hz, 2H), 2.01 (t, J=7.0 Hz, 2H), 1.67 (s,
6H), 1.59 ppm (s, 3H); 13C NMR (75 MHz, CDCl3): d=139.6, 131.6,
123.5 (t, JACTHNUTRGNE(NUG C,D)=23.0 Hz), 123.3, 59.4, 39.5, 26.3, 25.6, 17.6, 16.2 ppm.
6-[D]-Geranyl acetate ([D]-1):K2CO3 (0.30 g, 2.22 mmol), acetic anhy-
dride (0.21 mL, 2.22 mmol), and dimethylaminopyridine (0.02 g) were
added to a solution of 10 (0.23 g, 1.48 mmol) in ethyl acetate (5 mL).
After stirring at 258C for 1 h, the reaction mixture was filtered and the
filtrate was diluted with diethyl ether. The organic layer was washed with
2,2-[D2]-(E)-6-(tert-Butyldiphenylsilyloxy)-4-methylhex-4-en-1-ol
(7):
Acid 6 (4.3 g, 11.2 mmol) was added to a slurry of LiAlD4 (0.57 g,
Chem. Eur. J. 2009, 15, 11918 – 11927
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
11925

M. Stratakis et al.
a saturated aqueous solution of NaHCO3 and dried over MgSO4. Remov-
al of the solvent afforded [D]-1 (275 mg, 98%). 1H NMR (300 MHz,
CDCl3): d=5.34 (t, J=7.0 Hz, 1H), 4.58 (d, J=7.0 Hz, 2H), 2.08 (m,
4H), 2.04 (s, 3H), 1.70 (s, 3H), 1.67 (s, 3H), 1.59 (s, 3H); 13C NMR
Schulte-Elte, J. Org. Chem. 1992, 57, 955–960. Irradiation of the
more upfield absorption (d=1.47 ppm) among the diastereotopic Ha
and Hb of a-cyclogeranyl acetate (1b), results in a significant signal
enhancement to one of the two diastereotopic CH2OAc protons res-
onating at d=4.16 ppm. No signal enhancement of the CH2OAc
protons was found upon irradiation of the more downfield absorp-
tion resonating at d=1.20 ppm, in accordance with the 1H NMR
spectral assignments shown below:
(75 MHz, CDCl3): d=171.1, 142.3, 131.8, 123.7 (t, JACHTUNTRGNEU(GN C,D)=23.0 Hz),
118.2, 61.4, 39.5, 26.3, 21.1, 17.6, 16.5; MS (EI): m/z (%): 155
[MÀAcOH]+ (2), 137 (11), 122 (12), 108 (7), 70 (100).
Acknowledgements
This project was partially co-funded by the European Social Fund and
national resources (program PYTHAGORAS-II).
[1] a) D. W. Christianson, Chem. Rev. 2006, 106, 3412–3442; b) A. Es-
chenmoser, D. Arigoni, Helv. Chim. Acta 2005, 88, 3011–3050;
c) R. A. Yoder, J. N. Johnston, Chem. Rev. 2005, 105, 4730–4756;
d) K. U. Wendt, G. E. Schulz, E. J. Corey, D. R. Liu, Angew. Chem.
2000, 112, 2930–2952; Angew. Chem. Int. Ed. 2000, 39, 2812–2833;
e) I. Abe, M. Rohmer, G. D. Prestwich, Chem. Rev. 1993, 93, 2189–
2206.
[2] a) T. Hoshino, T. Sato, Chem. Commun. 2002, 291–301; b) I. Abe,
Nat. Prod. Rep. 2007, 24, 2812–2833.
[3] P. F. Vlad, Pure Appl. Chem. 1993, 65, 1329–1336.
[4] P. J. Linares-Palomino, S. Salido, J. Altarejos, A. Sanchez, Tetrahe-
dron Lett. 2003, 44, 6651–6655.
[5] a) K. Ishihara, H. Ishibashi, H. Yamamoto, J. Am. Chem. Soc. 2001,
123, 1505–1506; b) H. Yamamoto, K. Futatsugi, Angew. Chem.
2005, 117, 1958–1977; Angew. Chem. Int. Ed. 2005, 44, 1924–1942.
[6] A. Sakakura, A. Ukai, K. Ishihara, Nature 2007, 445, 900–903.
[7] Y.-J. Zhao, T.-P. Loh, J. Am. Chem. Soc. 2008, 130, 10024–10029.
[8] A. Dyer, An Introduction to Zeolite Molecular Sieves, Wiley, New
York, 1988.
[9] a) S. E. Sen, S. M. Smith, K. A. Sullivan, Tetrahedron 1999, 55,
12657–12698; b) N. J. Turro, Acc. Chem. Res. 2000, 33, 637–646;
c) V. Ramamurthy, P. Lakshminarasimhan, C. P. Grey, L. J. Johnston,
Chem. Commun. 1998, 2411–2418.
[10] V. J. Rao, S. R. Uppili, D. R. Corbin, S. Schwarz, S. R. Lustig, V.
Ramamurthy, J. Am. Chem. Soc. 1998, 120, 2480–2481.
[11] V. J. Rao, D. L. Perlstein, R. J. Robbins, P. H. Lakshminarasimhan,
H.-M. Kao, C. P. Grey, V. Ramamurthy, Chem. Commun. 1998, 269–
270.
[26] C. Raptis, H. Garcia, M. Stratakis, Angew. Chem. 2009, 121, 3179–
3182; Angew. Chem. Int. Ed. 2009, 48, 3133–3166.
[27] S. E. Sen, S. L. Roach, S. M. Smith, Y. Z. Zhang, Tetrahedron Lett.
1998, 39, 3969–3972.
[28] Y. J. Hong, D. J. Tantillo, J. Am. Chem. Soc. 2009, 131, 7999–8015.
[29] To test whether Brønsted acidity is responsible for the catalytic ac-
tivity of NaY in terpene cyclization, the zeolite (1 g) was treated for
30 min with D2O (2ꢂ5 mL). Then it was filtered and dried at 1208C
under vacuum, in the usual way zeolite NaY is dried/activated to
perform cyclization experiments. This isotopically modified zeo-
lite,[30] which we refer to as “NaDY”, catalyzed the cyclization of
geranyl acetate (1) to form after 2 h at 708C a mixture of a- and g-
cyclogeranyl acetate in a ratio ꢀ15:1. The loading level of 1@NaDY
1
was kept low (nꢀ0.1–0.2). Examination of the products by H NMR
spectroscopy revealed an approximately 30–35% incorporation of
deuterium atom on the C-5 of the a-cyclogeranyl framework next to
the gem-dimethyl group. Furthermore, a partial exchange of the hy-
drogen atoms from the allylic methyl group and the vinylic hydro-
gen atom on C-3 was obvious. Thus, in addition to the non-deuteri-
um-exchanged vinylic methyl group (CH3) absorptions due to
(CH2D) and (CHD2) were seen, and most likely a completely ex-
changed (CD3) might be formed to a minor extent. The partial deu-
[12] C. J. A. Mota, D. L. Bhering, N. Rosenbach, Jr., Angew. Chem. 2004,
116, 3112–3115; Angew. Chem. Int. Ed. 2004, 43, 3050–3053.
[13] C. Tsangarakis, M. Stratakis, Adv. Synth. Catal. 2005, 347, 1280–
1284.
[14] C. Tsangarakis, E. Arkoudis, C. Raptis, M. Stratakis, Org. Lett. 2007,
9, 583–586.
[15] C. Tsangarakis, C. Raptis, E. Arkoudis, M. Stratakis, Adv. Synth.
Catal. 2008, 350, 1587–1600.
[16] C. Efe, C. Tsangarakis, I. N. Lykakis, C. Raptis, M. Stratakis, Synlett
2008, 1635–1638.
[17] C. Tsangarakis, I. N. Lykakis, M. Stratakis, J. Org. Chem. 2008, 73,
2905–2908.
[18] W. Yu, F. Bian, Y. Gao, L. Yang, Z.-L. Liu, Adv. Synth. Catal. 2006,
348, 59–62.
[19] E. J. Corey, D. D. Staas, J. Am. Chem. Soc. 1998, 120, 3526–3527.
[20] E. E. van Tamelen, D. R. James, J. Am. Chem. Soc. 1977, 99, 950–
952.
[21] R. Rajamani, J. Gao, J. Am. Chem. Soc. 2003, 125, 12768–12781.
[22] B. A. Hess, Jr., L. Smentek, Org. Lett. 2004, 6, 1717–1720.
[23] S. P. T. Matsuda, W. K. Wilson, Q. Xiong, Org. Biomol. Chem. 2006,
4, 530–543.
[24] C. Tsangarakis, M. Stratakis, Eur. J. Org. Chem. 2006, 4435–4439.
[25] The a-cyclogeranyl derivatives adopt a preferred conformation in
which the CH2OAc group possesses a pseudoaxial orientation: R. L.
Snowden, J.-C. Eichenberger, S. M. Linder, P. Sonnay, C. Vial, K. H.
terium incorporation on the C-5 of the a-cyclogeranyl carbon skele-
ton ring is reasonable because during the cyclization, protons are
gradually eliminated from the deprotonation of the intermediate cy-
clogeranyl carbocation, which essentially compete with the intrazeo-
lite D+ for substrate protonation (versus deuteration). These results
indicate that Brønsted acidity (bridging hydroxyl groups, Al-OH-Si),
plays a dominant role in the NaY-promoted cyclization of terpenes.
1
In addition, from the H NMR spectrum of the cyclized product (see
the Supporting Information) it is obvious that D is selectively incor-
porated cis to the CH2OAc functionality, in agreement to the stereo-
chemistry of the cyclization of [D]-1 shown in Scheme 5. However,
the exact diastereomeric ratio could not be accurately measured due
to the overlapping absorptions of the non-deuterium incorporation
on C-5 a-cyclogeranyl acetate.
[30] For a similar procedure to prepare the deuterium exchanged zeolites
HZSM-5 and HSAPO-34, see: D. M. Marcus, M. J. Hayman, Y. M.
Blau, D. R. Guenther, J. O. Ehresmann, P. W. Kletnieks, J. F. Haw,
11926
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 11918 – 11927

Terpene Cyclization
FULL PAPER
Angew. Chem. 2006, 118, 1967–1969; Angew. Chem. Int. Ed. 2006,
45, 1933–1935.
[31] O. Kronja, M. Orlovic, K. Humski, S. Borcic, J. Am. Chem. Soc.
1991, 113, 2306–2308.
[32] a) V. J. Shiner, Jr., J. Am. Chem. Soc. 1953, 75, 2925–2929; b) M.
Orlovic, O. Kronja, K. Humski, S. Borcic, E. Polla, J. Org. Chem.
1986, 51, 3253–3256.
[33] a) E. Semprini, G. Perez, F. Stefani, P. Cafarelli, A. De Stefanis,
A. A. G. Tomlison, Adv. Sci. Technol. 2006, 51, 140–144; b) Y.
Huang, J. Leech, J. Phys. Chem. A 2001, 105, 6965–6970; c) N. Siva-
sankar, S. Vasudevan, Catal. Lett. 2004, 97, 53–58.
[34] G. Vitale, C. F. Mellot, A. K. Cheetham, J. Phys. Chem. B 1997, 101,
9886–9891.
[35] N. D. Ungur, N. P. Popa, N. V. Tuen, P. F. Vlad, Chem. Nat. Prod.
1993, 29, 473–478.
[36] Combined NOE experiments on bicyclic products 18 and 19, estab-
[37] C. Raptis, M. Stratakis, unpublished results.
[38] T. Hoshino, Y. Kumai, T. Sato, Chem. Eur. J. 2009, 15, 2091–2100.
1
lished the following H NMR assignments (see the Supporting Infor-
mation):
Received: June 9, 2009
Published online: September 24, 2009
Chem. Eur. J. 2009, 15, 11918 – 11927
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
11927