Cascading pericyclic reactions: Building complex carbon frameworks for natural product synthesis

Article abstract of DOI:10.1039/b700054p

Tandem reactions have emerged as powerful strategies to synthesize complex structures, in particular, processes involving pericyclic reactions. This article describes recent advancement by our group in the development of novel tandem pericyclic reactions

Full text of DOI:10.1039/b700054p

View Online / Journal Homepage / Table of Contents for this issue
FEATURE ARTICLE
www.rsc.org/chemcomm | ChemComm
Cascading pericyclic reactions: building complex carbon frameworks for
natural product synthesis
Steve Arns and Louis Barriault*
Received (in Cambridge, UK) 2nd January 2007, Accepted 12th January 2007
First published as an Advance Article on the web 19th February 2007
DOI: 10.1039/b700054p
Tandem reactions have emerged as powerful strategies to synthesize complex structures, in
particular, processes involving pericyclic reactions. This article describes recent advancement by
our group in the development of novel tandem pericyclic reactions aimed at constructing
diterpene frameworks.
Introduction
of this sort quickly assemble non-trivial carbon skeletons, they
also offer an atom-economy not observed with single-step
processes, inherent to this also being a notable minimization of
waste produced in the given process.
The quest to develop methods for the assembly of complex
carbon frameworks quickly and efficiently has long been a
challenge and goal for synthetic organic chemists. One of the
driving forces for the development of these methods is their
application to the total synthesis of natural products, the
majority of which have a carbon skeleton which forms the
backbone of the molecule.
The use of cascading reactions in organic chemistry, either
by chance or design, is not a new concept, and many excellent
1
reviews have been published on the subject. Despite the
abundance of reaction combinations, pericyclic reactions
occurring in cascade constitute one the most fascinating
transformations. One might consider that the concerted nature
of pericyclic reactions render them particularly attractive due
to the relative ease of predicting the stereochemical outcome of
the final product. Undoubtedly, this represents a true
advantage when considering the elaboration of complex
carbon frameworks bearing multiple stereogenic centers. For
example, Nicolaou et al.’s synthesis of endiandric acid B (3)
constitutes probably one of the most spectacular tandem
pericyclic reaction transformations reported in the literature
Recent years have seen the emergence of tandem reactions
as a powerful tool to construct complicated molecular
scaffolds. In the literature there are reports involving the
terms tandem, domino or cascade reactions where the
reactions were performed separately. In our opinion, these
are not tandem processes but rather sequential reactions.
Tandem or domino reactions can be defined as two or more
distinct chemical transformations that occur in quick succes-
sion in the same reaction vessel. If the reactions involve a
sequential addition of reagents, part of these reagents must be
embedded in the final product in order to be considered a
tandem reaction. For instance, the 1,4-addition of an
organocuprate to an enone followed by the enolate trapping
with an electrophile to give the corresponding a,b-substituted
ketone is considered a tandem reaction. Not only can reactions
2
,3
(Scheme 1).
Upon heating, polyene 1 was converted to
endiandric B methyl ester 2 in a single operation.
In the past decade, tandem pericyclic reactions have gained
in popularity, particularly in the biomimetic synthesis of
1
a
complex natural diterpenes.
In fact, numerous clever
transformations were reported, demonstrating the power and
the usefulness of these tandem processes. This account will
outline the recent progress in this field, having a focus on our
contributions that we have made since 2000 at the University
of Ottawa.
Department of Chemistry, 10 Marie Curie, University of Ottawa,
Ottawa, Ontario, Canada K1N 6N5. E-mail: lbarriau@uottawa.ca;
Fax: +1-613-562-5170; Tel: +1-613-562-5800
Steve Arns was born in St. John’s, Newfoundland, Canada, in
1
(2)-polycavernoside A. In May 1999, he accepted a position as
Assistant Professor at the University of Ottawa where he has
been promoted to Associate Professor (2003). Professor
Barriault’s research involves the development of novel strategies
using tandem pericyclic reactions to construct complex bioactive
natural products. Recently, Professor Barriault received the John
Polanyi Award in Chemistry (2000), Premier’s Research
Excellence Award (2002), Ottawa Life Science Michael Smith
Award (2002), Boehringer Ingelheim Young Investigator Award
(2002), University of Ottawa young researcher of the year
(2005), University of Ottawa excellence in education award
(2005), Merck Research Laboratory academic development
award (2006) and was appointed University Research Chair in
Synthetic Organic Chemistry (2006).
980 and received his BSc (chemistry) in 2003 at Memorial
University of Newfoundland. Since then he has been working
under the guidance of Professor Louis Barriault, engaged in the
synthesis of several natural products using the oxy-Cope–
Claisen–ene reaction. He has been awarded both NSERC
PGS-A and PGS-D postgraduate scholarships.
Louis Barriault was born in Armagh, Qu e´ bec, Canada, in 1970.
He received his BSc (chemistry) in 1993 at the University of
Sherbrooke. He pursued his PhD studies under the supervision of
Professor Pierre Deslongchamps at the same institution. After
completing his doctorate in 1997, he joined the group of Professor
Leo A. Paquette at the Ohio State University (OSU) as a FCAR
postdoctoral fellow where he completed the total synthesis of
This journal is ß The Royal Society of Chemistry 2007
Chem. Commun., 2007, 2211–2221 | 2211

View Online
Only a trace of desired compound 9 was isolated. This tandem
process leading to 11 was also observed in other cases as an
7
unwanted side reaction in both thermal and anionic oxy-
8
Cope reactions. We thought, using the appropriate starting
1
,2-divinylcyclohexanols, that we would be able to take
advantage of this process to specifically build decalin systems
containing a tertiary alcohol at the ring junction. This type of
framework is embedded in various natural sesqui- and
diterpenes.
Scheme 1 Synthesis of endiandric acid B.
Results and discussion
Heating the 1,2-divinylcyclohexanols 12–20, readily pre-
pared in a few steps from commercially available starting
material, in a sealed tube in toluene for 5 h at 220 uC yielded a
variety of cyclic compounds in moderate to good yields
Our group has long been interested in using various cascading
pericyclic reactions to access complicated carbon frameworks.
Our long term goal is to use these processes to fabricate
various carbon skeletons which correspond to the cores of
various natural products, culminating in their total synthesis.
In addition to this, through both experiment and calculation,
we strive to understand these reactions on a mechanistic level,
enabling us to accurately predict their outcome and explain the
formation of the compounds we observe as products.
(
Scheme 3). However, in some cases we isolated retroene by-
products 24, 26 and 28. We found that increasing electron
density in the allylic double bond favors the retroene manifold
to become the main pathway up to the point where it is the
only product observed, such as in the case of 1,2-divinylcy-
clohexanol 15. This is in agreement with the results reported by
9
Paquette, Houk et al.
In each case where the oxy-Cope–ene product was isolated,
only a single diastereomer was observed. This highly selective
process can be rationalized by first assuming that the initial
1
. Oxy-Cope–transannular ene reaction
Our first foray into investigating tandem pericyclic processes
was centered on the oxy-Cope–transannular ene reaction of
4
,2-divinylcyclohexanols (eqn (1)). We imagined that alcohol
thermal oxy-Cope reaction occurs through a chair-like
1
transition state, as in the conversion of A to B (Scheme 4).
The latter then undergoes a stereospecific tautomerization to
give C. At this point, ketone C can undergo a chair-like
4, upon heating, would be transformed into the corresponding
decalin framework 6. We expected that this tandem process
would be highly diastereoselective on the basis of preferential
conformations of macrocycle 5 at the transition state of the ene
reaction. Our exploitation of this process was not without
precedent. There are very few examples in the literature of the
5
tandem oxy-Cope–ene reaction. To the best of our knowl-
edge, the first example of this tandem process was reported by
6
Sworin and Lin (Scheme 2).
ð1Þ
During the course of their investigation of tandem oxy-
Cope–S
N
29 reactions to create hydroazulenoid scaffolds, they
observed that heating of 7 at 204 uC gave mainly macrocycle
1
0 and decalin 11 (from 10) in 49% and 20% yields respectively.
Scheme 2 Attempt to create hydroazulenoid scaffolds.
212 | Chem. Commun., 2007, 2211–2221
Scheme 3 Tandem oxy-Cope–ene reaction of 1,2-divinylcyclohexanols.
This journal is ß The Royal Society of Chemistry 2007
2

View Online
Table 1 DBU and NMP assisted oxy-Cope–ene reactions
Additive
(equivalents)
Entry
Alcohol
Products
Ratio (yield)
1
2
3
4
5
a
15
15
15
16
16
DBU (0.5)
DBU (50)
25, 26
25, 26
27, 28
27, 28
27, 28
1 : 7.9 (80%)
1 : 3 (89%)
11 : 1 (18%)
20 : 1 (75%)
.25 : 1 (98%)
NMP (solvent)
DBU (20)
DBU (60)
a
Microwave irradiation for 40 min at 220 uC.
Scheme 4 Proposed mechanism for the tandem oxy-Cope–ene
reaction.
1
0
transannular ene reaction to afford compound E or do a ring
flip (D) followed by an ene reaction to provide F.
Assuming a rapid equilibrium between conformers C and D,
the product ratio will be determined by the energy difference
1
1
between TS1 and TS2. In the case of cyclic allylic alcohols 14
and 16, TS1 is favored over TS2 due to the presence of severe
1
,3-diaxial interactions, methylene–methyl, in TS2, thereby
giving tricycles 23 and 27 as single diastereomers. In the case of
the tandem oxy-Cope–ene reaction of 1,2-divinylcyclohexanol
1
3, only decalin 22 was isolated after heating at 220 uC for 5 h.
This indicates that the transannular ene reaction proceeds
through TS2 where the methyl group is oriented in the pseudo-
equatorial position. In the case of a cyclic ene donor such as
1
7–20, no macrocyclic ring flip (C A D) is possible. This
Scheme 5 Tandem oxy-Cope–ene reactions of 1-alkynyl-2-vinylcy-
ensures the stereoselective formation of tricycles 29–32.
Although the diastereoselectivity of the process was excep-
tional for generating decalins and tricycles, we were still faced
with the problem of the retroene reaction manifold competing
with the desired oxy-Cope–ene pathway.
clohexanols.
2
. Oxy-Cope–ene reaction in the synthesis of (+)-arteannuin M
(
43)
Having developed a reliable and predictable reaction cascade
for generating trans-decalin ring systems, we then set our sights
towards applying this methodology in a total synthesis. Of
particular interest were natural products containing a trans-
decalin ring system with a tertiary alcohol at one of the ring
junctions. (+)-Arteannuin M (43), a natural product isolated
Literature precedent suggested the rate of the oxy-Cope
1
2
rearrangement can be accelerated by additive solvents. It was
postulated that hydrogen bonding of solvent to the hydroxyl
proton increases the electron density on the oxygen, allowing
for an ‘‘electron push’’ that accelerates the C–C bond breaking
in the [3,3] sigmatropic rearrangement. In light of this
observation, we scanned various solvent additives hoping to
find something that would increase the rate of the oxy-Cope
rearrangement sufficiently so as to suppress the retroene
reaction. To our delight, we discovered that the use of DBU as
an additive suppressed or diminished isolation of retroene
1
5
from Artemisia annua L (Fig. 1). proved to be the perfect
target, containing all the desired structural features we wished
to construct using our cascading sequence. It is important to
point out that the absolute configuration of the molecule and
the stereochemistry at C9 were unknown.
1
3
reaction products (Table 1).
Finally, we demonstrated that the tandem oxy-Cope–ene
reaction can be performed on 1-alkynyl-2-vinylcyclohexanols
(
Scheme 5). The conversion of 33–37 to their respective dienes
8–42 was accomplished by microwave irradiation in the
3
1
4
presence of 5–20 equivalents of DBU in toluene. In all cases,
the desired product was isolated as the sole diastereomer in fair
to good yield.
Fig. 1 Structure of (2)-arteannuin M (43).
Chem. Commun., 2007, 2211–2221 | 2213
This journal is ß The Royal Society of Chemistry 2007

View Online
Scheme 6 Synthesis of ent-arteannuin M (43).
Our synthesis commenced with the conversion of limonene
1
4 (ee . 98%) to 1,2-divinylcyclohexanol 45 (Scheme 6).
6
4
Thermal rearrangement of 45 gave 46 in 60% yield as the sole
diastereomer with an enantiomeric excess of 78%. The high
diastereoselectivity of the process is rationalized using the
mechanism depicted in Scheme 7. The process is triggered by a
Scheme 8 Planar chirality.
. Conservation of planar chiral information in the oxy-Cope–
3
ene reaction
[
3,3] sigmatropic rearrangement of 45 to afford enol 47 which
after tautomerization led to 48. The latter is poised to undergo
a transannular ene reaction through transition states A and B.
A close examination of transition states A and B reveals that
the alkyl group in A is oriented pseudo-axially, whereas in B
the alkyl chain occupies a pseudo-equatorial position. On the
basis of this observation, transition state B is favored over A,
explaining why we observed 46 as the sole product.
Intrigued by the enantioselectivity leakage in the tandem
process, we initiated a more in-depth investigation into the
diastereo- and enantioselectivity of the tandem reaction. This
began with the preparation of several optically pure 1,2-
divinylcyclohexanols 50a–e, which were then subjected to
1
8
microwave irradiation at 220 uC for one hour to give decalins
51a–e and 52a–e (Scheme 9).
1
9
The loss of enantiopurity can be accounted for by a ring flip
of 47 to give ent-47 (Scheme 8). We initially proposed that the
chirality present in 45 would be transferred, by virtue of the
We scanned a variety of basic additives in an effort to
evaluate their effects on the enantioselectivity. We found that
in the majority of cases we isolated decalins 51 and 52 with
ratios ranging from 8.5 : 1 to 25 : 1, regardless of the additive,
in moderate to excellent yields. In addition, the majority of the
1
7
planar chirality of the intermediate macrocyclic enol 47,
through to the final product. We thought that the three double
bonds embedded in enol intermediate 47 would create a rigid
structure that would prevent the enol moiety from rotating and
thereby producing ent-47. However, we observed in our
substrate that such an inversion is feasible, hence accounting
for the loss of enantiopurity. Of course, ent-47 leads ultimately
to ent-46 via the same manifold by which 47 leads to 46. This
synthesis firmly established the use of the oxy-Cope–ene
reaction as a viable synthetic method for generating decalin
ring systems.
2
0
bases used resulted in essentially no erosion of enantiopurity.
However, DBU represents an exception. Heating of 50a–e in
the presence of 10 equivalents of DBU gave the corresponding
decalins 51a–e having enantiomeric excesses of 93%, 35%,
70%, 74% and 84% respectively. We discovered that an
increase in the DBU proportion was directly related to a
lowering of enantiomeric excess. Interestingly, when we
2
1
performed the anionic oxy-Cope rearrangement this resulted
in a complete racemization of the substrate.
This suggests that enol 53 possesses a high energy barrier to
invert to ent-53, whereas enolate 55 has a lower energy barrier
for ring inversion (Scheme 10). The ring inversion can also
compete with the tautomerization process which after the
transannular ene reaction leads to ent-51. Therefore, when a
weak base or no strong base is present in the reaction mixture,
a complete conservation of the planar chirality information
through the process is observed. However, when a portion of
Scheme 7 Proposed mechanism for the tandem oxy-Cope–ene
reaction.
Scheme 9 Enantioselective tandem oxy-Cope–ene reaction.
This journal is ß The Royal Society of Chemistry 2007
2
214 | Chem. Commun., 2007, 2211–2221

View Online
Scheme 12 Tandem oxy-Cope–ene–Claisen.
Scheme 10 Proposed mechanism for the tandem oxy-Cope–ene
Claisen rearrangement, to the sequence (Scheme 12). This would
give access to one or more quaternary carbon centers in the final
product, depending on the substitution pattern of the allyl ether
moiety. Acting on this possibility, we constructed a variety of
simple allyl-1,2-divinylcyclohexanols 58 which reacted through
a highly diastereoselective oxy-Cope–ene–Claisen cascade. This
gives products of type 59 with an all-carbon quaternary center
reaction.
2
3
now embedded in the decalin system. As shown in Scheme 12,
thermal rearrangement of an allyl ether monosubstituted at the
terminus position afforded the corresponding lactol with high
diastereoselectivity.
Scheme 11 Proposed mechanism for the tandem oxy-Cope–ene
After the oxy-Cope rearrangement and tautomerization of
58, the ketone intermediate can undergo a transannular ene
reaction through two transition states A or B (Scheme 13). A
closer examination of transition state B reveals the presence of
a pseudo 1,3-diaxial interaction between the allyl ether group
and the macrocyclic ring. Consequently, this favors transition
state A over B as the reactive conformer to provide enol ether
60. Finally, the latter undergoes a Claisen rearrangement on
the opposite side of the tertiary alcohol to afford 59 as the sole
isomer. The stereoselectivity of the process is thus explained.
We also found that propargyl ethers 63 were ideal substrates
for the tandem oxy-Cope–ene–Claisen rearrangement, giving
products 64 or 65 in excellent yields and with high diastereo-
selectivity, as illustrated in Scheme 14. Allene products 64 were
favored in the cases where R was an alkyl group or a proton;
whereas cyclic ethers 65 were favored when R was an electron-
withdrawing or aryl substituent. Compound 65 arises from a
5-exo-dig cyclization onto the previously generated allene 64.
reaction.
the enol 53 is converted to enolate 55, we observe a leak in ee.
On the basis of the results reported above, this strongly
suggests that the pK of the proton on the enol 53 combined
a
with the strength of the base used are directly responsible for
the loss of chirality.
We also found that the electronic nature of the R group
affected the diastereomeric ratio of products 51 and 52. While
most substituents still favor products 51 greatly over type 52, it
was observed that electron-withdrawing substituents lower the
ratio of 51 : 52. For instance, when R = CF3 (50c), the
diastereomeric ratio fell to 8.5 : 1 favoring 51c. However, when
R = OEt (50d), the formation of 52d was slightly favored over
5
1d in a ratio of 1 : 2.3.
As demonstrated above, the diastereoselectivity of the
transannular ene reaction is usually governed by the Curtin–
Hammett principle. This is true when electron-donating or
neutral groups (R) are adjacent to the carbonyl moiety on 56
(
Scheme 11). On the other hand, it seems unlikely that the ene
reaction is controlled by the Curtin–Hammett principle when
R is an electron-withdrawing group. One might propose that
the ring inversion (56 A 57) competes with the ene reaction
since the latter reaction is accelerated by electron-withdrawing
2
2
groups such as R = CF
3
and OEt.
4
. The oxy-Cope–ene–Claisen cascade
One obvious limitation of the oxy-Cope–ene sequence resides
in the inability to install all-carbon quaternary centers on a
decalin system. To remedy this problem, we imagined that the
addition of an allyloxy ether fragment to the isopropenyl unit
would result in the addition of a third pericyclic reaction, a
Scheme 13 Mechanism of the tandem oxy-Cope–ene–Claisen.
Chem. Commun., 2007, 2211–2221 | 2215
This journal is ß The Royal Society of Chemistry 2007

View Online
Scheme 14 Tandem oxy-Cope–ene–Claisen reaction of propargyl
ethers.
This is the result of a thermodynamically controlled cyclization
where the least strained 5-membered-ring acetal is formed.
Having realized the potential of this tandem sequence, we
sought to install two contiguous quaternary carbon centers in
the molecule. This could be achieved by thermal rearrange-
ment of 58 possessing an allyl ether disubstituted at the distal
1 2
end of the alkene (where R and R ? H). Much to our
surprise, submitting compound 66 to the standard conditions
did not give the expected oxy-Cope–ene–Claisen product, but
compound 69 in 75% yield with an E : Z ratio of the newly
2
4
formed olefin of 89 : 11 (Scheme 15). Microwave irradiation
of compounds 67 and 68 resulted in the formation of
structurally related products 70 and 72 respectively. We also
isolated unreacted E olefin 71 in the reaction of 67, and dimer
Scheme 15 Tandem oxy-Cope–ene–Claisen reaction of allyl ethers.
7
3 in the reaction of 68.
These results suggest that there is a radical [1,3]-shift in
competes with the rate of recombination. This is evident by the
fact that different E : Z ratios are obtained for 69 and 70 which
should have the same allyl radical intermediate. Moreover, the
combination of two of the allyl radicals 74 or 76 explains the
formation of dimer 73.
competition with the [3,3] Claisen rearrangement. Incidentally,
the isolation of the E olefin 71 serves to confirm the high
selectivity of the transannular ene portion of the reaction
cascade (vide supra).
It has been previously suggested that the Claisen rearrange-
ment could proceed through the formation of an oxallyl–allyl
radical pair instead of the concerted mechanism, especially
In summary, we have observed that disubstituted allyl ether
starting materials in the oxy-Cope–ene–Claisen rearrangement
proceed through the expected reaction pathway. However, a
different pathway is manifested in the case of more sterically
demanding trisubstituted allyl ethers for the Claisen step of the
cascade. In this case, a [1,3]-shift radical pathway is favored
over the congested [3,3] rearrangement.
2
5
when there are radical-stabilizing substituents present. In our
situation, the intermediate after the oxy-Cope–ene portion of
the cascade 60 undergoes homolytic C–O bond cleavage to
afford 74 and 75 (Scheme 16). Allyl radical 74 can undergo an
olefin isomerization to afford 76, both of which can recombine
with rearranged alkyl radical 77 to give the formal [1,3]-shift
product 78.
5
. Oxy-Cope–ene–Claisen in total synthesis
The scrambling of the geometry in 78 is accounted for by the
isomerization of 74 to 76 where the rate of isomerization
Having undertaken an extensive study of the oxy-Cope–ene–
Claisen rearrangement and having devised a predictable
Scheme 16 Mechanism of the 1,3-radical shift.
2
216 | Chem. Commun., 2007, 2211–2221
This journal is ß The Royal Society of Chemistry 2007

View Online
Scheme 17 Synthesis of tetrodecamycin.
method for the generation of decalin systems containing a
quaternary carbon, we sought to apply our method in total
synthesis.
Scheme 18 Tandem oxy-Cope–Claisen–ene.
Tetrodecamycin 79, an antibiotic isolated from
depend on the conformational preferences of the macrocyclic
intermediates 89 and 90 at the transition states.
2
6
Streptomyces nashvillensis MJ885-mF8, is a structurally
intriguing natural product containing a trans-decalin and a
tetronic acid moiety as its major structural features
In pursuit of making this cascading reaction viable, a variety
of precursors were synthesized through several routes, giving a
series of structurally related compounds designated 91, with
(
Scheme 17). Its biological activity against several Gram-
positive bacteria has resulted in multiple synthetic pursuits, but
1 2 3 4
variation throughout R , R , R and R . As we expected,
2
7
no total synthesis has yet been reported. Our approach to
this compound involved the two disconnections superimposed
on 79, revealing decalin 81 and tetronic acid residue 80 as the
subjection of the majority of these compounds to thermal
rearrangement conditions resulted in the formation of decalin
products 92 in very good yields and excellent diastereo-
selectivity, as indicated with the several illustrative examples
2
8
major constituents we needed to construct.
3
2
From cyclohexene oxide 82, the precursor for the pivotal
oxy-Cope–ene–Claisen rearrangement 83 was readily available
in five steps. As expected, decalin 84, containing the
quaternary center found in the natural product, was isolated
after treatment of 83 with microwaves at 220 uC. This was then
elaborated to several precursors that we deemed appropriate as
intermediates for the coupling with tetronic acid residues of
type 80. Whilst we have yet to complete the total synthesis, we
have been successful in synthesizing the decalin cores 81a and
shown in Scheme 19.
The excellent diastereoselectivity of this process can be
explained by taking a closer look at the reaction mechanism,
where we find the diastereoselection is ultimately dictated by
the conformational preferences of intermediates 99 and 101
3
3
(Scheme 20). The oxy-Cope rearrangement of 91 through a
chair-like transition state affords intermediate 99. At this
point, two things can happen: it can undergo a ring inversion
to 100, or rearrange to 101 via the Claisen reaction. We found
that the rate of the ring inversion is slow compared to that of
the Claisen rearrangement, leading to exclusive formation of
8
1b of tetrodecamycin along with coupling partner 80. Efforts
are ongoing towards the completion of the synthesis.
6
. The oxy-Cope–Claisen–ene reaction
One of the problems encountered in the synthesis of polycyclic
compounds is the stereoselective formation of quaternary
2
9
carbon centers. As an example, bioactive natural diterpenes
3
such as LL-S491b (85) and myrocin C (86) possess a
0
31
quaternary center at C9 and a tertiary alcohol at C10
(
Scheme 18). The retrosynthetic analysis of these particular
molecules reveals that the main framework can be generated
from a common decalin intermediate 88, which bears the
requisite quaternary carbon center at C9 adjacent to a tertiary
alcohol at C10. We envisaged the formation of 88 via a thermal
cyclization of the allylic ether 87. Heating of the allyl ether 87
produces the enol intermediate 89 via an oxy-Cope reaction.
The latter is transformed into ketone 90 through a Claisen
rearrangement. The resulting macrocyclic ketone can undergo
a transannular ene reaction to give 88. It is important to notice
that the stereochemical outcome of the final product 88 will
Scheme 19 Synthesis of decalin frameworks.
Chem. Commun., 2007, 2211–2221 | 2217
This journal is ß The Royal Society of Chemistry 2007

View Online
Scheme 20 Proposed mechanism for the tandem oxy-Cope–Claisen–ene reaction.
macrocycle 101. As a result, diastereomeric products 107 and formation of compound 110 (105) is favored over 111 (106)
1
08 are never observed.
thus explaining the diastereoselectivity of the tandem process.
All other examples of this tandem sequence illustrate
Curtin–Hammett control in the ene reaction as well. While
we cannot discuss all the results in this account, we can state
generally that the favored diastereomer observed is the one
arising from the ene reaction where the most sterically
demanding substituents are in equatorial conformations.
In order to confirm that the Claisen rearrangement is faster
than the ring inversion, we performed the tandem reaction
using enantiomerically enriched compounds 91 (R ? H; R ,
From intermediate 101, the ene product 106 can be
produced directly, or 101 can undergo a ring inversion to
02, which after transannular cyclization gives 105. At this
1
stage of the reaction, one might consider a fast equilibrium
between conformers. Consequently, the product ratio 105 : 106
is governed by the Curtin–Hammett principle. For the
majority of our substrates, we observed compounds structu-
rally related to 106 as the major product of the reaction,
although there are exceptions.
1
2
3
5
Much circumstantial evidence has been collected to confirm
the aforementioned mechanism is what is actually occurring in
3 4
R , R = H; ee = 98%). In these cases, intermediates 99 and
100 are enantiomers, rather than diastereomers, by virtue of
planar chirality. Therefore, ring inversion of 99 will result in a
complete racemization of the final products 105 and 106. For
every substrate investigated, there was no loss of ee in any case,
proving, without a shadow of a doubt, that the ring inversion
of 99 to 100 is a higher energy process than the Claisen
rearrangement (99 A 101).
3
4
this tandem sequence. A general survey of our results clearly
shows that products 107 and 108 were never observed in any
case. This strongly supports the conclusion that the ring
inversion of 99 to 100 is a much slower process than the
Claisen rearrangement of 99 to 101.
This is exemplified in the tandem reaction of 109 to give 110
and 111 in a 4 : 1 ratio (Scheme 21). Initial oxy-Cope
rearrangement of 109 gives directly 112, which is analogous to
compound 99. One might propose that conformer 112 should
have a higher energy transition state for the Claisen rearrange-
ment than conformer 113. If the inversion process of 112 to
1
1
13 was possible, we should have observed products of type
07/108. Since we did not observe any of these products, one
can suggest that the rotation barrier of the tetrasubstituted
enol ether in 112 (or 99) is a high energy process.
The product distribution of 110 and 111 can be rationalized
by Curtin–Hammett control of the ene reaction. The transition
1 3
state leading to 111 (106, R and R = Me) in this case requires
that both methyl groups be situated axially, whereas that
leading to 110 (105, R and R = Me) has the two methyl
1
Scheme 21 Tandem oxy-Cope–Claisen–ene reaction of allyl ether
109.
3
groups situated equatorially. As a consequence of this, the
2
218 | Chem. Commun., 2007, 2211–2221
This journal is ß The Royal Society of Chemistry 2007

View Online
Scheme 22 Proposed mechanism for the tandem oxy-Cope–Claisen–
ene reaction of allyl ether 114.
Scheme 23 Ring inversion.
transannular ene reactions 120 A 115 and 122 A 116 has a
21
7
. Theoretical evidence supporting the oxy-Cope–Claisen–ene
2
1
mechanism
higher energetic cost at 34.0 kcal mol and 34.6 kcal mol
respectively. This points to a rapid ring inversion that does not
compete with the ene reactions.
While there was ample experimental evidence supporting the
proposed oxy-Cope–Claisen–ene mechanism, we felt it neces-
sary to provide theoretical proof that the possible ring
inversion before the Claisen rearrangement is a high energy
process, and that the ene reaction is under Curtin–Hammett
control. Furthermore, a theoretical investigation of this
process would constitute a quantitative assessment of the
energy barriers associated with the processes in our proposed
reaction mechanism above and beyond the general conclusions
that we have drawn through rationalization of the reaction
We ultimately determined from our computational results
that the ene reaction is under Curtin–Hammett control, and
that the ratio of diastereomers arising from the ene portion of
the reaction mechanism can be calculated from the relative
transition state energies of the ene reactions with reasonable
3
5
precision. These calculations reinforce our original mechan-
ism as the correct depiction of the reaction manifold that is
followed in the oxy-Cope–Claisen–ene cascade.
3
5
outcomes.
8
. Oxy-Cope–Claisen–ene in total synthesis
We decided to model the reaction of 114, which we had pre-
viously examined experimentally (Scheme 22). When consider-
ing the ring inversion of the initial oxy-Cope reaction product
Having developed an in-depth understanding of the oxy-
Cope–Claisen–ene reaction, we sought to apply this powerful
sequence to the synthesis of several structurally challenging
natural products. Our first target was the natural product
wiedemannic acid (124), an abietane diterpene isolated from
1
17, we must take into account that it is a two-step process
where each double bond rotates through the 10-membered ring
separately (117 A 118 A 119). Therefore, two transition states
must be examined to evaluate the transformation.
3
6
Salvia wiedemannii (Scheme 24). A cursory analysis of the
molecule reveals five contiguous stereogenic centers based
around a trans-decalin core, making this the apparent major
synthetic challenge. Reviewing our previous studies of the oxy-
Cope–Claisen–ene reaction, we found that we had previously
prepared compound 94 from 123, which possesses several
architectural features of wiedemannic acid (124). Taking
advantage of this, we could quickly elaborate an analog of
DFT calculations at the B3LYP level of theory using the
6
-31G(d,p) basis set showed a barrier for the rotation of 117 to
2
18 of 17.3 kcal mol , which, as we would expect, is an
1
1
allowed process at 200 uC. The second rotation of 118 to 119
was found, however, to have an enormous barrier of rotation
2
1
of 160.5 kcal mol , thereby making it an unfeasible
conformational change. While we expected one of the two
olefin rotations to have a high barrier of inversion based on
experimental evidence that supports no ring inversion, we were
3
7
the natural product (125).
Unfortunately, despite the close resemblance of our analog
1
25 to the reported structure of wiedemannic acid (124), the H
surprised to find
a barrier of such magnitude for a
1
1
3
tetrasubstituted olefin. In addition, we found that the
transition state energy of the Claisen rearrangement step itself
and C NMR data of 125 were significantly different from
those reported for 124. In particular, the portion of 125 that is
2
1
is 24.1 kcal mol , which is just one seventh that of the
inversion of 118 to 119; hence, it is predicted to be the favored
process over ring inversion. This is in agreement with our
experimental findings.
We also examined the energies of the second ring inversion
versus the two possible transannular ene reactions in the
second half of this tandem reaction. The ring inversion is again
a two-step process, with both an outside rotation of the
carbonyl 120 to 121, which has a negligible energy cost, and a
rotation of the trisubstituted olefin 121 to 122 (Scheme 23).
The latter represents the rate determining step of the overall
ring flip. We calculated the barrier of the ring inversion of 121
2
1
to 122 to be 14.0 kcal mol , whereas the energy of both
Scheme 24 Synthesis of wiedemannic acid analog.
Chem. Commun., 2007, 2211–2221 | 2219
This journal is ß The Royal Society of Chemistry 2007

View Online
Acknowledgements
We thank the University of Ottawa, Natural Science and
Engineering Research Council (NSERC), Ontario Innovation
Trust (OIT), Canada Foundation for Innovation (CFI),
Government of Ontario (PREA), Merck-Frosst Canada,
Merck Research Laboratories, AstraZeneca, Bristol Myers
Squibb and Boehringer Ingelheim for their financial support.
S. A. thanks NSERC and OGS for postgraduate scholarships.
Most importantly, we thank each and every student that was
involved in the research and development of the projects
described above, without them this work would not be possible.
References
Scheme 25 Approach toward the synthesis of teucrolivin A.
1 (a) K. C. Nicolaou, D. J. Edmonds and P. G. Bulger, Angew.
Chem., Int. Ed., 2006, 45, 7134; (b) L. F. Tietze, G. Brasche and
K. Gericke, Domino Reactions in Organic Synthesis, Wiley-VCH,
Weinheim, 2006; (c) H. Pellisier, Tetrahedron, 2006, 62, 1619; (d)
L. F. Tietze, Chem. Rev., 1996, 96, 115; (e) P. J. Parsons,
C. S. Penkett and A. J. Shell, Chem. Rev., 1996, 96, 195; (f)
T.-L. Ho, Tandem Organic Reactions, Wiley, New York, 1992.
close to the equatorial methyl group possesses different
chemical shifts and coupling constants. The structure of 125
was confirmed by X-ray crystallography, we compared our
data to other natural abietane diterpenes and we concluded
that the initial structural assignment for wiedemannic acid
2
For a biomimetic approach, see: (a) W. M. Bandaranayake,
J. E. Banfield, D. St. C. Black, G. D. Fallon and B. M. Gatehouse,
J. Chem. Soc., Chem. Commun., 1980, 162; (b) W. M.
Bandaranayake, J. E. Banfield and D. St. C. Black, J. Chem.
Soc., Chem. Commun., 1980, 902.
(
124) was probably incorrect, prompting us to abandon the
total synthesis at that point.
Our most recent synthetic pursuit using the oxy-Cope–
Claisen–ene cascade is directed towards the synthesis of
3 (a) K. C. Nicolaou, N. A. Petasis, R. E. Zipkin and J. Uenishi,
J. Am. Chem. Soc., 1982, 104, 5555; (b) K. C. Nicolaou,
N. A. Petasis, J. Uenishi and R. E. Zipkin, J. Am. Chem. Soc.,
3
8
teucrolivin A 130, a neoclerodane diterpenoid isolated from
Teucrium oliverianum, a plant whose genus is a rich source of
such compounds (Scheme 25). While this compound is not
known to exhibit any notable biological activity, the highly
oxygenated core, coupled with an additional all-carbon
quaternary center at the decalin ring junction not accessible
with the pericyclic cascade, presents a significant synthetic
challenge.
1
982, 104, 5557; (c) K. C. Nicolaou, R. E. Zipkin and N. A. Petasis,
J. Am. Chem. Soc., 1982, 104, 5558; (d) K. C. Nicolaou, N. A.
Petasis and R. E. Zipkin, J. Am. Chem. Soc., 1982, 104, 5560.
J. M. Warrington, G. P. A. Yap and L. Barriault, Org. Lett., 2000,
4
5
2
, 663.
(a) K. Rajagopalan and R. Srinivasan, Tetrahedron Lett., 1998, 39,
133; (b) P. Shanmugam, B. Devan, R. Srinivasan and
4
K. Rajagopalan, Tetrahedron, 1997, 53, 12637; (c)
K. Rajagopalan and P. Shanmugam, Tetrahedron, 1996, 52, 7737.
M. Sworin and K.-C. Lin, J. Am. Chem. Soc., 1989, 111, 1815.
A. P. Chorlton, G. A. Morris and J. K. Sutherland, J. Chem. Soc.,
Perkin Trans. 1, 1991, 1205.
6
7
We prepared the cascade precursor 127 in just 10 steps from
1
,3-cyclohexadiene 126. We were successful in implementing
8
(a) L. A. Paquette, S. Nakatani, T. M. Zydowsky,
S. D. Edmondson, Q.-L. Sun and R. Skerlj, J. Org. Chem., 1999,
64, 3244; (b) K. Rajagopalan, S. Janardhanam and A. Balakumar,
J. Org. Chem., 1993, 58, 5482 and references therein.
the cascade reaction to access the decalin core 128 of the
natural product, albeit in a moderate yield due to the
instability of the 3-substituted furan to the reaction condi-
3
9
tions. To date, we have achieved the synthesis of the
advanced intermediate 129 and we are currently investigating
the installation of the last remaining all-carbon quaternary
center at the ring junction. Significant progress has been made
and will be reported in due course.
9 They observed that electron-rich double bonds cause the anionic
oxy-Cope reaction to decelerate, see: F. Haeffner, Y. R. Reddy,
K. N. Houk and L. A. Paquette, J. Am. Chem. Soc., 1999, 121,
1
1880.
10 T. Terada and S. Yamamura, Tetrahedron Lett., 1979, 20, 1623.
11 (a) D. Y. Curtin, Rec. Chem. Prog., 1954, 15, 111; (b) S. Winstein
and N. J. Holness, J. Am. Chem. Soc., 1955, 77, 5562; (c) E. L. Eliel,
Stereochemistry of Carbon Compounds, McGraw-Hill, New York,
1962; (d) L. P. Hammett, Physical Organic Chemistry, McGraw-
Hill, New York, 1970, ch. 5; (e) J. I. Seeman, Chem. Rev., 1983, 83,
83; (f) J. I. Seeman, J. Chem. Educ., 1986, 63, 42.
Conclusions
Over the past several years we have developed numerous
pericyclic reaction cascades that constitute efficient and
reliable methods for the generation of various carbocyclic ring
systems with high diastereoselectivity at one or more stereo-
genic centers. Accompanying these methodological develop-
ments we have invested significant effort to gain a deep
mechanistic understanding of these processes and have used
this knowledge in application to the total synthesis of various
natural targets. Undoubtedly we, and chemists as a whole, will
continue to develop cascading reactions in the interest of
greater reaction efficiency and facile access to more and more
complicated systems in a single transformation.
1
2 (a) Y. Fujita, T. Onishi and T. Nishida, J. Chem. Soc., Chem.
Commun., 1978, 651; (b) Y. Fujita, T. Onishi and T. Nishida,
Synthesis, 1978, 934; (c) A. Utagawa, H. Hirota, S. Ohno and
T. Takahashi, Bull. Chem. Soc. Jpn., 1988, 61, 1207.
13 D. Deon, MSc thesis, University of Ottawa, 2001.
1
1
4 P. R. MacLean, MSc thesis, University of Ottawa, 2003.
5 G. D. Brown, L.-K. Sy and R. Haynes, Tetrahedron, 1998, 54,
4
345. The natural enantiomer is actually (2)-arteannuin M, whose
absolute configuration we assigned through its total synthesis.
16 L. Barriault and D. Deon, Org. Lett., 2001, 3, 1925.
1
7 (a) M. Nakazaki, K. Yamamoto and K. Naemura, Top. Curr.
Chem., 1984, 125, 1; (b) E. L. Eliel and S. H. Wilen,
Stereochemistry of Organic Compounds, Wiley-Interscience, New
York, 1994.
2
220 | Chem. Commun., 2007, 2211–2221
This journal is ß The Royal Society of Chemistry 2007

View Online
1
8 Microwave heating was found to be more desirable compared to
conventional heating methods, with our cascading reaction
sequences occurring with a 10–300 fold increase in rate under
these conditions. For reviews on the use of microwaves in organic
synthesis, see: (a) A. Loupy, A. Petit, J. Hamelin, F. Texier-Boullet,
P. Jacquault and D. Math e´ , Synthesis, 1998, 1213; (b) G. Majetich
and R. J. Hichs, Journal of Microwave Power and Electromagnetic
Energy, 1995, 30, 27; (c) A. Loupy and L. Perreux, Tetrahedron,
347, 1473; (c) Quaternary Stereocentres: Challenges and Solutions
for Organic Synthesis, ed. J. Christoffers and A. Baro, Wiley-VCH,
Weinheim, 2005; (d) D. J. Ramon and M. Yus, Curr. Org. Chem.,
2004, 8, 149; (e) C. J. Douglas and L. E. Overman, Proc. Natl.
Acad. Sci. U. S. A., 2004, 101, 5363; (f) L. Barriault and
I. Denissova, Tetrahedron, 2003, 59, 10105; (g) J. Christoffers
and A. Baro, Angew. Chem., Int. Ed., 2003, 42, 1688; (h)
J. Christoffers and A. Mann, Angew. Chem., Int. Ed., 2001, 40,
4591; (i) E. J. Corey and A. Guzman-Perez, Angew. Chem., 1998,
2001, 57, 9199; (d) P. Lidstr o¨ m, J. Tierny, B. Wathey and
J. Westman, Tetrahedron, 2001, 57, 9225; (e) A. Loupy,
Microwaves in Organic Synthesis, Wiley-VCH, Weinheim, 2006.
9 All reactions were performed in toluene, using different additives
as noted above depending on the purpose of the experiment.
0 Bases used: DBU, TMEDA, Et N, pyridine, 2,6-di-tert-butyl-
3
pyridine, DMAP, sparteine and 2-tert-butyl-1,1,3,3-tetramethyl-
guanidine.
1 Conditions: KHMDS, DME, 110 uC. Quench with AcOH to
isolate the macrocyclic ketone, the precursor for the ene reaction in
the thermal sequence.
2 (a) B. B. Snider, in Comprehensive Organic Synthesis, Combining
C–C p Bonds, ed. L. A. Paquette, Pergamon Press, Oxford, 1991,
vol. 5, ch. 1.1; (b) B. B. Snider, in Comprehensive Organic Synthesis,
Addition To C–X p Bonds, ed. C. H. Heathcock, Pergamon Press,
Oxford, 1991, vol. 2, ch. 2.1.
1
10, 402.
3
0 G. A. Ellestad, M. P. Kunstmann, P. Mirando and G. O. Morton,
J. Am. Chem. Soc., 1972, 94, 6206.
1 (a) M. Nakagawa, Y.-H. Hsu, A. Hirota, S. Shima and
M. Nakayama, J. Antibiot., 1989, 42, 218; (b) M. Nakagawa,
Y.-H. Hsu, A. Hirota, S. Shima, M. J. Nakayama, T. Adachi and
H. Nozaki, J. Antibiot., 1989, 42, 223.
2 E. L. O. Sauer and L. Barriault, J. Am. Chem. Soc., 2004, 126,
8
3 Seminal work in the field of diastereoselective control through
1
2
3
2
3
569.
3
2
macrocyclic conformation was performed by W. C. Still et al., see:
(
a) W. C. Still and I. Galynker, Tetrahedron, 1981, 37, 3981; (b)
W. C. Still, J. Am. Chem. Soc., 1977, 99, 4186; (c) W. C. Still, J. Am.
Chem. Soc., 1979, 101, 2493; (d) W. C. Still, Curr. Trends Org.
Synth., Proc. Int. Conf. 4th 1982 (1983), 233; (e) W. C. Still,
S. Murata, G. Revial and K. Yoshihara, J. Am. Chem. Soc., 1983,
1
1
1
4 There is a wealth of experimental evidence supporting this
mechanism; however it is impossible to include all results in a
review: key ideas have therefore been presented. A full account of
the results leading to these conclusions can be found in ref. 32 and
2
2
3 L. Barriault and I. Denissova, Org. Lett., 2002, 4, 1371.
4 L. Barriault, J. A. Farand and I. Denissova, Heterocycles, 2004, 62,
05, 625; (f) W. C. Still and V. J. Novack, J. Am. Chem. Soc., 1984,
06, 1148; (g) W. C. Still and A. G. Romero, J. Am. Chem. Soc.,
986, 108, 2105.
735.
2
5 (a) J. J. Gajewski and N. D. Conrad, J. Am. Chem. Soc., 1979, 101,
2
1
747; (b) F. Barluenga, A. R. Liz and M. Bayod, J. Org. Chem.,
987, 52, 5190; (c) J. J. Gajewski, Acc. Chem. Res., 1997, 30, 219.
3
2
6 (a) T. Tsuchida, H. Iinuma, C. Nishida, N. Kinoshita, T. Sawa,
M. Hamada and T. Takeuchi, J. Antibiot., 1995, 48, 1104; (b)
T. Tsuchida, H. Iinuma, R. Sawa, T. Takahashi, H. Nakamura,
K. T. Nakamura, T. Sawa, H. Naganawa and T. Takeuchi,
J. Antibiot., 1995, 48, 1110.
3
5.
3
3
5 E. L. O. Sauer, J. H. Hooper, T. Woo and L. Barriault, J. Am.
Chem. Soc., 2007, 129, 2112.
6 (a) A. Ulubelen, G. Topcu and B. Terem, Phytochemistry, 1987, 26,
2
7 (a) F. F. Paintner and G. Bauschke, Tetrahedron Lett., 2003, 44,
2549; (b) F. F. Paintner, L. Allmendinger, G. Bauschke, C. Berns
1534; (b) A. Ulubelen and G. Topcu, Phytochemistry, 1990, 29,
2346.
and P. Heisig, Bioorg. Med. Chem., 2003, 11, 2823; (c) F. F.
Paintner, L. Allmendinger, G. Bauschke and K. Polborn, Synlett,
2002, 1308; (d) F. F. Paintner, G. Bauschke and M. Kestel,
Tetrahedron Lett., 2000, 41, 9977.
8 J. M. Warrington and L. Barriault, Org. Lett., 2005, 7, 4589.
9 For recent reviews, see: (a) B. M. Trost and C. Jiang, Synthesis,
37 E. L. O. Sauer and L. Barriault, Org. Lett., 2004, 6, 3329.
38 M. Bruno, A. A. Omar, A. Perales, F. Piozzi, B. Rodriguez,
G. Savona and M. C. De la Torre, Phytochemistry, 1991, 30,
275.
2
2
2006, 369; (b) J. Christoffers and A. Baro, Adv. Synth. Catal., 2005,
39 S. Arns and L. Barriault, J. Org. Chem., 2006, 71, 1809.
This journal is ß The Royal Society of Chemistry 2007
Chem. Commun., 2007, 2211–2221 | 2221