Titanocene-Catalyzed Cascade Cyclization of Epoxypolyprenes: Straightforward Synthesis of Terpenoids by Free-Radical Chemistry

Article abstract of DOI:10.1002/chem.200305647

The titanocene-catalyzed cascade cyclization of epoxypolyenes, which are easily prepared from commercially available polyprenoids, has proven to be a useful procedure for the synthesis of C₁₀, C₁₅, C ₂₀, and C₃₀

Full text of DOI:10.1002/chem.200305647

FULL PAPER
Titanocene-Catalyzed Cascade Cyclization of Epoxypolyprenes:
Straightforward Synthesis of Terpenoids by Free-Radical Chemistry
Josÿ Justicia,[a] Antonio Rosales,[a] Elena BuÊuel,[c] Juan L. Oller-LÛpez,[a]
MÛnica Valdivia,[a] Ali HaÔdour,[b] J. Enrique Oltra,[a] Alejandro F. Barrero,[a]
Diego J. Cµrdenas,*[c] and Juan M. Cuerva*[a]
Abstract: The titanocene-catalyzed cas-
cade cyclization of epoxypolyenes,
which are easily prepared from com-
mercially available polyprenoids, has
proven to be a useful procedure for the
synthesis of C10, C15, C20, and C30 terpe-
noids, including monocyclic, bicyclic,
and tricyclic natural products. Both
theoretical and experimental evidence
suggests that this cyclization takes
place in a nonconcerted fashion via dis-
crete carbon-centered radicals. Never-
theless, the termination step of the
process seems to be subjected to a kind
of water-dependent control, which is
unusual in free-radical chemistry. The
catalytic cycle is based on the use of
the novel combination Me3SiCl/2,4,6-
collidine to regenerate the titanocene
catalyst. In practice this procedure has
several advantages: it takes place at
room temperature under mild condi-
tions compatible with different func-
tional groups, uses inexpensive re-
agents, and its end step can easily be
controlled to give exocyclic double
bonds by simply excluding water from
the medium.
Keywords: cyclization ¥ domino re-
actions ¥ homogeneous catalysis ¥
radical reactions ¥ titanium
Introduction
tion of (S)-2,3-oxidosqualene into lanosterol has received
considerable attention in recent years[2] and there is now
solid theoretical and experimental evidence to support its
carbocationic nature.[3] Mimicking this natural transforma-
tion, Goldsmith, van Tamelen, and Corey, among others,
have exploited the acid-induced cascade cyclization of epox-
ypolyprenes as a very useful procedure in the building of
polycyclic terpenoids through carbocationic chemistry.[4]
This method involves certain drawbacks, however, such as
the need to attach extra groups to the polyene substrate to
stabilize carbocationic intermediates and control the termi-
nation steps. An alternative concept, radical cascade cycliza-
tion, introduced by Breslow and Julia[5] more than thirty
years ago, has also proven to be an excellent method for the
stereoselective synthesis of polycyclic compounds from dif-
ferent acyclic precursors.[6] To the best of our knowledge,
however, this concept was never applied to the cyclization
of epoxypolyprenes during the last century, probably owing
to the lack of a suitable protocol for the radical opening of
epoxides. Nevertheless, the titanocene(iii)-based procedure
discovered by Nugent and RajanBabu and the catalytic ver-
sion subsequently developed by Gans‰uer and co-workers
has filled this gap,[7] thus opening up the possibility of mim-
icking lanosterol synthase with free-radical chemistry. The
aim of our work here has been to take advantage of such a
method to develop a straightforward procedure for the syn-
thesis of terpenoids with a wide range of carbocyclic skele-
tons.
The increasing demand for selectivity and atom- and step-
economy in organic synthesis will presumably have a deci-
sive influence on the strategies employed by chemists in
coming years.[1] The biosynthesis of lanosterol from squalene
fits these requirements admirably, taking place as it does in
only two steps: the enantioselective epoxidation of squalene
followed by the stereoselective cascade cyclization of 2,3-ox-
idosqualene. Only one proton is lost during this process, to
form the double bond at D8. The enzyme-catalyzed cycliza-
[a] J. Justicia, A. Rosales, J. L. Oller-LÛpez, M. Valdivia, Dr. J. E. Oltra,
Prof. Dr. A. F. Barrero, Dr. J. M. Cuerva
Departamento de QuÌmica Orgµnica
Facultad de Ciencias, Universidad de Granada
Granada, 18071 (Spain)
Fax: ( +34)958-248-437
E-mail: jmcuerva@platon.ugr.es
[b] Dr. A. HaÔdour
Scientific Instrument Center
Granada, 18071 (Spain)
[c] Dr. E. BuÊuel, Dr. D. J. Cµrdenas
Departamento de QuÌmica Orgµnica
Facultad de Ciencias, Universidad AutÛnoma de Madrid
Cantoblanco, 28049 Madrid (Spain)
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author. Spectroscopic
data of some minor products and copies of selected H and 13C NMR
1
spectra.
1778
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/chem.200305647
Chem. Eur. J. 2004, 10, 1778 1788

1778 1788
Results and Discussion
The effects of water upon tita-
nocene-promoted radical cycli-
zations of epoxypolyprenes: In
preliminary experiments em-
ploying stoichiometric propor-
tions of [Cp2TiCl][8] we ob-
tained encouraging results, but
excess quantities of [Cp2TiCl2]
were required[9] and varying
amounts of reduction products
such as 14 were formed, dis-
turbing the chromatographic
isolation of the main com-
pounds and endangering the re-
producibility of the results. As
collateral observations suggest-
ed that these products might
derive
from
adventitious
water[10] we treated epoxypoly-
prene 1 with [Cp2TiCl] under
strictly anhydrous conditions. In
this manner we obtained a sub-
stantially increased yield of bi-
cyclic alkene 11 (40% isolated
product versus roughly 25% in
our preliminary experiments),[8]
together with lesser amounts of
acyclic 4 (23%) and monocyclic
7 (10%); no 14 was detected.
Moreover, when D2O was
added to the medium, deuterat-
ed isotopomer 15[11] was ob-
tained instead of 14. These re-
sults pointed to a cascade cycli-
zation via discrete carbon-cen-
tered radicals (Scheme 1), and
confirmed that the termination
step of the process can be easily
Scheme 1. Proposed mechanism for the titanocene(iii)-mediated cyclization of 1. a) [Cp2Ti(Cl)H] elimination
under anhydrous conditions; b) acidic quenching after the [Cp2Ti(Cl)H] elimination.
controlled to give either alkenes (as 11) or reduction prod-
ucts (as 14) by simply excluding or adding water to the
medium. The discovery of this water-dependent phenomen-
on, which is unusual in free-radical chemistry, guaranteed
further reproducible results.[12]
of the reaction is shown in Figure 1. Both the first (I!II)
and the second (II!III) 6-endo cyclizations are exothermic,
with reaction energies of ꢀ7.5 kcalmolꢀ1 and ꢀ8.9 kcalmolꢀ1
respectively, and both steps have moderate activation ener-
gies (11.3 and 10.6 kcalmolꢀ1 respectively). These energies
are considerably higher than those calculated for cationic
cyclizations in model systems.[3e] In these systems the second
cyclization has been calculated to proceed with activation
energies of about 1 kcalmolꢀ1, suggesting a concerted mech-
anism for the acid-catalyzed formation of A and B rings
from 2,3-oxidosqualene. In turn, the concerted process of
oxirane opening and ring A formation from the protonated
epoxide takes place with even lower barriers (about
0.6 kcalmolꢀ1).[3h] In our case, however, the values of the ac-
tivation energy barriers suggest a two step mechanism. In-
terestingly, there exists an energy minimum for radical I
with the appropriate conformation to give the first cycliza-
tion product. This type of structure was also detected at the
AM1 semiempirical level. Nevertheless, no interaction be-
Theoretical calculations supporting the nonconcerted nature
of the radical cascade cyclization: Because some controversy
remains as to whether radical cascade cyclizations take
place in a concerted or stepwise fashion,[13] we made compu-
tational studies on the cyclization of the model radical I
(closely related to 2) to gain more information about the
nature of our process. Both concerted and stepwise mecha-
nisms were considered and the pathways were carried out at
DFT level. After careful inspection of the potential energy
surface, no transition state for a concerted reaction from I
to III could be found. The theoretical calculations pointed
instead to a reaction following a two-step mechanism, in ac-
cordance with the experimental evidence. An energy profile
1779
Chem. Eur. J. 2004, 10, 1778 1788
www.chemeurj.org
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

D. J. Cµrdenas*, J. M. Cuerva et al.
FULL PAPER
Figure 1. An energy profile of the cyclization reaction of the model radical I.
tween the carbon-centered radical and the double bond
exists at this stage because the distance is too large. On the
other hand, a similar conformer for 2E,6E-10,11- epoxyfar-
nesol could not be located. Radical II exhibited an even
better pre-organization towards cyclization, which may ac-
count for the lower activation energy of the second step. All
these theoretical results strongly support the stepwise mech-
anism depicted in Scheme 1. Assuming the nonconcerted
nature of our radical cyclizations, the stereoselectivity ob-
served can be explained in terms of Beckwith Houk rules
described elsewhere.[13]
as 10 (Scheme 2). To check this hypothesis we treated epox-
ypolyprene 1 (prepared from commercially available 2E,6E-
farnesol by van Tamelen×s procedure)[16] with a substoichio-
metric quantity of [Cp2TiCl2] (0.2 equiv), Mn dust, and the
mixture of Me3SiCl and collidine in dry THF (10ꢀ1 m sub-
strate concentration) (Scheme 2). In this way we obtained
the expected exocyclic alkene[17] 11 (after fluoride workup)
at the same yield (40%) as that under stoichiometric condi-
tions but employing lower [Cp2TiCl2] proportions and dilu-
tion levels by one and two orders of magnitude respectively.
This result supported the main features of the catalytic cycle
depicted in Scheme 2.
Development of the titanocene-catalyzed version: With val-
uable mechanistic data available to us, we envisaged the de-
velopment of a catalytic version to reduce the considerable
proportions of [Cp2TiCl2] and the high dilutions required in
our preliminary experiments.[14] Our starting hypothesis was
Synthesis of terpenoids with various carbocyclic skeletons:
Once we were confident about the viability of the titano-
cene-catalyzed cyclization and the experimental conditions
required to control the end step of the process, we decided
that with a judicious choice of starting material this method
might be a useful tool for the synthesis of terpenoids with
different carbon skeletons, including monocyclic compounds
such as 18, 24, and 25, bicyclic sesquiterpenoids (such as 26)
based on the use of the novel combination Me3SiCl/2,4,6-
[15]
collidine,[10b]
,
which is compatible with oxiranes and
should be capable of regenerating [Cp2TiCl2] from both
[Cp2Ti(Cl)H] and oxygen-bonded titanium derivatives such
1780
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 1778 1788

Titanocene-Catalyzed Synthesis of Terpenoids
1778 1788
sequent treatment of 21 with vinylmagnesium bromide pro-
vided tertiary alcohol 22 as a mixture (9R* and 9S* epimers)
in a 3:2 isomeric ratio. Selective esterification of the secon-
dary alcohol of 22, followed by allylic hydroxylation of 23
afforded a 3:2 mixture of the 9R* and 9S* epimers 24.
Originally Marco and co-workers did not establish the C-
9 stereochemistry of the metabolite found in A. chamaemeli-
folia.[18] Recently, however, Uttaro et al. have demonstrated
the 9R stereochemistry of the natural product by means of
chemical synthesis and X-ray crystallographic analysis.[20] In
our epimeric mixture (24) the NMR signals corresponding
to the major component matched those of the natural me-
tabolite,[18] whereas the signals of the minor one agreed
closely with those of the 9S isomer.[20] Therefore we com-
pleted the total synthesis of the natural product in seven
steps in 23% overall yield, confirming the usefulness of our
method for the preparation of cyclofarnesane-type monocy-
clic sesquiterpenoids. Ketone 21 also proved to be a valua-
ble intermediate for the total synthesis of the monocyclic tri-
terpenoid achilleol A (25) (Scheme 3) following the conver-
gent strategy recently developed in our laboratory.[21]
Scheme 2. Hypothetical catalytic cycle for the TiIII mediated cyclization
of 1 to 11.
and diterpenoids (such as 30),
as well as tricyclic products as
the isocopalane diterpenoid 36.
As we expected, the titano-
cene-catalyzed cyclization of
6,7-epoxygeranyl acetate[8] (17)
under anhydrous conditions se-
lectively gave 1,3-cis-disubsti-
tuted monoterpenoid 18 with
an exocyclic double bond
(Scheme 3). The initial results
obtained in the synthesis of 18
encouraged us to extend our
method to the preparation of
more complexmonocyclic ter-
penoids. Cyclofarnesane sesqui-
terpenoid 24 was discovered by
Marco et al.[18] in the plant Ar-
temisia chamaemelifolia togeth-
er with other polyoxygenated
metabolites. We started its syn-
thesis (Scheme 3) with commer-
cial geranylacetone, which was
easily transformed into epoxy-
ketal 19 by conventional
chemistry (see Experimental
Section). Unlike ketones, the
ketal group of 19 proved to be
inert
toward
free-radical
Scheme 3. Titanocene-catalyzed synthesis of monocyclic terpenoids. DMAP = 4-(dimethylamino)pyridine.
chemistry (at least under our
conditions) and remained un-
changed after titanocene-catalyzed cyclization of 19 to 20
(61% yield). The deprotection of the carbonyl group with
cerium(iii) chloride[19] avoided extensive isomerization of
the exocyclic double bond of 20 (promoted by other acids),
and an excellent 95% yield of ketone 21 was obtained. Sub-
Drimanes constitute a family of bicyclic sesquiterpenoids
with interesting biological properties.[22] The simple saponifi-
cation of 11 (obtained from commercial farnesol as descri-
1
bed above) gave synthetic drimane 26 (Scheme 4), with H
and 13C NMR data in accordance with those of natural iso-
1781
Chem. Eur. J. 2004, 10, 1778 1788
www.chemeurj.org
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

D. J. Cµrdenas*, J. M. Cuerva et al.
FULL PAPER
drimenediol excreted by the fungus Polyporus arcularius.[23]
We thus achieved the total synthesis of isodrimenediol in
just four steps, high regio- and stereoselectivity degrees, and
in a considerable overall yield of 21%. To the best of our
techniques. Apart from optical rotation, synthetic 30 had the
same physical properties as natural (+)-3b-hydroxyma-
nool[24] and thus the first total synthesis of this terpenoid
was achieved in five steps in an overall yield of 6%. It
should be noted that the rela-
tive proportions of products 30
and 31 obtained from the reac-
tion with vinylmagnesium bro-
mide revealed that the nucleo-
philic attack by the Si face of
ketone 29 was faster than that
by the Re face.
Dinor-labdane 33 was recent-
ly isolated from copaiba oil and
its structure elucidated by
NMR spectroscopy, but the rel-
ative stereochemistry at C-13
had not so far been deter-
mined.[27] We attempted its syn-
thesis by reducing ketone 29
with NaBH4 (Scheme 4). We
thus obtained a mixture of two
epimeric alcohols, 32 and 33, in
relative proportions of 6:5 re-
spectively (1H NMR analysis).
When L-Selectride was used in-
stead of NaBH4 the stereoselec-
tivity of the reduction in-
creased, and the product ratio
was 32:33 = 3:1. Since the Si
face of ketone 29 proved to be
more reactive than the opposite
face against nucleophilic re-
agents (see above) we tenta-
tively assigned the 13R* rela-
tive configuration (derived
from the hydride attack by the
Si face) to the major product
(32) and, consequently, the
Scheme 4. Titanocene-catalyzed synthesis of bicyclic terpenoids
13S* to the minor one (33).
knowledge this is the first total synthesis reported for isodri-
menediol and confirms the structure 26 proposed by Fleck
et al. for the fungal metabolite.[23]
Both diastereomers 32 and 33 were isolated (45% and 37%
yields respectively) and their NMR spectra were compared
with those of the copaiba oil component. The 13C NMR
spectrum of the minor isomer 33 virtually matched that re-
ported for the natural compound,[27] whereas in the spec-
trum of 32 slight but significant differences were observed in
the chemical shifts of carbons C-8, C-9, C-11 to C-14, and
C-17 (see Table 1). Therefore, we propose the relative ster-
eochemistry 13S* depicted in 33 for the bicyclic terpenoid
isolated from copaiba oil.
The marine metabolite stypoldione (37) has attracted the
attention of chemists owing both to its pharmacological
properties[28] and its challenging chemical structure. Recent-
ly Xing and Demuth reported an elegant total synthesis of
stypoldione via the tricyclic intermediate 36.[29] Because of
the biological interest of stypoldione, we selected the isoco-
palane diterpenoid 36 as a target to prove the efficiency of
our method for the synthesis of tricyclic terpenoids from ep-
oxypolyene 34, previously prepared from commercially
We then addressed the chemical preparation of 3b-hy-
droxymanool (30), a bicyclic diterpenoid with a labdane
skeleton from the fern Gleichenia japonica.[24] As starting
material we chose commercial farnesylacetone,[25] which was
successively transformed into epoxyketal 27, cyclic deriva-
tive 28, and ketone 29 (Scheme 4), in the same way that ger-
anylacetone was transformed into ketone 21 (see Scheme 3).
Interestingly, the NMR data of synthetic ketone 29 matched
those of one of the components of copaiba oil (a commer-
cial mixture of natural oleoresins used both for cosmetics
and medicinal purposes),[26] confirming the chemical struc-
ture of this natural product. The treatment of ketone 29
with vinylmagnesium bromide provided 30 (39% isolated
yield) together with a lesser quantity of its 13S* epimer 31
(23% yield). Fortunately both isomers could be easily isolat-
ed by flash chromatography and analyzed by spectroscopic
1782
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 1778 1788

Titanocene-Catalyzed Synthesis of Terpenoids
1778 1788
Table 1. 13C NMR data[a] for a natural dinor-labdane terpenoid (33) isolated from copaiba oil and the synthetic compounds 32 and 33.
Carbon
Natural
Synthetic 33
Dd
Synthetic 32
Dd
1
2
3
4
5
6
7
8
37.16
27.98
78.94
39.18
54.69
24.06
38.23
148.19
56.80
39.52
20.05
38.54
68.90
23.60
106.78
14.47
15.46
28.37
37.22
28.04
78.96
39.22
54.76
24.10
38.27
148.23
56.88
39.57
20.09
38.60
68.91
23.62
106.79
14.49
15.47
28.39
0.06
0.06
0.02
0.04
0.07
0.04
0.04
0.04
0.08
0.05
0.04
0.06
0.01
0.02
0.01
0.02
0.01
0.03
37.24
28.05
78.98
39.23
54.79
24.08
38.26
148.02
56.59
39.50
19.72
38.38
68.47
23.79
107.00
14.50
15.46
28.40
0.08
0.07
0.04
0.05
0.10
0.02
0.03
0.17
0.21
0.02
0.33
0.16
0.43
0.19
0.22
0.03
0.00
0.04
9
10
11
12
13
14
17
18
19
20
[a] The most significant data are in bold characters.
available geranylgeraniol by van Tamelen×s procedure.[30] Ti-
tanocene-catalyzed cyclization of 34 gave tricyclic alkene 35
in a moderate 31% yield (Scheme 5). This yield can be re-
radical-based method constitutes an especially convenient
alternative to conventional carbocationic chemistry when
the synthetic targets are cyclic terpenoids bearing exocyclic
double bonds.[17]
Titanocene-catalyzed
cycliza-
tion of 2,3-oxidosqualene, mim-
icking the enzyme lanosterol
synthase by free-radical chemis-
try: Finally, the possibility of
achieving the first radical cycli-
zation of 2,3-oxidosqualene
(38) encouraged us to prepare
this epoxide from commercially
available squalene[31] and treat
it with a catalytic quantity of ti-
tanocene (Scheme 6). In this
manner we obtained malabari-
cane 39[32] and its 13b-epimer
Scheme 5. Titanocene-catalyzed synthesis of tricyclic terpenoids.
garded as satisfactory, however,
if we bear in mind that the syn-
thesis of 35 selectively afforded
a
product containing three
fused (trans/anti/trans) six-mem-
bered rings, an exocyclic double
bond, and sixstereogenic cen-
ters, among 192 potential regio-
and stereoisomers. Catalytic hy-
drogenation of 35 gave 36
(73% yield) and thus the
formal synthesis of stypoldione
was completed.
All the above results confirm
the value of our procedure for
synthesizing terpenoids with
different carbon skeletons, in-
cluding monocyclic, bicyclic,
and tricyclic products. Our free- Scheme 6. Titanocene-catalyzed cyclization of 2,3-oxidosqualene.
1783
Chem. Eur. J. 2004, 10, 1778 1788
www.chemeurj.org
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

D. J. Cµrdenas*, J. M. Cuerva et al.
FULL PAPER
40, together with minor amounts of the acyclic alcohol 41
and achilleol A (25). Bicyclic compounds or Wagner Meer-
wein rearrangement products, as described for the acid-in-
duced cyclization of 38,[31] were not detected.
Apart from the preparative interest (total synthesis of ma-
labaricanes in only two steps), the above results also have
mechanistic relevance and merit some further comment. As
in the acid-induced rearrangement of 38,[31] the main prod-
ucts (39 and 40) derive from a 6-endo/6-endo/5-exo cycliza-
tion process[33] (Scheme 7), but under our conditions the 5-
numbering), which is similar to a nonclassical carbocation.[3f]
Through free-radical chemistry, however, it seems unlikely
that the double bond at D17 could give anchimeric assistance
to facilitate ring-C expansion and D-ring formation from 45.
Therefore, this radical has no option but to evolve towards
malabaricatrienes (39 and 40). This intrinsic tendency of
free-radical chemistry to give malabaricanes from 2,3-oxi-
dosqualene (and possibly from squalene also) is intriguing
from a biogenetic point of view. The recent discovery of ma-
labaricanes in marine sediments,[34] for example, is especially
relevant because it is believed that they are synthesized by
organisms living under anoxic conditions similar to those
provided by the strictly deoxygenated solvents required for
free-radical chemistry.
Conclusion
We have developed a novel procedure for the straightfor-
ward total synthesis of terpenoids with different carbon skel-
etons by means of free-radical chemistry. This method has
proven to be useful for synthesizing C10, C15, C20, and C30 ter-
penoids, including monocyclic, bicyclic, and tricyclic natural
products. The key step of the process is the titanocene-cata-
lyzed cascade cyclization of epoxypolyenes, easily prepared
from commercially available polyprenoids. The cyclization
proceeds with high regio- and stereoselectivity and provides
yields which can generally be regarded as satisfactory.
Mechanistically the reaction is likely to occur via discrete
carbon-centered radicals, but the termination step of the
process seems to be subject to a type of water-dependent
control that is unusual in free-radical chemistry. In practice
the method has many advantages: it proceeds at room tem-
perature under mild conditions compatible with several
functional groups, uses inexpensive reagents, and the termi-
nation step can easily be controlled to give exocyclic al-
kenes. Moreover, as epoxypolyprenes can be enantioselec-
tively obtained by asymmetric catalysis, an enantioselective
version of our method seems plausible. We are currently
working on this task and the application of our procedure to
the synthesis of marine terpenoids containing seven-mem-
bered rings.
Experimental Section
Scheme 7. Proposed mechanism for the titanocene-catalyzed cyclization
of 2,3-oxidosqualene.
General: For the reactions employing titanocene all solvents and addi-
tives were thoroughly deoxygenated prior to use. The numbering used in
the NMR assignments corresponds to the cyclofarnesane, drimane, lab-
dane, and isocopalane systems and not the IUPAC nomenclature. Epox-
ides 1,[16] 17,[8] 34,[30] and 38[31] were prepared according to known proce-
dures. The following known compounds were isolated as pure samples
and showed identical NMR spectra to the reported compounds: 11,[8]
exo cyclization step giving the protomalabaricane radical 45
seems to be specially fast, thus avoiding the generation of
bicyclic byproducts (see ref. [30b]). It is generally accepted
nowadays that the biosynthesis of lanosterol takes place via
a carbocation intermediate with a tricyclic skeleton contain-
ing a five-membered C-ring closely related to 45.[2b] In this
context, recent theoretical calculations suggest that this in-
termediate undergoes a C-ring expansion and concomitant
D-ring formation through a transition structure involving
the double bond between C-17 and C-18 (malabaricane
14,[6c] 18,[35] 24,[18] 25,[36] 26,[23] 29,[26] 30,[24] 33,[27] and 36.[29] Other general
experimental details have been reported elsewhere.[8]
,
[10a]
General procedure for the titanocene-catalyzed cyclization of epoxypoly-
prenes: Strictly deoxygenated THF (20 mL) was added to a mixture of
[Cp2TiCl2] (0.5 mmol) and Mn dust (20 mmol) under an Ar atmosphere
and the suspension was stirred at room temperature until it turned lime
green (after about 15 min). Then, a solution of epoxide (2.5 mmol) and
2,4,6-collidine (20 mmol) in THF (2 mL), and Me3SiCl (10 mmol) were
1784
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 1778 1788

Titanocene-Catalyzed Synthesis of Terpenoids
1778 1788
added and the solution was stirred for 8 h. The reaction was then
quenched with 2n HCl and extracted with tBuOMe. The organic layer
was washed with brine, dried (anhydrous Na2SO4), and the solvent re-
moved. The residue was dissolved in THF (20 mL) and stirred with
Bu4NF (10 mmol) for 2 h. The mixture was then diluted with tBuOMe,
washed with brine, dried (anhydrous Na2SO4), and the solvent removed.
Products obtained were isolated by column chromatography of the resi-
due on silica gel (hexane/tBuOMe) and characterized by spectroscopic
techniques. The main polycyclic compounds were isolated in the follow-
ing yields: 11 (40%), 18 (51%), 20 (61%), 28 (42%), 35 (31%), and 39
40 (39%).
Synthesis of monocyclic diol 22: Vinylmagnesium bromide (1m in THF,
0.22 mL, 0.22 mmol) was added to a solution of 21 (14 mg, 0.07 mmol) in
THF (3 mL) at 08C. The reaction was stirred for 2 h and then quenched
with ice-water, extracted with tBuOMe, dried (anhydrous Na2SO4), and
the solvent removed. Flash chromatography (hexane/tBuOMe 3:2) of the
residue gave a mixture of the 9R* and 9S* epimeric alcohols 22 (14 mg,
90%) in a 3:2 ratio. 1H NMR (300 MHz, CDCl3): d = 5.90 (dd, J =
17.3, 10.7 Hz, 1H; 9R* isomer), 5.89 (dd, J = 17.3, 10.7 Hz, 1H; 9S*
isomer), 5.20 (d, J
= 17.3 Hz, 1H), 5.05 (d, J = 10.7 Hz, 1H; 9S*
isomer), 5.04 (d, J = 10.7 Hz, 1H; 9R* isomer), 4.85 (s, 1H), 4.58 (s, 1H;
9R* isomer), 4.55 (s, 1H; 9S* isomer), 3.39 (dd, J = 9.7, 4.3 Hz, 1H),
2.30 (dt, J = 12.9, 5.0 Hz, 1H), 2.00 0.80 (m, 8H), 1.26 (s, 3H), 1.03 (s,
3H; 9S* isomer), 1.02 (s, 3H; 9R* isomer), 0.71 ppm (s, 3H); 13C NMR
Synthesis of deuterium-labeled drimane 15: A mixture of [Cp2TiCl2]
(703 mg, 2.85 mmol) and Mn dust (412 mg, 7.56 mmol) in THF (25 mL)
was stirred at room temperature until the red solution turned green. Sub-
sequently, the green solution was slowly added to a mixture of 1 (100 mg,
0.36 mmol) and D2O (128 mg, 7.14 mmol) in THF (20 mL), and was stir-
red at room temperature for 24 h. The reaction was then quenched with
5% aqueous NaH2PO4 and extracted with tBuOMe. The organic layer
was washed with brine, dried (anhydrous Na2SO4), and the solvent re-
moved. The residue was chromatographed (hexane/tBuOMe 7:3) afford-
(75 MHz, CDCl3, DEPT): d
= 147.40 (C), 147.34 (C), 145.36 (CH),
145.19 (CH), 111.79 (CH2), 111.69 (CH2), 108.73 (CH2), 108.64 (CH2),
77.32, (CH), 73.63 (C), 73.51 (C), 52.26 (CH), 52.15 (CH), 41.62 (CH2),
41.56 (CH2), 40.83 (C), 34.52 (CH2), 32.94 (CH2), 28.16 (CH3), 27.78
(CH3), 26.49 (CH3), 19.76 (CH2), 15.76 ppm (CH3), (some signals were
not observed); HRMS (FAB): calcd for C15H26O2Na 261.1830, found
261.1835.
Preparation of acetate 23: A mixture of 22 (35 mg, 0.15 mmol), Ac2O
(17 mg, 0.16 mmol), and DMAP (20 mg, 0.16 mmol) in CH2Cl2 (5 mL)
was stirred at room temperature for 1 h. The mixture was then diluted
with tBuOMe and washed with 2n HCl, saturated NaHCO3, and brine.
The organic layer was dried (anhydrous Na2SO4) and the solvent re-
moved. Flash chromatography (hexane/tBuOMe 4:1) of the residue gave
a mixture of the 9R* and 9S* epimeric acetates 23 (33 mg, 80%) in a 3:2
1
ing 15 (36 mg, 36% yield). H and 13C NMR spectra of 15 matched those
of the isotopomer 14,[6c] except for the following significant signals: d =
0.92 (brs; H3 12), 28.97 ppm (small t, 1J(13C,D) = 20.1 Hz; C-8); MS
(70 eV, EI): m/z (%): 265 (3), 222 (10), 121 (100); HRMS (FAB): calcd
for C17H29DO3Na: (M+) 306.2155, found 306.2160. Minor amounts of a
1
C-8 epimer could also be detected. H and 13C NMR signals of this minor
[37]
isomer agreed closely with a related structure described elsewhere.[6c]
,
1
ratio. H NMR (300 MHz, CDCl3): d = 5.90 (dd, J = 17.3, 10.7 Hz, 1H;
9R* isomer), 5.89 (dd, J = 17.3, 10.7 Hz, 1H; 9S* isomer), 5.19 (d, J =
17.3 Hz, 1H), 5.01 (d, J = 10.7 Hz, 1H), 4.85 (s, 1H), 4.64 (dd, J = 9.5,
4.1 Hz, 1H), 4.60 (s, 1H; 9R* isomer), 4.57 (s, 1H; 9S* isomer), 2.30 0.80
(m, 9H), 2.02 (s, 3H), 1.25 (s, 3H), 0.91 (s, 3H), 0.76 ppm (s, 3H); 13C
Preparation of epoxide 19: Powdered NBS (1.73 g, 9.75 mmol) was grad-
ually added to a solution of geranylacetone ethylene ketal[38] (1.50 g,
6.34 mmol) in a mixture of DME/water (100 mL, 3:2) at 08C. The reac-
tion was stirred for 30 min, diluted with tBuOMe, washed with water,
dried (anhydrous Na2SO4), and the solvent removed. The residue was dis-
solved in 0.5m methanolic K2CO3 (20 mL) and stirred for 10 min. The
methanolic solution was then diluted with tBuOMe, washed with water,
dried (anhydrous Na2SO4), and concentrated to dryness, giving a residue
which was submitted to flash chromatography (hexane/tBuOMe 4:1) to
NMR (75 MHz, CDCl3, DEPT): d
= 170.68 (C), 146.86 (C), 145.38
(CH), 145.34 (CH), 111.65 (CH2), 111.61 (CH2), 109.54 (CH2), 109.52
(CH2), 78.60 (CH), 78.51 (CH), 73.51 (C), 73.39 (C), 53.46 (CH2), 52.63
(CH), 41.40 (CH2), 41.32 (CH2), 39.44 (C), 28.66 (CH2), 28.31 (CH3),
27.88 (CH3), 26.30 (CH3), 26.22 (CH3), 21.31 (CH3), 19.98 (CH2),
17.97 ppm (CH3), (some signals were not observed); MS (70 eV, EI): m/z
(%): 262 (1), 205 (28), 96 (100); HRMS (EI): calcd for C17H28O3Na
303.1936, found 303.1929.
1
give 19 (1.15 g, 71%) as a colorless oil. H NMR (400 MHz, CDCl3): d =
5.18 (t, J = 5.1 Hz, 1H), 3.94 (m, 4H), 2.71 (t, J = 6.3 Hz, 1H), 2.21
2.04 (m, 5H), 1.63 (s, 3H), 1.73 1.55 (m, 3H), 1.33 (s, 3H), 1.30 (s, 3H),
1.26 ppm (s, 3H); 13C NMR (75 MHz, CDCl3, DEPT): d = 134.36 (C),
124.68 (CH), 109.94 (C), 64.71 (CH2), 64.21 (CH), 58.35 (C), 39.11 (CH2),
36.34 (CH2), 27.49 (CH2), 24.95 (CH3), 23.86 (CH3), 22.73 (CH2), 18.81
(CH3), 15.99 ppm (CH3); HRMS (FAB): calcd for C15H26O2Na [M+]
277.1779, found 277.1773.
Data for the cyclic alcohol 20: Colorless oil; 1H NMR (300 MHz,
CDCl3): d = 4.85 (brs, 1H), 4.60 (brs, 1H), 3.91 (m, 4H), 3.39 (dd, J =
9.7, 4.3 Hz, 1H), 2.30 (dt, J = 13.0, 4.6 Hz, 1H), 2.20 0.75 (m, 8H), 1.30
(s, 3H), 1.03 (s, 3H), 0.69 ppm (s, 3H); 13C NMR (75 MHz, CDCl3,
DEPT): d = 147.26 (C), 110.40 (C), 108.66 (CH2), 77.35 (CH), 64.67
(CH2), 51.85 (CH), 40.76 (C), 38.23 (CH2), 33.14 (CH2), 32.30 (CH2),
25.98 (CH3), 23.92 (CH3), 19.80 (CH2), 15.47 ppm (CH3); MS: m/z (%):
254 (1) [M+], 239 (1) [M+ꢀCH3], 221 (1) [M+ꢀCH3ꢀH2O, 159 (12), 87
(100). Minor signals (13:1 relationship) for an endocyclic regioisomer
could be observed in the 1H NMR spectrum (300 MHz, CDCl3): (only
distinctive signals) d = 5.26 (brs, 1H), 3.43 ppm (dd, J = 7.7, 5.6 Hz,
1H).
Synthesis of monocyclic sesquiterpene 24: SeO2 (4 mg, 0.012 mmol) and
tbutyl hydroperoxide (70 wt% in water, 0.056 mL, 0.35 mmol) were
added to a solution of 23 (33 mg, 0.12 mmol) in CH2Cl2 (2 mL) and the
mixture was stirred at room temperature for 48 h. It was then diluted
with tBuOMe and washed with 10% KOH and brine. The organic layer
was dried (anhydrous Na2SO4) and the solvent removed. Flash chroma-
tography (hexane/tBuOMe 3:2) of the residue gave a mixture of the 9R*
and 9S* epimeric sesquiterpenes 24 (19 mg, 55%) in a 3:2 ratio.
Synthesis of isodrimenediol 26:
A sample of acetate 11 (440 mg,
1.60 mmol) was dissolved in 0.5m methanolic K2CO3 (50 mL) and stirred
at room temperature for 15 h. Then tBuOMe was added and the mixture
was washed with aqueous 2n HCl and brine, dried (anhydrous Na2SO4),
and the solvent removed. Flash chromatography (hexane/tBuOMe 1:1) of
the residue gave 26 (374 mg, quantitative yield).
Preparation of epoxide 27: Powdered NBS (324 g, 1.82 mmol) was gradu-
ally added to a solution of farnesylacetone ethylene ketal[39] (500 mg,
1.65 mmol) in a mixture of DME/water (100 mL, 3:2) at 08C. The reac-
tion was stirred for 30 min, diluted with tBuOMe, washed with water,
dried (anhydrous Na2SO4), and the solvent removed. The residue was dis-
solved in 0.5m methanolic K2CO3 (20 mL) and stirred for 40 min. The
methanolic solution was then diluted with tBuOMe, washed with water,
dried (anhydrous Na2SO4), and concentrated to dryness. The residue was
submitted to flash chromatography (hexane/tBuOMe 4:1) affording col-
orless oil 27 as a mixture of four stereoisomers (220 mg, 42%). 1H NMR
(300 MHz, CDCl3): d = 5.20 5.00 (m, 2H), 3.90 3.80 (m, 4H), 2.70 2.60
(m, 1H), 1.62 (s, 3H), 1.57 (s, 3H), 1.26 (s, 3H), 1.21 (s, 3H), 1.20 ppm (s,
3H); 13C NMR (75 MHz, CDCl3, DEPT): d = 135.80 (C), 135.01 (C),
134.67 (C), 134.92 (C), 125.71 (CH), 125.04 (CH), 124.90 (CH), 124.81
(CH), 124.20 (CH), 124.12 (CH), 109.93 (C), 64.65 (CH2), 64.16 (CH),
64.08 (CH), 58.64 (C), 39.86 (CH2), 39.60 (CH2), 39.42 (CH2), 39.14
(CH2), 36.34 (CH2), 32.02 (CH2), 31.81 (CH2), 28.56 (CH2), 28.52 (CH2),
Obtention of ketone 21:
A solution of 20 (160 mg, 0.69 mmol),
[CeCl3¥7H2O] (726 mg, 1.95 mmol), and NaI (57 mg, 0.38 mmol) in
MeCN (50 mL) was stirred at room temperature for 16 h. The mixture
was diluted with tBuOMe, washed with brine, dried (anhydrous Na2SO4),
and the solvent removed. Flash chromatography (hexane/tBuOMe 1:1) of
the residue afforded 21 (125 mg, 95%) as a white solid. M.p. 38 408C;
1H NMR (400 MHz, CDCl3): d = 4.84 (brs, 1H), 4.50 (brs, 1H), 3.38
(dd, J = 8.9, 4.1 Hz, 1H), 2.51 (ddd, J = 14.3, 9.1, 5.0 Hz, 1H), 2.35 2.23
(m, 2H), 2.08 (s, 3H), 2.00 1.40 (m, 6H), 1.01 (s, 3H), 0.74 ppm (s, 3H);
13C NMR (75 MHz, CDCl3, DEPT): d = 209.27 (C), 147.23 (C), 108.78
(CH2), 76.94 (CH), 51.38 (CH), 42.95 (CH2), 40.55 (C), 32.24 (CH2),
32.05 (CH2), 30.05 (CH3), 26.11 (CH3), 19.69 (CH2), 16.25 ppm (CH3);
HRMS (FAB): calcd for C13H22O2Na 233.1517, found 233.1519.
1785
Chem. Eur. J. 2004, 10, 1778 1788
www.chemeurj.org
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

D. J. Cµrdenas*, J. M. Cuerva et al.
FULL PAPER
27.51 (CH2), 27.49 (CH2), 26.60 (CH2), 26.53 (CH2), 26.49 (CH2), 26.30
(CH2), 24.90 (CH3), 23.83 (CH3), 23.40 (CH3), 22.68 (CH2), 22.59 (CH2),
18.77 (CH3), 18.74 (CH3), 15.97 ppm (CH3); HRMS (FAB): calcd for
C20H34ONa 345.2405, found 345.2408.
phere, and the suspension was stirred at room temperature until it turned
lime green (about 15 min). A solution of 2,3-oxidosqualene 38 (500 mg,
1.17 mmol) and 2,4,6-collidine (1.0 mL, 8.19 mmol) in THF (2 mL) and
Me3SiCl (0.60 mL, 4.68 mmol) were then added and the solution was stir-
red for 4 h. The reaction was then quenched with 2n HCl and extracted
with tBuOMe. The organic layer was washed with brine, dried (anhy-
drous Na2SO4), and the solvent removed. The residue was dissolved in
THF (20 mL) and stirred with Bu4NF (1.1 g, 3.51 mmol) for 2 h. The mix-
ture was then diluted with tBuOMe, washed with brine, dried (anhydrous
Na2SO4), and the solvent removed. Column chromatography of the resi-
due on 20% AgNO3/silica gel (hexane/tBuOMe 9:1) afforded malabari-
canes 39 (75 mg, 15%) and 40 (120 mg, 24%), as well as allylic alcohol
41 (trace) and achilleol A (25) (trace).
Data for 39: 1H NMR (300 MHz, CDCl3): d = 5.10 (t, J = 6.0 Hz, 1H),
5.08 (t, J = 6.0 Hz, 1H), 4.87 (brs, 1H), 4.58 (brs, 1H), 3.20 (dd, J =
12.0, 5.8 Hz, 1H), 2.10 1.95 (m, 9H), 1.67 (s, 3H), 1.59 (s, 6H), 0.96 (s,
3H), 0.94 (s, 3H), 0.84 (s, 3H), 0.77 ppm (s, 3H); 13C NMR: (75 MHz,
CDCl3): d = 154.5 (C), 135.2 (C), 131.3 (C), 124.5 (CH), 124.3 (CH),
108.9 (CH2), 79.3 (CH), 56.4 (CH), 55.8 (CH), 55.5 (CH), 45.3 (C), 40.3
(C), 39.8 (CH2), 39.3 (CH2), 38.8 (CH2), 37.2 (C), 36.6 (CH2), 28.1 (CH3),
27.7 (CH2), 27.4 (CH2), 26.9 (CH2), 26.8 (CH2), 25.7 (CH3), 24.8(CH3),
20.8 (CH2), 19.1 (CH2), 17.7 (CH3), 16.1 (CH3), 15.7 (CH3), 15.4 ppm
(CH3); MS (70 eV, EI): m/z (%): 426 (3), 411 (1), 247 (10), 207 (30), 189
(20), 135 (25), 69 (100).
Data for cyclic alcohol 28: White solid; m.p. 95 1008C; 1H NMR
(300 MHz, CDCl3): d = 4.81 (brs, 1H), 4.55 (brs, 1H), 3.95 3.85 (m,
4H), 3.22 (dd, J = 11.5, 4.1 Hz, 1H), 2.37 (ddd, J = 12.7, 4.0, 2.5 Hz,
1H), 2.05 0.80 (m, 13H), 1.28 (s, 3H), 0.96 (s, 3H), 0.75 (s, 3H),
0.66 ppm (s, 3H); 13C NMR (75 MHz, CDCl3; DEPT): d = 148.01 (C),
110.45 (C), 106.95 (CH2), 78.94 (CH), 64.71 (CH2), 64.69 (CH2), 56.71
(CH), 54.72 (CH), 39.50 (C), 39.21 (C), 38.23 (CH2), 37.99 (CH2), 37.18
(CH2), 28.39 (CH3), 28.02 (CH2), 24.06 (CH2), 23.88 (CH3), 18.04 (CH2),
15.49 (CH3), 14.47 ppm (CH3); MS (70 eV, EI): m/z (%): 322 (1) [M]+,
289 (1), 260 (8), 135 (15); HRMS (FAB): calcd for C20H34O3N: 345.2405,
found 345.2399.
Preparation of bicyclic ketone 29: A solution of 28 (66 mg, 0.20 mmol),
[CeCl3¥7H2O] (273 mg, 0.73 mmol), and NaI (22 mg, 0.14 mmol) in
MeCN (10 mL) was stirred at room temperature for 16 h. The mixture
was diluted with tBuOMe, washed with brine, dried (anhydrous Na2SO4),
and the solvent removed. Flash chromatography (hexane/tBuOMe 7:3) of
the residue afforded 29 (49 mg, 85%).
Synthesis of 3b-hydroxymanool (30): Vinylmagnesium bromide (1m in
THF, 0.5 mL, 0.5 mmol) was added to
a solution of 29 (22 mg,
0.08 mmol) in THF (5 mL) at 08C and stirred for 30 min. The reaction
was quenched with ice water, extracted with tBuOMe, dried (anhydrous
Na2SO4), and the solvent removed. Flash chromatography (hexane/
tBuOMe 1:1) of the residue gave epimeric alcohols 30[24] (9.5 mg, 39%)
and 31 (5.5 mg, 23%). Data for 31: vitreous solid; 1H NMR (300 MHz,
CDCl3): d = 5.89 (dd, J = 17.3, 10.7 Hz, 1H), 5.20 (d, J = 17.3, 1H),
5.05 (d, J = 10.7 Hz, 1H), 4.81 (brs, 1H), 4.48 (brs, 1H), 3.23 (dd, J =
11.6, 4.6 Hz, 1H), 2.38 (ddd, J = 12.8, 6.7, 2.6 Hz, 1H) 2.00 0.80 (m,
13H), 1.26 (s, 3H), 0.98 (s, 3H), 0.76 (s, 3H), 0.67 ppm (s, 3H); 13C NMR
(75 MHz, CDCl3, DEPT): d = 148.21 (C), 145.19 (CH), 111.78 (CH2),
106.86 (CH2), 78.98 (CH), 73.71 (C), 57.05 (CH), 54.77 (CH), 41.36
(CH2), 38.67 (C), 38.33 (C), 38.28 (CH2), 37.18 (CH2), 28.39 (CH3), 28.21
(CH2), 28.04 (CH3), 24.10 (CH2), 17.92 (CH2), 15.47 (CH3), 14.55 ppm
(CH3); MS (70 eV, EI): m/z (%): 273 (6), 255 (8), 135 (100); HRMS (EI):
calcd for C20H34O2 306.2558, found 306.2563.
Data for 40: 1H NMR (300 MHz, CDCl3): d = 5.11 (t, J = 6 Hz, 1H),
5.10 (t, J = 6.0 Hz, 1H), 4.89 (brs, 1H), 4.73 (brs, 1H), 3.20 (t, J =
7.8 Hz, 1H), 2.20 1.95 (m, 9H), 1.67 (s, 3H), 1.59 (s, 6H), 0.97 (s, 3H),
0.84 (s, 3H), 0. 78 (s, 3H), 0.65 ppm (s, 3H); 13C NMR (75 MHz, CDCl3):
d = 149.3 (C), 135.0 (C), 131.3 (C), 124.5 (CH), 124.3 (CH), 110.1(CH2),
79.3 (CH), 63.2 (CH), 57.1 (CH), 56.3 (CH), 43.6 (C), 41.0 (C), 39.8
(CH2), 38.8 (C), 38.4 (CH2), 37.6 (CH2), 36.9 (C), 28.2 (CH3), 27.3 (CH2),
27.1 (CH2), 26.8 (CH2), 25.7 (CH3), 25.4 (CH2), 19.6 (CH2), 19.3 (CH2),
17.7 (CH3), 16.1 (CH3), 15.5 (CH3), 15.3 (CH3), 15.1 ppm (CH3); MS
(70 eV, EI): m/z (%): 426 (3), 411 (1), 247 (10), 207 (30), 189 (20), 135
(25), 69 (100); HRMS (EI): calcd for C30H50ONa 449.3760, found
449.3759.
Computational methods: Calculations were made with the GAUSSIAN
98 series of programs.[40] The geometries of all intermediates were opti-
mized at the DFT level employing the B3LYP hybrid functional,[41] using
the standard 6 31G(d) basis set for C, H, and O. Harmonic frequencies
were calculated at the same level of theory to characterize the stationary
points and to determine the zero-point energies (ZPE). More accurate
energies were determined by single-point calculations at the same level
using the 6 311+G(d,p) basis set. Final energies include ZPE correction.
The bonding characteristics of the local minima were analyzed by means
of the Natural Bond Orbital (NBO) analysis of Weinhold et al.[42]
Synthesis of dinor-labdane alcohols 32 and 33: A sample of NaBH4
(50 mg, 1.31 mmol) was added to a solution of 29 (12 mg, 0.04 mmol) in
EtOH (5 mL), and was stirred at 08C for 1 h. The mixture was then dilut-
ed with tBuOMe, washed with brine, dried (anhydrous Na2SO4), and the
solvent removed. Flash chromatography (hexane/tBuOMe 3:7) of the res-
idue gave epimeric alcohols 32 (5.5 mg, 45%) and 33[27] (4.5 mg, 37%).
Data for 32: white solid; m.p. 130 1358C; 1H NMR (300 MHz, CDCl3): d
= 4.83 (brs, 1H), 4.54 (brs, 1H), 3.76 (m, 1H), 3.24 (dd, J = 11.6,
4.5 Hz, 1H), 2.39 (ddd, J = 12.8, 4.2, 2.5 Hz, 1H), 1.95 (dt, J = 12.5,
5.0 Hz, 1H), 1.85 0.80 (m, 12H), 1.16 (d, J = 6.1 Hz, 3H), 0.98 (s, 3H),
0.76 (s, 3H), 0.68 ppm (s, 3H); 13C NMR (75 MHz, CDCl3): in Table 1;
MS (70 eV, EI): m/z (%): 262 (1), 247 (1), 207 (4), 135 (100); HRMS
(EI): calcd for[C18H32O2 280.2402, found 280.2397.
Acknowledgement
1
This research was supported by the Spanish DirecciÛn General de Inves-
tigaciÛn CientÌfica y Tÿcnica (Project PB 98 1365). J. Justicia thanks the
Spanish Ministerio de Ciencia y Tecnologia, and A. Rosales and Juan L.
Oller-Lopez the Spanish Ministerio de Educacion Cultura y Deporte for
the grants enabling them to pursue these studies. We thank also to our
English colleague Dr. J. Trout for revising our English text.
Data for tricyclic isocopalane 35: White solid; m.p. 132 1358C; H NMR
(300 MHz, CDCl3): d = 4.84 (brs, 1H), 4.51 (brs, 1H), 4.34 (dd, J =
11.0, 3.6 Hz, 1H), 4.17 (dd, J = 11.0, 9.4 Hz, 1H), 3.21 (dd, J = 11.4,
4.8 Hz, 1H), 2.39 (brd, J = 12.7 Hz, 1H), 2.10 1.90 (m, 1H), 2.01 (s,
3H), 1.87 1.80 (m, 1H), 1.75 1.25 (m, 12H), 0.98 (s, 3H), 0.81 (s, 3H),
0.76 (s, 3H), 0.74 ppm (s, 3H); 13C NMR (75 MHz, CDCl3, DEPT): d =
171.52, (C), 146.55 (C), 107.13 (CH2), 78.84 (CH), 61.52 (CH2), 59.64
(CH), 55.27 (CH), 55.02 (CH), 40.68 (CH2), 39.14 (C), 38.89 (C), 38.56
(CH2), 37.57 (C), 37.45 (CH2), 28.05 (CH3), 27.33 (CH2), 22.49 (CH2),
21.20 (CH3), 18.70 (CH2), 16.37 (CH3), 16.06 (CH3), 15.37 (CH3); MS
(70 eV, EI): m/z (%): 288 (1), 207 (17), 189 (14), 93 ppm (100); HRMS
(FAB): calcd for C22H36O3Na 371.2562, found 371.2561.
[1] a) B. M. Trost, Science 1991, 254, 1471 1477; b) B. M. Trost, Angew.
Chem. 1995, 107, 285 307; Angew. Chem. Int. Ed. Engl. 1995, 34,
259 281; c) A. F¸rstner, Synlett 1999, 1523 1533; d) A. F¸rstner,
A. Leitner, Angew. Chem. 2003, 115, 320 323; Angew. Chem. Int.
Ed. 2003, 42, 308 311.
[2] a) I. Abe, M. Rohmer, G. D. Prestwich, Chem. Rev. 1993, 93, 2189
2206; b) 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.
Synthesis of saturated isocopalane 36:
A mixture of 35 (11 mg,
0.03 mmol) and 5% Pd/C (5 mg) in MeOH (5 mL) was stirred under H2
(1 atm) for 6 h. The mixture was filtered and the solvent removed from
the filtrate, giving 36 (8 mg, 73%) as a 3:2 mixture of epimers at C-13.
Titanocene-catalyzed cyclization of 2,3-oxidosqualene 38: Strictly deoxy-
genated THF (20 mL) was added to a mixture of [Cp2TiCl2] (58 mg,
0.23 mmol) and Mn dust (512 mg, 9.30 mmol) under an argon atmos-
[3] a) E. J. Corey, S. C. Virgil, J. Am. Chem. Soc. 1991, 113, 4025 4026;
b) E. J. Corey, S. C. Virgil, S. Sarshar, J. Am. Chem. Soc. 1991, 113,
1786
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 1778 1788

Titanocene-Catalyzed Synthesis of Terpenoids
1778 1788
8171 8172; c) E. J. Corey, S. C. Virgil, H. Cheng, C. H. Baker,
S. P. T. Matsuda, V. Singh, S. Sarshar, J. Am. Chem. Soc. 1995, 117,
11819 11820; d) E. J. Corey, H. Cheng, Tetrahedron Lett. 1996, 37,
2709 2712; e) C. Jenson, W. L. Jorgensen, J. Am. Chem. Soc. 1997,
119, 10846 10854; f) B. A. Hess Jr. , J. Am Chem. Soc. 2002, 124,
10286 10287; g) B. A. Hess Jr. , Org. Lett. 2003, 5, 165 167; h) D.
Gao, Y.-K. Pan, J. Am. Chem. Soc. 1998, 120, 4045 4046
For a review on this item see: c) A. F¸rstner, Chem. Eur. J. 1998, 4,
567 570.
[16] E. E. van Tamelen, A. Storni, E. J. Hessler, M. Schwartz, J. Am.
Chem. Soc. 1963, 85, 3295 3297.
[17] In contrast with carbocationic cyclizations, in which mainly endocy-
clic double bonds are formed,[10a, 30] both stoichiometric and catalytic
versions of TiIII-based cyclization show a high regioselectivity to-
wards exocyclic alkenes. This preference might be related to the
free rotation of methyl groups (C-12 in 9), which would allow a syn
disposition between hydrogen and titanium atoms and thus facilitate
the elimination of [Cp2Ti(Cl)H].
[18] J. A. Marco, J. F. Sanz-Cervera, M. D. Morante, V. Garcia-Lliso, J.
Vallÿs-Xirau, J. Jakupovic, Phytochemistry 1996, 41, 837 844.
[19] E. Marcantoni, F. Nobili, G. Bartoli, M. Bosco, L. Sambri, J. Org.
Chem. 1997, 62, 4183 4184.
[20] J. P. Uttaro, G. Audran, E. Palombo, H. Monti, J. Org. Chem. 2003,
68, 5407 5410.
[21] A. F. Barrero, J. M. Cuerva, E. J. çlvarez-Manzaneda, J. E. Oltra, R.
Chahboun, Tetrahedron Lett. 2002, 43, 2793 2796.
[22] a) J. D. Connolly, R. A. Hill, Dictionary of Terpenoids, Vol. 1, Chap-
man and Hall, London, 1991, pp. 453 462; b) B. M. Fraga, Nat.
Prod. Rep. 2002, 19, 650 672 and previous issues in this series.
[23] W. F. Fleck, B. Schlegel, P. Hoffmann, M. Ritzau, S. Heinze, U.
Grafe, J. Nat. Prod. 1996, 59, 780 781.
[4] For a seminal work, see: a) D. J. Goldsmith, J. Am. Chem. Soc. 1962,
84, 3913 3918. For some reviews, see: b) E. E. van Tamelen, Acc.
Chem. Res. 1975, 8, 152 158; c) S. K. Taylor, Org. Prep. Proced. Int.
1992, 24, 247 284; d) S. T. Dennison, D. C. Harrowven, J. Chem.
Educ. 1996, 73, 697 701; e) C. M. Marson, Tetrahedron 2000, 56,
8779 8793. For recent selected examples, see: f) A. X. Huang, Z.
Xiong, E. J. Corey, J. Am. Chem. Soc. 1999, 121, 9999 10003; g) J.
Zhang, E. J. Corey, Org. Lett. 2001, 3, 3215 3216; h) M. Yuan, J. V.
Schreiber, E. J. Corey, J. Am. Chem. Soc. 2002, 124, 11290 11291.
[5] a) R. Breslow, E. Barrett, E. Mohacsi, Tetrahedron Lett. 1962, 3,
1207 1211; b) R. Breslow, S. S. Olin, J. T. Groves, Tetrahedron Lett.
1968, 1837 1840; c) J. Y. Lallemand, M. Julia, D. Mansuy, Tetrahe-
dron Lett. 1973, 14, 4461 4464.
[6] For a recent review on the synthesis of polycyclic compounds by
means of radical cascade reactions see: a) A. L. Dhimane, L. Fen-
sterbank, M. Malacria in Radicals in Organic Synthesis, Vol. 2 (Eds.:
P. Renaud, M. P. Sibi), Wiley-VCH, Weinheim, Germany, 2001, pp.
350 382. For selected references see: b) S. Handa, G. Pattenden, J.
Chem. Soc. Perkin Trans. 1 1999, 843 844.; c) U. Hoffmann, Y. Gao,
B. Pandey, S. Klinge, K.-D. Warzecha, C. Kr¸ger, H. D. Roth, M.
Demuth, J. Am. Chem. Soc. 1993, 115, 10358 10359; d) B. B.
Snyder, J. Y. Kiselgoc, B. M. Foxman, J. Org. Chem. 1998, 63, 7945
7952; e) P. A. Zoretic, H. Fang, A. A. Ribeiro J. Org. Chem. 1998,
63, 4779 4785.
[7] a) W. A. Nugent, T. V. RajanBabu, J. Am. Chem. Soc. 1988, 110,
8561 8562; b) T. V. RajanBabu, W. A. Nugent, J. Am. Chem. Soc.
1994, 116, 986 997; c) A. Gans‰uer, M. Pierobon, H. Bluhm,
Angew. Chem. 1998, 110, 107 109; Angew. Chem. Int. Ed. 1998, 37,
101 103; d) A. Gans‰uer, H. Bluhm, M. Pierobon, J. Am. Chem.
Soc. 1998, 120, 12849 12859; e) A. Gans‰uer, M. Pierobon, H.
Bluhm, Synthesis 2001, 2500 2520; f) A. Gans‰uer, B. Rinker, Tetra-
hedron 2002, 58, 7017 7026; g) A. Gans‰uer, H. Bluhm, B. Rinker,
S. Narayan, M. Schick, T. Lauterbach, M. Pierobon, Chem. Eur. J.
2003, 9, 531 542.
[24] K. Munesada, H. L. Siddiqui, T. Suga, Phytochemistry 1992, 31,
1533 1536.
[25] Farnesylacetone was bought from Fluka in the form of a mixture of
four stereoisomers containing 31% of the all-trans isomer (GC-MS
analysis with an authentic sample as standard). This mixture was
used as received because it is easier to isolate cyclic product 28 than
the stereoisomers of the starting material. Nevertheless, as it is
known that cis isomers do not undergo cyclization under our experi-
mental conditions (see the case of epoxyneryl acetate in ref. [8]), the
yields are based on the all-trans isomer contained in the starting
mixture.
[26] H. Monti, N. Tiliacos, R. Faure, Phytochemistry 1996, 42, 1653
1656.
[27] H. Monti, N. Tiliacos, R. Faure, Phytochemistry 1999, 51, 1013
1015.
[28] a) S. J. White, R. S. Jacobs, Mol. Pharmacol. 1983, 24, 500 508;
b) E. T. O’Brien, D. J. Asai, R. S. Jacobs, L. Wilson, Mol. Pharmacol.
1989, 35, 635 642.
[29] a) X. Xing, M. Demuth, Synlett 1999, 987 990; b) X. Xing, M.
Demuth, Eur. J. Org. Chem. 2001, 537 544.
[30] E. E. van Tamelen, R. C. Nadeau, J. Am. Chem. Soc. 1967, 89, 176
177.
[31] a) E. E. van Tamelen, J. Willet, M. Schwartz, R. Nadeau, J. Am.
Chem. Soc. 1966, 88, 5937 5938; b) K. B. Sharpless, E. E. van Tame-
len, J. Am. Chem. Soc. 1969, 91, 1848 1849.
[32] Both 3b-hydroxymalabarica-14(26), 17E, 21-triene (39) and the cor-
responding acetate are natural metabolites from plants. NMR data
of synthetic 39 and its acetyl derivative were in accordance with
those of the natural products, confirming the structures proposed on
the basis of chemical degradation and spectroscopic techniques; see:
a) J. Jakupovic, F. Eid, F. Bohlmann, S. El-Dahmy, Phytochemistry
1987, 26, 1536 1538; b) F. J. Marner, A. Freyer, J. Lex, Phytochemis-
try 1991, 30, 3709 3712.
[33] A further 5-endo cyclization of radical 45 is disfavored by Baldwin×s
rules; see: a) J. E. Baldwin, J. Chem. Soc. Chem. Commun. 1976,
734 736; b) J. E. Baldwin, J. Cutting, W. Dupont, L. Kruse, L. Sil-
berman, R. C. Thomas, J. Chem. Soc. Chem. Commun. 1976, 736
738.
[8] A. F. Barrero; J. M. Cuerva, M. M. Herrador, M. V. Valdivia, J. Org.
Chem. 2001, 66, 4074 4078.
[9] Titanocene(iii) is generated in situ as the active form [Cp2TiIIICl] by
stirring commercially available [Cp2TiIVCl2] and Mn dust; see: a) D.
Sekutowski, R. Jungst, G. D. Stucky, Inorg. Chem. 1978, 17, 1848
1855; b) R. J. Enemaerke, G. H. Hjollund, K. Daasbjerg, T. Skrydstr-
up, C. R. Acad. Sci. Ser II 2001, 4, 435 438.
[10] a) A. F. Barrero, J. E. Oltra, J. M. Cuerva, A. Rosales, J. Org. Chem.
2002, 67, 2566 2571; b) A. F. Barrero, A. Rosales, J. M. Cuerva,
J. E. Oltra, Org. Lett. 2003, 5, 1935 1938.
[11] GC-MS analysis indicated a 77% deuterium incorporation.
[12] Despite of the free-radical nature of TiIII mediated epoxide open-
ings, the water-dependent control of termination steps seems to be a
general feature when tertiary radicals are involved in this process;
see ref. [10].
[13] D. P. Curran, N. A. Porter, B. Giese, Stereochemistry of Radical Re-
actions, VCH, Weinheim, 1995.
[14] In initial experiments using an excess of titanocene (ref. [8]) dilu-
tions levels to the order of 10ꢀ3 m were needed to avoid increased
proportions of byproducts such as 4 and 7, derived from the prema-
ture trapping of intermediate radicals such as 2 and 5.
[15] The use of chlorosilanes alone, originally described by F¸rstner and
Hupperts to regenerate titanium catalysts in McMurry-type reac-
tions, is not suitable when epoxypolyenes are involved in the reac-
tion, because, under treatment with acidic reagents, these substrates
undergo carbocationic cyclization instead of the desired radical
process. For more detailed discussions about this subject see
refs. [7c] and [10]. For the original reports on the use of chlorosi-
lanes to regenerate titanium and chromium catalysts see: a) A.
F¸rstner, A. Hupperts, J. Am. Chem. Soc. 1995, 117, 4468 4475;
b) A. F¸rstner, N. Shi, J. Am. Chem. Soc. 1996, 118, 12349 12357.
[34] A. Behrens, P. Schaeffer, S. Bernasconi, P. Albrecht, Org. Geochem.
1999, 30, 379 383; b) S. Schouten, M. J. L. Hoefs, J. S. Sinninghe
Damste, Org. Geochem. 2000, 31, 509 521; c) J. P. Werne, D. J. Hol-
lander, A. Behrens, P. Schaeffer, P. Albrecht, J. S. Sinninghe Damste,
Geochim. Cosmochim. Acta 2000, 64, 1741 1751.
[35] A. F. Barrero, E. J. Alvarez-Manzaneda, P. L. Palomino, Tetrahedron
1994, 50, 13239 13250.
[36] A. F. Barrero, E. J. Alvarez-Manzaneda, R. Alvarez-Manzaneda,
Tetrahedron Lett. 1989, 30, 3351 3352.
[37] See Supporting Information.
1787
Chem. Eur. J. 2004, 10, 1778 1788
www.chemeurj.org
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

D. J. Cµrdenas*, J. M. Cuerva et al.
FULL PAPER
[38] J. B. Arterburn, M. C. Perry, Org. Lett. 1999, 1, 769 771.
[39] A. S. Gopalan, R. Prieto, B. Mueller, D. Peters, Tetrahedron Lett.
1992, 33, 1679 1682.
Liashenko, P. Piskorz, I. Komaromi, R. Gomperts, R. L. Martin,
D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng, A. Nanayakkara,
C. Gonzalez, M. Challacombe, P. M. W. Gill, B. G. Johnson, W.
Chen, M. W. Wong, J. L. Andres, M. Head-Gordon, E. S. Replogle,
J. A. Pople, Gaussian, Inc., Pittsburgh, PA, 1998.
[40] Gaussian 98 (Revision A.7), M. J. Frisch, G. W. Trucks, H. B. Schle-
gel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, V. G. Zakrzewski,
J. A. Montgomery, Jr., R. E. Stratmann, J. C. Burant, S. Dapprich,
J. M. Millam, A. D. Daniels, K. N. Kudin, M. C. Strain, O. Farkas, J.
Tomasi, V. Barone, M. Cossi, R. Cammi, B. Mennucci, C. Pomelli,
C. Adamo, S. Clifford, J. Ochterski, G. A. Petersson, P. Y. Ayala, Q.
Cui, K. Morokuma, D. K. Malick, A. D. Rabuck, K. Raghavachari,
J. B. Foresman, J. Cioslowski, J. V. Ortiz, B. B. Stefanov, G. Liu, A.
[41] a) A. D. Becke, Phys. Rev. A 1988, 38, 3098 3100; b) C. Lee, W.
Yang, R. G. Parr, Phys. Rev. B 1988, 37, 785 789.
[42] A. E. Reed, L. A. Curtiss, F. Weinhold, Chem. Rev. 1988, 88, 899
926.
Received: October 21, 2003 [F5647]
1788
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 1778 1788