An approach to mimicking the sesquiterpene cyclase phase by nickel-promoted diene/alkyne cooligomerization

Article abstract of DOI:10.1021/jo202314a

Artificially mimicking the cyclase phase of terpene biosynthesis inspires the invention of new methodologies, since working with carbogenic frameworks containing minimal functionality limits the chemist's toolbox of synthetic strategies. For example, the construction of terpene skeletons from five-carbon building blocks would be an exciting pathway to mimic in the laboratory. Nature oligomerizes, cyclizes, and then oxidizes γ,γ-dimethylallyl pyrophosphate (DMAPP) and isopentenyl pyrophosphate (IPP) to all of the known terpenes. Starting from isoprene, the goal of this work was to mimic Nature's approach for rapidly building molecular complexity. In principle, the controlled oligomerization of isoprene would drastically simplify the synthesis of terpenes used in the medicine, perfumery, flavor, and materials industries. This article delineates our extensive efforts to cooligomerize isoprene or butadiene with alkynes in a controlled fashion by zerovalent nickel catalysis building off the classic studies by Wilke and co-workers.

Full text of DOI:10.1021/jo202314a

Featured Article
pubs.acs.org/joc
An Approach to Mimicking the Sesquiterpene Cyclase Phase
by Nickel-Promoted Diene/Alkyne Cooligomerization
Dane Holte,† Daniel C. G. Gotz,† and Phil S. Baran*
̈
Department of Chemistry, The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, California 92037, United States
S
* Supporting Information
ABSTRACT: Artificially mimicking the cyclase phase of
terpene biosynthesis inspires the invention of new method-
ologies, since working with carbogenic frameworks containing
minimal functionality limits the chemist’s toolbox of synthetic
strategies. For example, the construction of terpene skeletons
from five-carbon building blocks would be an exciting pathway to mimic in the laboratory. Nature oligomerizes, cyclizes, and then
oxidizes γ,γ-dimethylallyl pyrophosphate (DMAPP) and isopentenyl pyrophosphate (IPP) to all of the known terpenes. Starting
from isoprene, the goal of this work was to mimic Nature’s approach for rapidly building molecular complexity. In principle, the
controlled oligomerization of isoprene would drastically simplify the synthesis of terpenes used in the medicine, perfumery,
flavor, and materials industries. This article delineates our extensive efforts to cooligomerize isoprene or butadiene with alkynes
in a controlled fashion by zerovalent nickel catalysis building off the classic studies by Wilke and co-workers.
Organic chemists have held a long-term fascination with
terpenoids because of their structural diversity, challenging
molecular frameworks, and opportunities for the invention of
methods and strategies in organic synthesis.1,6 In Nature, γ,γ-
dimethylallyl pyrophosphate (DMAPP, 5) and isopentenyl
pyrophosphate (IPP, 6) are first oligomerized to relatively
simple linear precursors like geranyl pyrophosphate (7) and
farnesyl pyrophosphate (8, Scheme 1). Additional oligomeriza-
tions with IPP (6) can lead to higher order linear terpenes.7
Following prenyl chain extension, carbocationic cyclizations can
afford monocyclic compounds, such as β-bisabolene (9),
germacrene A (10), and humulene (11, Scheme 1). After
building the carbocyclic skeleton, Nature employs both oxidase
enzymes and nonenzymatic oxidation processes to take the
unadorned carbocycles to various oxidation states.
INTRODUCTION
■
In perhaps the most dramatic example of divergent synthesis,
Nature utilizes simple five-carbon (C5) building blocks to
fashion tens of thousands of natural products,1 known
collectively as terpenes and possessing a diverse array of
physical and biological properties (for selected examples of
terpene structures, see Figure 1).2,3 Terpenes have long been
In contrast, a recapitulation of the biosynthetic route in the
laboratory setting remains an unmet challenge. As a modest
step in that direction, our laboratory recently outlined a two-
phase biomimetic approach to terpenes consisting of the ef-
ficient construction of a minimally oxidized terpenoid followed
by stepwise, chemoselective oxidations (Figure 2, left).8,9 The
full development of this approach may be broadly applied to
the synthesis of terpene natural products, and as such, studies
have been initiated to probe artificial versions of both “cyclase”
and “oxidase” phase.
Figure 1. Selected examples of terpene natural products: polyisoprene
(natural rubber, 1), dihydrojunenol (2), vinigrol (3), and Taxol (4).
Although our previous approach to the eudesmane
sesquiterpenes (e.g., pygmol (12) and dihydrojunenol (2)) is
efficient and scalable (Figure 2, left),9,10 it is strategically
inferior to Nature’s union of simple five-carbon building blocks.
In further contrast to biosynthesis, the target-oriented synthesis
used to access dihydrojunenol (2) from ketone 13 and
utilized as flavors and fragrances, from floral to citrus and
peppermint. The direct use of isoprene in the form of natural
rubber, polyisoprene (1, Figure 1), by the ancient Mesoamer-
icans dates back to as early as 1600 B.C.4 Today the majority of
isoprene is obtained from the C5 cracking fractions of
petroleum refining and is mostly used in the production of
synthetic polymers.5
Received: November 10, 2011
Published: January 3, 2012
© 2012 American Chemical Society
825
dx.doi.org/10.1021/jo202314a | J. Org. Chem. 2012, 77, 825−842

The Journal of Organic Chemistry
Featured Article
Scheme 1. Biosynthesis of Terpenoid Natural Products
starting material. Attempts to achieve this goal through a closely
related diene/alkyne cooligomerization are described in this
article.
Planning and Historical Context.
“[Nickel] catalysis may also be expected to be used in the
partial syntheses of natural products (e.g. terpenes from
isoprene).” Paul Heimbach, 197312
Isoprene is underutilized in terpene total synthesis, often
because more advanced terpenoid starting materials are
available (e.g., limonene, farnesol, or steroid skeletons).
Additionally, there is a lack of methods to rapidly process
isoprene into synthetically useful building blocks. The initial
inspiration for harnessing the reactivity of isoprene arose from
the well-established fact that dienes react with a variety of
transition metals through an oxidative cyclization to form
metallacyclopentenes. These metallacycles are known to be key
intermediates in the oligomerization of dienes. If successfully
controlled, a cyclo-oligomerization approach could rapidly
furnish macrocyclic terpene natural products, like the simple
sesquiterpenes germacrene6 or humulene,13 which currently
require multistep total syntheses.
Identification of the metal−ligand system with the best
precedence for rapidly building medium-sized rings was the first
priority. In principle, any metal capable of two-electron redox
chemistry with a high affinity for π-donating ligands would be
able to catalyze isoprene oligomerization. Indeed, a plethora of
transition metals have been used to promote diene oligome-
rization,14,15 the most prominent of which are Pd,16−22 Fe,23−25
Co,23,26 Cr,27−29 Ti,28−31 Ru,32 and Zr.30
Eventually, the chemistry of Wilke and Heimbach using
zerovalent nickel for the oligomerization of dienes piqued our
interest.12,33−35 Utilizing dienes, Wilke and co-workers have
established methods for the creation of 4,36−38 5,39−41 6,42
8,36−38 10,43−45 12,46 and >12-membered rings12,47 (Scheme 2A).
In particular, the synthesis of 10-membered ring systems by
cooligomerization with a C2 component, either alkenes48,49 or
alkynes,44,45,50,51 has been developed in great detail (vide infra).
There have been extensive studies on the effects of solvent,
temperature, pressure, ligands, and substrates (for both the diene
and cooligomerization partner).52 In the case of cyclo-cool-
igomerization with ethylene, Wilke was able to achieve 78% crude
Figure 2. Retrosynthetic analysis of eudesmane natural products
utilizing a two-phase approach to terpenes: direct comparison between
the previously published, target-oriented method (left) and the
proposed biomimetic route (right).
aldehyde 14 does not permit divergent entry to numerous
scaffolds. Thus, a route from a simple C5 unit, isoprene, to
macrocyclic terpene natural products was pursued. Such a plan
would rapidly forge the all-carbon germacrane skeleton (10), in
the lowest possible oxidation state (Figure 2, right). In
principle, such a structure could then be processed to a
multitude of different terpenes, including the eudesmanes. This
strategy would represent an ideal entry to many terpene
families, as it does not require the divergent prefunctionaliza-
tion of a five-carbon fragment (like the conversion of mevalonic
acid pyrophosphate into DMAPP (5) and IPP (6) in Nature11)
but would rely on the feedstock chemical isoprene as the sole
826
dx.doi.org/10.1021/jo202314a | J. Org. Chem. 2012, 77, 825−842

The Journal of Organic Chemistry
Featured Article
Scheme 2. (A) Synthesis of Monocyclic Carbon Skeletons
from Butadiene by Nickel Catalysis and (B) the Nickel-
Bound C8-Intermediate 15 as a “Geranyl Synthon”
Scheme 3. Development of a Controlled Ni0-Catalyzed
Cooligomerization of Isoprene: Rapid Access to
Germacrene-Derived Natural Products
yield on multikilogram scale.53 This seems ideal, as all of the
skeletons targeted in this research program would be carried on to
further oxidase-phase studies that require scalable access to the
lowly oxidized germacrene scaffold. Heimbach’s striking quote
above was particularly inspiring in this regard. To our knowledge,
despite Wilke’s and Heimbach’s pioneering studies on the Ni0-
catalyzed cyclo-oligomerization, chemists have not explored this
prospect in the context of natural product synthesis.54−57
To further expand the concept of alkyne/diene cool-
igomerization, the catalytically formed metal-bound C8-
intermediate 15 (Scheme 2B) was identified as a reactive
building block with differentiated termini58 with the potential
to be intercepted by a variety of C1, C2, and C3 units to
selectively access caryophyllane, germacrane, and humulane
frameworks, respectively (Scheme 2A). Establishing such a
cooligomerization as a method for terpene synthesis would
require a great expansion of Wilke’s methodologyalready
well worked out for butadieneby judicious choice and fine-
tuning of both the catalytic system and cooligomerization
components.
Zerovalent nickel stands out for its pronounced tendency to
form medium-sized rings due to its innate size.59 In the absence
of ligands, Ni0 will engage three butadiene molecules and
rapidly form 1,5,9-trans,trans,trans-cyclododecatriene (CDT)
catalytically and in greater than 80% yield.60,61 Although
handling Ni0 complexes requires an inert atmosphere, there are
reasonably stable zerovalent nickel sources such as Ni(cod)2
which can be purchased or easily prepared in multigram
quantities.62 In addition, the respective NiII precursors (e.g.,
Ni(acac)2) are air-stable, inexpensive, and are easily converted
in situ to catalytically active Ni0 species by a variety of common
reducing agents.63
afford different functional handles on the terpenoid backbone
and thus allow access to related sesquiterpenes like β-elemene
(18), α-eudesmol (19), or constunolide (20) within a
minimum number of synthetic manipulations.
The field of catalytic C−C coupling processes has undergone
a stunningly rapid development since the early work of Wilke
and his co-workers. Thus, a reinvestigation to overcome some
of the limitations that have to date precluded the application of
Ni0-catalyzed ring-forming processes of isoprene to significantly
more complex terpenoid frameworks was pursued. Several
issues needed to be addressed in order to realize this plan; these
are highlighted within the catalytic cycle of cooligomerization
proposed by Wilke and co-workers (Scheme 4).64 In the first
step, two isoprene units must be oxidatively dimerized
regioselectively, in a head-to-tail fashion, as found in Nature.
In an experiment with stoichiometric nickel, Wilke has shown
that the regioselectivity in the formation of 21 can be altered
by employing a suitable ligand. Strikingly, the use of PMe3 led
to a 50:50 mixture of tail-to-tail and head-to-tail dimers, while
changing the phosphine ligand to PPh3 or PiPr3 yielded
exclusively the desired coupling product with head-to-tail
connectivity.64,65 Similarly, van Leeuwen et al. have demonstrated
modest control of regioselectivity in the cyclodimerization of
isoprene through extensive exploration of phosphine ligands.66
Second, the hapticity in the key intermediate 22 (previously
characterized by X-ray diffractometry40,67−69) has to be
controlled since it determines the regioselective outcome of
the alkyne incorporation.65,70 The Ni0 intermediate 22 has been
shown to exist as the η1,η3-complex (as drawn) irrespective of
the nature of the phosphine ligand.64 Since the two reactive
positions in key intermediate 22 are clearly differentiated
both sterically and electronically58regioselective incorpora-
tion of unsymmetrically disubstituted alkynes was deemed a
worthy pursuit.
If successfully adapted to isoprene, the application of this
method employing simple alkynes 16a−c could produce the
germacrane skeleton in a single step (17a−c, Scheme 3).
Sophisticated choice of the alkyne “coupling” partner should
The most serious drawback of the alkyne/diene cool-
igomerization as originally reported by Wilke is that the
827
dx.doi.org/10.1021/jo202314a | J. Org. Chem. 2012, 77, 825−842

The Journal of Organic Chemistry
Featured Article
higher-order oligomers and polymers, both linear and cyclic.76
For simple systems, Wilke was able to optimize the ligand,
temperature, and reactant ratio in order to partially suppress
such nonproductive pathways.44,45,77
Scheme 4. Catalytic Cycle as Proposed by Wilke, and
Challenges (blue) Associated with the Co-oligomerization of
Isoprene and Alkynes
Finally, germacrene sesquiterpenes are notoriously unstable
to acidic conditions (leading to cyclized products) and thermal
conditions (leading to Cope rearrangements)78 and can often
exist as conformationally stable isomers at ambient temper-
atures.79−81
All of the above issues lead to further serious, albeit technical
obstacles since the copolar nature of the products creates
problems with regard to purification of crude reaction mixtures
and subsequent analysis of inseparable mixtures of isomers.
Because Heimbach conducted most of his work with butadiene
(instead of isoprene) and symmetric alkynes, many of the above
difficulties were avoided. Furthermore, his research group often
simplified the resulting mixture of crude products by hydro-
genation or Cope rearrangement.82,83 Our proposal, utilizing
isoprene and unsymmetrical alkynes, and accounting for all
possible combinations of E and Z olefins, leads to 32 possible
isomers84 (as opposed to only six possible isomers for the
cooligomerization of symmetric alkynes with butadiene). While
Wilke and co-workers usually purified complex product
mixtures by gas chromatography and/or spinning-band
distillation on preparative scale,43,51 this was deemed unsuitable
for the synthesis of terpenoid natural products in multigram
quantities. Despite these tremendous obstacles, the benefits of
such a simple and rapid synthesis were still enticing, and thus
research endeavors were carried out.
cyclodecatrienes are always obtained as the 4Z,7Z,10E isomers,71
whereas most naturally occurring germacrenes feature two
endocyclic 4E,10E double bonds.72 It was conjectured that
fine-tuning the catalytic system using modern ligands capable
of stabilizing low-valent transition metals and/or the addition of
suitable additives might resolve this issue.73 As an alternative, the
possibility of delaying the olefin isomerization to a subsequent
transformation was considered.74,75
RESULTS AND DISCUSSION
■
Butadiene Cooligomerizations. Initial forays utilized
butadiene as a means to probe the scope of alkyne substitution.
Wilke’s reported procedure45 was repeated at first: Ni(cod)2
(3.4 mol %), PPh3 (3.4 mol %), butadiene (5 equiv), and
4-octyne (16d), neat, and sealed in a pressure vessel at room
temperature for ∼18 h (Scheme 5A). With similar dialkyl
alkynes, Wilke reported a ∼90% yield of 10-membered rings
based on GC which was immediately distilled to afford the
respective Cope products.85 In our hands, macrocycle 24d
Additionally, if intermediate 23 reenters the catalytic cycle,
the cooligomerization reaction has the potential to produce
a
Scheme 5. Major Products of Selected Butadiene/Alkyne Cooligomerization Reactions
a
Reagents and conditions: (a) 5:1 butadiene/alkyne, Ni(cod)2 (10 mol %), PPh3 (10 mol %), 16 h, 83%; (b) 120 °C, 1 h, quant; (c) 40:1 butadiene/
alkyne, Ni(cod)2 (10 mol %), PPh3 (10 mol %), 16 h.
828
dx.doi.org/10.1021/jo202314a | J. Org. Chem. 2012, 77, 825−842

The Journal of Organic Chemistry
Featured Article
could be isolated (along with trace impurities) by silica gel
chromatography, eluting with hexanes, in 83% yield. As
anticipated, 4Z,7Z,10E-cyclodecatriene 24d was conformation-
ally flexible at room temperature and required low-temperature
NMR for reliable structure elucidation; this turned out to be
a recurring issue for the analysis of any 4Z,7Z,10E-cyclo-
decatriene system. Compound 24d underwent quantitative
Cope rearrangement at 120 °C to give divinylcyclohexene 25.86
It was found that simply increasing the catalyst loading from
3.4 mol % to 10 mol % shortened the necessary reaction time.
The use of a pressure vessel was essential for the reaction to pro-
ceed quickly and with good conversion, as previously demon-
strated by Wilke. Unfortunately, the calibrated and heatable
glass autoclave routinely used by Wilke to monitor internal
pressure, volume loss, and temperature over the course of the
reactions was not available to us.37,44,51,87 Although volume loss
was observed during the reactions, its quantification was not
possible. To obtain comparable results, reactions were typically
stopped after ∼16−20 h.
Reppe-type cycloadducts (e.g., 26, Scheme 5B).88,89 This well-
precedented cycloaddition demonstrates one background reaction
that occurred with many different substrates (cf. formation of 29,
Scheme 5C). Many reactions also afforded hetero-[2 + 2 + 2]
products, in which two alkynes and a single butadiene molecule
had reacted (e.g., rac-27 and 28, Scheme 5B). In addition, linear
products were observed repeatedly (e.g., 31, Scheme 5D). As
expected, the isolation and structure elucidation of these
repeatedly occurring motifs was challenging, but their character-
ization enabled the establishment of guidelines for the analysis of
the conducted cyclo-cooligomerization reactions (Figure 3).
Because product purification was extraordinarily tedious, rapid
analysis of crude reaction mixtures by means of GC−MS and/or
crude NMR according to the flowchart shown in Figure 3 was
crucial.
Immediately, after running the reactions, crude GC/MS
analysis quickly showed if the reaction fit into category A or B:
no reaction at all or too complex to be practical. In both of
these cases, the reactions were abandoned at that point. If the
reaction appeared to show a limited number of major products,
but the desired product was not the main component by gas
chromatography (category C), the reaction was worked up.
After a preliminary substrate screen, several interesting, albeit
undesired, side products that plagued these reactions were
identified. Specifically, the use of alkynes with conjugated electron-
withdrawing groups (e.g., 16e) gave exclusively [2 + 2 + 2],
1
The crude reaction mixture was first analyzed by H and 13C
Figure 3. Flowchart that demonstrates the method of reaction analysis, illustrated with selected gas chromatograms representative for reactions
belonging to categories A−E.
829
dx.doi.org/10.1021/jo202314a | J. Org. Chem. 2012, 77, 825−842

The Journal of Organic Chemistry
Featured Article
chelation to Ni0 or the instability of zerovalent nickel in the
presence of protic substrates. Allylic alkyne 16k resulted in a
complex, inseparable mixture (entry 5, Table 1). As briefly
stated above, alkynes with electron-withdrawing substituents
(16e and 16l, entries 6 and 7, Table 1) reacted exclusively in
[2 + 2 + 2] fashion, which is well documented in the
literature.88 Under standard conditions, substrates 16b, 16m,
and 16n (entries 8−10, Table 1) also primarily underwent
[2 + 2 + 2] cyclotrimerization. This was not entirely unexpected,
as Reppe-type trimerizations were generally observed as a
background reaction. The use of excess butadiene (20 equiv)
with protected alcohol 16n (entry 11, Table 1) suppressed the
[2 + 2 + 2] reaction. Unfortunately, the primary 10-membered
ring was only observed in trace amounts by GC−MS; instead,
the follow-up Cope product predominated. In the presence of
homopropargyl alcohol (16c, entry 12, Table 1) the primary
product was found to be a linear adduct between two alkynes
and a single butadiene (31, Scheme 5d). This may be due to
the ability of the free alcohol to coordinate to the nickel center
bringing two alkynes closer to the reactive metal.90,91 Simple
alkyl alkynes 16a, 16d, or 16o are incorporated efficiently into
10-membered rings, as previously demonstrated by Wilke
(entries 13−15, Table 1). In the case of 1-octyne (16o), the
crude mixture was distilled directly and analyzed as its respective
Cope-rearranged product (structure not shown, see the Exper-
imental Section). It is worth mentioning that the reaction with
1-octyne (16o) is the first reported case of successful incorporation
of a terminal alkyne into a 10-membered ring system under Wilke’s
reported conditions. Utilizing isopropylacetylene (16a), the
product was isolated following derivatization by cycloaddition
with 1,1-dibromoformaldoxime (see the Experimental Section). To
the best of our knowledge, α-branched alkyl alkynes have not been
previously used in the Ni0-catalyzed cooligomerization with dienes.
To fully suppress the formation of linear side products (cf. entry
12, Table 1), the unsymmetrically disubstituted alkynes 16p and
16q were investigated (entries 16 and 17, Table 1). In each case,
the desired product was obtained, and crude analysis of the
reaction mixture was sufficient for structural elucidation. Product
24p (entry 16, Table 1) was isolated as a 1.6:1 mixture of
regioisomers (with respect to the alkyne incorporation) as
determined by low-temperature 2D-NMR.92 This encouraging
result indicated that regiochemical incorporation of the alkyne
might be even more effectively controlled by judicious choice of
the substrate.
Although surprising at first glance, it was observed that
disubstituted products 24p and 24q (entries 16 and 17, Table 1)
underwent Cope rearrangement at unusually low tempera-
tures. It is difficult to deduce simple rules for the propensity of
the 10-membered macrocycles to rearrange, but there are two
trends, both of which are supported by theses from the
Heimbach group.49 First, 4Z,7Z,10E-cyclodecatrienes undergo
Cope rearrangement at a lower temperature (room temperature
to 60 °C) than 4Z,10E-cyclodecadienes (∼150 °C). Second,
the ease of [3,3]-rearrangement correlates with increasing
bulkiness and number of the substituents on the olefinic
carbons. This is, however, extremely dependent on the position
of the substituents. These trends are a result of the orbital
overlap of the π-systems as determined by the conformation of
the macrocycle, which is strongly influenced by the substitution
pattern of the 10-membered ring. A significant driving force for
the rearrangement is the strain release of the macrocycle.
On the basis of the above results, terminal alkyne 16f (entry
18, Table 1) appeared to be an ideal cooligomerization partner
Table 1. Ni0-Catalyzed Butadiene/Alkyne
Cooligomerization
a
b
entry
alkyne
R1
R2
TBS
product outcome
1
2
3
4
5
6
7
8
16g
16h
16i
CH2CH2OTBS
CH2CH2OTBS
CH2CH2OTBS
(CH2)4CO2H
CH2CHCH2
CO2Me
24g
24h
24i
A
A
A
A
B
TBDPS
TIPS
H
16j
24j
16k
16e
16l
H
24k
24e
24l
c
c
c
c
c
e
f
CO2Me
H
C
C
C
C
C
C
C
CO2Et
16m
16b
16n
16n
16c
16d
16o
16a
16p
16q
16f
CH2CH2Cl
C(CH3)2OH
C(CH3)2OTMS
C(CH3)2OTMS
CH2CH2OH
nPr
H
24m
24b
24n
24n
24c
24d
24o
24a
24p
24q
24f
d
9
H
10
H
d
11
H
12
H
nPr
g
h
13
D2, E
D2
14
15
16
17
nHex
iPr
H
H
D2
CH2CH2OTBS
CH2CH2OTBS
CH2CH2OTBS
Me
TMS
H
D1
D1
g
h
18
D2, E
a
Reagents and conditions: 5:1 butadiene/alkyne, Ni(cod)2 (10 mol
b
%), PPh3 (10 mol %), rt, 16 h. Refer to Figure 3 for reaction
outcomes. Main product: [2 + 2 + 2] adducts. 20:1 butadiene/
alkyne. Main product: Cope; GC/MS shows traces of 10-membered
c
d
e
f
ring. Main product: linear coupling products (31, cf. Scheme 5d).
g
h
40:1 butadiene/alkyne. The macrocyclic product was characterized
both directly and after derivatization.
NMR. If crude analysis was insufficient and separation was
possible by silica gel chromatography, purification was
undertaken and the obtained products were reanalyzed.
Frequently, isomeric mixtures were still inseparable and were
analyzed in a “semi-pure” state by 1D and 2D NMR, as well as
HRMS.
In the event that the product was the major component after
crude analysis (categories D1, D2, and E), purification was
attempted by silica gel chromatography and preparative thin-
layer chromatography, often in combination and repeatedly. It
is worth noting that the selection of alkyne substrate was at
least partially influenced by the projected facility of purification.
Nonpolar alkynes such as isopropylacetylene (16a) could
rapidly lead to germacrene (10), but unless the reaction
performed perfectlyin terms of yield, regio- and stereo-
selectivitycompletely intractable mixtures of nonpolar
compounds would result. In addition, distillation was hampered
by the facile [3,3]-Cope rearrangement. On the other hand,
10-membered ring systems with polar functional groups were
more easily handled and separated. Nevertheless, even for these
more polar products, secondary derivatization (e.g., removal of
the protecting group from protected alcohol 24f) was often
necessary for purification (category D2).
With a working reaction in hand, namely butadiene/4-octyne
(16d, Scheme 5A), the substrate scope was explored (Table 1).
It was quickly determined that bulky substituents on the alkyne
inhibited the reaction, presumably due to steric clash around
the metal center (16g−i, entries 1−3, Table 1). Carboxylic acid
16j did not react (entry 4, Table 1), perhaps due to either
830
dx.doi.org/10.1021/jo202314a | J. Org. Chem. 2012, 77, 825−842

The Journal of Organic Chemistry
Featured Article
since: (1) The alkyne is electronically and sterically more
differentiated than 16p and 16q (entries 16 and 17, Table 1)
and was thus expected to be incorporated with higher
regioselectivity; (2) the respective 10-membered ring 24f is
less prone to undergo Cope rearrangement under the reaction
conditions, therefore enabling its isolation; and (3) the desired
product 24f would contain a suitably functionalized side chain
for elaboration into terpene natural products when applied to a
cooligomerization reaction with isoprene.
Innate Reactivity of 4Z,7Z,10E-Cyclodecatrienes. With
4Z,7Z,10E-cyclodecatriene 24f in hand, some of its general
reactivity was explored. Macrocycle 24f was first deprotected
with TBAF to generate free alcohol 24c (Scheme 6). From
this common intermediate, various derivatizations were carried
out. A transannular cyclization in acetic acid afforded the 6,6-
fused bicyclic system rac-32. While the constitution of
macrocycle 24f had been fully elucidated, the regiochemical
outcome of the alkyne insertion remained unclear due to the
complexity of the NMR spectra, even employing low
temperature experiments. After extensive efforts to synthesize
crystalline derivatives of 24c or 24f (e.g., by epoxidation with
excess m-CPBA and esterification with various substituted
benzoyl chlorides), single crystals suitable for X-ray crystallo-
graphic analysis were finally obtained from the dibrominated
derivative 33 (Scheme 6). The crystal structure confirmed the
constitution, previously determined by NMR experiments
(see the Supporting Information), and unambiguously proved
the regiochemical outcome of the alkyne insertion. Thus, the
macrocycles 24c and 24f are 4Z,7Z,10E-cyclodecatrienes with a
side chain in the 7-position.
Crabtree’s catalyst effected a selective hydrogenation,
unexpectedly reducing the most strained 10E-olefin rather
than the olefin proximal to the alcohol to yield diene 34.
Continued hydrogenation led to the expected 7Z-double bond
being reduced, affording macrocycle rac-35. In contrast, Wilke
had demonstrated the reduction of both disubstituted olefins
with H2/Raney-Ni.45 Not surprisingly, under thermal con-
ditions, the 10-membered ring 24c underwent [3,3]-Cope
rearrangement to give divinylcyclohexene rac-36. Additionally,
Heimbach and co-workers have shown that 4Z,7Z,10E-
cyclodecatrienes can be further processed to saturated hydro-
carbons and diketones by selective hydrogenation and
ozonolysis, respectively.82,83 As such, a variety of synthetically
useful building blocks can be obtained from 4Z,7Z,10E-
cyclodecatrienes, as exemplified by some selected reactions of
macrocycle 24c (Scheme 6).
Under the standard conditions, the desired product 24f was
observed, albeit with a substantial number of higher order
oligomeric side products and Reppe-type cycloadducts. In order
to suppress these side reactions, optimization of the butadiene/
alkyne 16f cooligomerization was carried out. An extensive
ligand screen including mono- and bidentate phosphines (e.g.,
P(2-furyl)3, PBu3, SPhos, and rac-BINAP), amines (NPh3),
NHCs (e.g., 1,3-bis(2,4,6-trimethylphenyl)imidazolium chlo-
ride), and arsines (AsPh3) led to either inhibition or no
significant improvement when compared to the routine ligand
(PPh3). However, two general trends were deduced: (1) The
only ligands that afforded the desired product were derivatives
of PPh3 (e.g., P(3-MeOPh)3 or P(4-ClPh)3) and (2) the use of
bidentate ligands (e.g., dppp or rac-BINAP) showed no
consumption of alkyne. Changing the reaction temperature
from room temperature to 40 °C led to a more complex
mixture, while decreasing it to 4 °C led to a lowered reaction
rate without improving the reaction outcome. Finally, the
diene/alkyne ratio was modulated in the hope of inhibiting
the formation of both [2 + 2 + 2] products and other
nonproductive pathways. A decreased ratio of 2:1 resulted in
reduced formation of 10-membered ring and more side
products, while a 40:1 stoichiometry sufficiently suppressed
side reactions involving multiple alkynes. Similarly, diluting the
reaction with toluene, Et2O, or THF (5:1 diene/alkyne,
concentration: 0.5 M) led to comparable results. Finally, the
optimized conditions (vide supra) afforded the desired product
24f in 23% isolated yield (980 mg product). Most remarkably,
macrocycle 24f was shown to be a single regioisomer by 1D
and 2D NMR studies, demonstrating the excellent selectivity of
the alkyne insertion step (see 22 → 23, Scheme 4). With this
result, it was anticipated that substrate 16f could lead to terpene
natural products if cooligomerized with isoprene instead of
butadiene.
Isoprene Cooligomerizations. The reaction of simple test
substrate 4-octyne (16d) with isoprene afforded a crude
mixture of 10-membered ring products, which was distilled to
give two regioisomeric Cope products (rac-37 and meso-37′, 1:1
1
ratio, Scheme 7A) as determined by H, 13C, and APT NMR.
The structural elucidation of vinylcyclohexenes rac-37 and
meso-37′ provided indirect proof of the constitution of the
a
Scheme 6. General Reactivity of 4Z,7Z,10E-Cyclodecatriene 24c
a
Reagents and conditions (all yields unoptimized): (a) TBAF (1.1 equiv), THF, rt, 97%; (b) AcOH, H2SO4 (cat.), rt, 46% for 32, 28% 24ad;
(c) 4-BrPhNCO (4.0 equiv), NEt3 (6.0 equiv), DMAP (0.25 mol %), DCM, rt, 38%; (d) [Ir(cod)(PCy3)(py)]PF6 (5 mol %) H2 (150 psi),
EtOAc, rt, ∼60% for 34, ∼30% for rac-35; (e) neat, 100 °C, 80%.
831
dx.doi.org/10.1021/jo202314a | J. Org. Chem. 2012, 77, 825−842

The Journal of Organic Chemistry
Featured Article
primary products: rac-37 arises from the desired 4,10-
dimethylcyclodecatriene 17d (head-to-tail connectivity), while
meso-37′ traces back to the respective 1,4-regioisomer 17d’
(tail-to-tail connectivity). From the obtained data, the presence
of the head-to-tail 1,5-dimethyl regioisomer cannot be excluded
since it converges to yield the same Cope product (rac-37) as
the 4,10-dimethylcyclodecatriene 17d. The fact that the third
possible regioisomer (i.e., with head-to-head coupled isoprene sub-
units) was not observed, demonstrated that the regioselectivity
in the initial oxidative dimerization of two C5 building blocks
can indeed be controlledat least to a certain extent.93 The
head-to-tail connectivity and 4Z,10E-geometry of 17d was
corroborated by exhaustive epoxidation of the crude reaction
mixture and subsequent X-ray crystallographic analysis of
triepoxide product (rac-38, Scheme 7B). The reaction of 17d with
16v which was fully consumed, but in unproductive fashion (entry
5, Table 2). As opposed to the observations for butadiene, simple
a
Table 2. Ni0-Catalyzed Isoprene/Alkyne Cooligomerization
b
entry
alkyne
R1
R2
product outcome
1
2
16r
TMS
TMS
H
17r
A
A
A
A
B
B
B
B
B
16s
TMS
H
17s
c
3
16t
(CH2)2CO2H
CH2OH
17t
4
16u
16v
16w
16o
16a
16x
16y
16n
16z
16aa
16ab
16ac
16q
16l
CH2OH
H
17u
17v
17w
17o
17a
17x
17y
17n
17z
17aa
17ab
17ac
17q
17l
5
(CH2)4CO2Me
Et
nHex
iPr
iPr
6
H
a
7
H
Scheme 7. (A) Isoprene Cooligomerization Model Study,
8
H
and (B) Crystal Structure of Triepoxide rac-38
9
Et
d
10
11
12
13
14
15
16
Ph
Bu
H
B
B
B
B
B
B
B
C
C
C
C(CH3)2OTMS
CH2OTHP
CH2CH2OTHP
CH2CH2OMe
CH2CH2OTMS
CH2CH2OTBS
CO2Et
H
H
H
H
TMS
H
e
f
17
g
f
18
16b
16p
16f
C(CH3)2OH
CH2CH2OTBS
CH2CH2OTBS
nPr
H
17b
17p
17f
h
f i
19
Me
H
nPr
e,g
20
21
C ,
16d
17d
D2
a
Reagents and conditions: 5:1 isoprene/alkyne, Ni(cod)2 (10 mol %),
b
PPh3 (10 mol %), 60 °C, 16 h. Refer to Figure 3 for reaction
outcomes. Toluene was used as a cosolvent (0.5 M). A trace of Cope
product was observed. Alkyne was added dropwise; not under
c
d
e
f
g
pressure. Main product: [2 + 2 + 2] adduct. 40:1 butadiene/alkyne.
h
i
Main product: Cope product (see the Experimental Section). Main
product: linear trimer (see the Experimental Section).
terminal alkyl alkynes 16a, 16o, and 16w (entries 6−8, Table 2)
typically resulted in complex mixtures. Similarly, alkyl, phenyl, and
oxygenated unsymmetrically disubstituted alkynes (16n, 16q, and
16x-16ac, entries 9−16, Table 2) were consumed, but also
resulted in inseparable mixtures. As expected from the results
obtained for butadiene (cf. Table 1), substrates 16l and 16b
(entries 17 and 18, Table 2) showed [2 + 2 + 2] adducts as the
primary products. Disappointingly, the substrates which had
provided the most promising results for butadiene, namely alkynes
16p and 16f (entries 19 and 20, Table 2), yielded mixtures of [2 +
2 + 2], linear, Cope, and other unidentifiable products. Protected
alcohol 16p (entry 19, Table 2) afforded a Cope product as
the major component (see the Experimental Section), demon-
strating the intermediacy and thermal instability of the desired
10-membered ring. Being aware that dilution had significantly
improved the butadiene/16f cooligomerization, a dropwise
addition of protected alcohol 16f was undertaken (entry 20,
Table 2). As expected, the reaction performed poorly in the
absence of a pressure vessel, but there was a noticeable increase in
product formation by GC/MS analysis. Unfortunately, the mixture
was still too complex to allow for isolation, and further attempts to
improve the reaction were met with failure. In order to amend
these results, a variety of additives were screened, including
Brønsted (e.g., AcOH, Montmorillonite K10) and Lewis acids
a
Reagents and conditions: (a) 5:1 butadiene/alkyne, Ni(cod)2 (10
mol %), PPh3 (10 mol %), 16 h; (b) distillation.
excess m-CPBA afforded a mixture of three major components,
two of which were determined to be diastereomers by X-ray
crystallography, both arising from the desired head-to-tail coupled
10-membered ring 17d.94 At this point, the regioselectivity
problem arising from the alkyne insertion was successfully
overcome (cf. 24f), and the formation of regioisomeric mixtures
(head-to-tail selectivity) during the oxidative dimerization of
isoprene was partially solved (cf. 17d).
After successfully applying Wilke’s reaction conditions to
isoprene/4-octyne (16d), the substrate scope was explored in
further detail (Table 2). Substrates 16r−u (entries 1−4, Table 2)
were not reactive under these standard conditions. This can be
rationalized for the carboxylic acid and the diol, in agreement with
the results obtained with butadiene (entries 4 and 12, Table 1):
Coordination events and/or the presence of an acidic proton may
inhibit or shut down the reactivity of the catalytically active nickel
species. This hypothesis was substantiated by testing methyl ester
832
dx.doi.org/10.1021/jo202314a | J. Org. Chem. 2012, 77, 825−842

The Journal of Organic Chemistry
Featured Article
(e.g., MgBr2·Et2O, TiCl2(iOPr)2), bases (e.g., DBU, Cs2CO3),
cometals (e.g., PtCl4, AgNO3), solvents (e.g., DMSO, hexafluoro-
2-propanol), and salts (e.g., KI, Bu4NBr). However, none of these
led to any improvement in the reaction.
Scheme 8. Attempted Installation of Missing Germacrene
Carbons on a Macrocycle Derived from Butadiene
a
Despite all of the substrates and conditions explored, none of
the cooligomerizations employing isoprene proceeded with
sufficient selectivity. Even when utilizing the conditions
optimized for butadiene (vide supra), the isoprene/alkyne 16f
cooligomerization resulted in a complex mixture. Furthermore,
the reaction of isoprene with one of the simplest substrates,
1-octyne (16o), revealed a general limitation: While the
butadiene/1-octyne (16o) system had provided a clean reaction
to form the desired 10-membered ring, the outcome was
completely different with isoprene, in which case the desired
macrocyclic product 17o was not observed. Instead, formation
of different isomers of the rearranged Cope product
predominated. This showcases an inherent problem related to
the use of isoprene as the cooligomerization partner, since
substituted 4,10-dimethyl-4Z,7Z,10E-cyclodecatrienesthe de-
sired productscan be expected to undergo Cope rearrange-
ment under the reaction conditions (60 °C). For this reason,
any optimization cannot overcome the main obstacle: The
thermal instability of substituted 4,10-dimethyl-4Z,7Z,10E-
cyclodecatrienes. While significantly lowering the reaction
temperature may preclude Cope rearrangement, the rate of
the reaction would then become the limiting factor and would
prevent its synthetic utility.
Delayed Installation of C1 Units onto 4Z,7Z,10E-
Cyclodecatriene 24c. With the inherent flaw in the
isoprene/alkyne cooligomerization revealed, the butadiene
system 24c (obtained by deprotection of macrocycle 24f,
Scheme 6) was reevaluated. It was envisioned that subsequent
installation of the “missing” methyl groups on macrocycle 24c
could still enable access to germacrene natural products
(Scheme 8). Most importantly, this transformation would
have to proceed with high regio- and chemoselectivity to afford
only one out of four possible regioisomers (rac-39−rac-42,
Scheme 8). Epoxidation followed by subsequent SN2 reaction
with a methyl nucleophile was ruled out because backside
attack on the σ* orbitals is blocked by the macrocycle. Utilizing
a cycloaddition strategy would install the necessary carbon
atom while maintaining the oxidation state of the endocyclic
olefin. Cycloadditions with 1,1-dibromoformaldoxime have
been shown to proceed with high regioselectivity at room
temperature,95 which was crucial to avoid Cope rearrangement.
The intended endgame would then involve fragmentation of
the resulting isoxazoline moieties to the bis-β-hydroxynitrile
rac-43 with Raney-Ni96,97 or TMSCl/NaI.98 Following
elimination of the secondary alcohols, a vinyl nitrile cross-
coupling based closely on the work of Dankwardt could be
employed in order to access the lowly oxidized germacrene
skeleton 44.99
In practice, 1,1-dibromoformaldoxime afforded the desired
isoxazoline rac-45, unfortunately as the minor regioisomer
(2.7:1 ratio, Scheme 8). Regioisomers rac-45 and rac-46 were
separated and subjected to the same [3 + 2]-cycloaddtion
conditions.100 In both cases the respective diadducts (rac-39−
rac-42) were formed with no observed regioselectivity (1:1
ratio). Although good chemoselectivity was achieved (i.e., the
trisubstituted olefin remained intact), it proved impossible to
alter the regioselectivity in either cycloaddition by changing the
solvent, temperature, or base. Thus, the dicycloadducts were
not elaborated to natural products because of a lack of
a
Reagents and conditions (all yields unoptimized): (a) 1,1-
dibromoformaldoxime (excess), NaHCO3 (excess), EtOAc, rt, 21%
for rac-46, 58% for rac-47; (b) 1,1-dibromoformaldoxime (excess),
NaHCO3 (excess), EtOAc, rt, 47% for rac-39:rac-40 (1:1 ratio by
NMR) and 41% for rac-41:rac-42 (1:1 ratio by NMR).101
regiocontrol during the dipolar cycloaddition. More impor-
tantly, this sequence deviated significantly from the initial
proposal of an ideal approach, namely utilizing isoprene for the
total synthesis of terpenes.
CONCLUSION
■
In summary, this full account has traced our extensive studies
aiming to mimic Nature’s cyclase phase of terpene synthesis.
Wilke’s and Heimbach’s pioneering studies in both diene
oligomerization and organonickel chemistry were vital in our
efforts. In this publication, a comprehensive list of their original
publications related to diene (co)oligomerization, many of
which are not easily available, was compiled and efforts to
translate and summarize details from the original German texts
833
dx.doi.org/10.1021/jo202314a | J. Org. Chem. 2012, 77, 825−842

The Journal of Organic Chemistry
Featured Article
and either (1) p-anisaldehyde in ethanol/aqueous H2SO4/CH3CO2H
or (2) 2% KMnO4 in 4% aqueous sodium bicarbonate and heat as
developing agents. Nuclear magnetic resonance (NMR) spectra were
calibrated using residual undeuterated solvent signals as an internal
reference (CHCl3: 7.26 ppm for 1H NMR, 77.16 ppm for 13C NMR).
1D and 2D NMR spectra were recorded at room temperature unless
otherwise stated. All NMR peak assignments given in the Supporting
Information were determined by 1H, 13C, APT, COSY, HMQC, and/
or HMBC experiments. The following abbreviations were used to
assign NMR signal multiplicities: s = singlet, d = doublet, t = triplet,
q = quartet, m = multiplet, br = broad. High-resolution mass spectra
(HRMS) were recorded on an LC/MSD TOF mass spectrometer by
electrospray ionization time-of-flight reflectron experiments (ESI-TOF).
For IR spectra the major absorbance bands are reported in
wavenumbers. Melting points (m.p.) are uncorrected. GC/MS
analyses were carried out applying the following method: Agilent
19091S-433 column (30 m × 320 μm × 0.25 μm); H2 30 mL/min, air
400 mL/min, He 25 mL/min; 50 °C for 2.25 min, ramp to 300 °C
(60 °C/min), and hold at 300 °C for 4 min.
were made. Reinvestigating their work in a more modern
context, an attempt was made to push the limits of their
methods to the total synthesis of terpene natural products. The
ambitious goal was to establish a preparatively useful method
for the creation of simple, yet synthetically challenging terpenes
such as the germacrenes, in two to three steps by a controlled
Ni0-mediated isoprene/alkyne cooligomerization. Lasting les-
sons from these studies include:
1 It was realized through experimental findings that the
vision of terpene synthesis from isoprene/alkyne cool-
igomerization, in the present form, is inherently flawed.
Most significantly, the products, 4Z,7Z,10E-cyclodeca-
trienes, are thermally unstable under the reaction
conditions. The number of regio- and stereoisomers
and inseparable side products led to tedious and often
impossible separations that resulted in serious analytical
issues.
Preparation of Bis(1,5-cyclooctadiene)nickel(0).62 Since the
outcome of the Ni(cod)2 formation is strongly dependent on the
quality of Ni(acac)2, the latter was first azeotropically distilled in
refluxing toluene (130 °C) with a Dean−Stark apparatus to remove
trace amounts of water. The flask was cooled to ∼90 °C and filtered
through Celite while hot. Toluene was removed in vacuo to afford
anhydrous Ni(acac)2 (forest green solid). Freshly dehydrated
Ni(acac)2 (2.0 g, 7.78 mmol) were dissolved in toluene (4.8 mL, if
not fully soluble the azeotropic distillation should be repeated) in a
flame-dried Schlenk flask. 1,5-cyclooctadiene (4.78 mL, 38.9 mmol,
5 equiv) was added and the mixture degassed by either three freeze−
pump−thaw cycles or by sonication for 15 min with concomitant
vigorous argon bubbling. The degassed homogeneous mixture was
cooled to −78 °C and butadiene (∼0.5 mL, 5.91 mmol, 0.76 equiv)
was condensed into the flask. The solution was warmed to −10 °C and
AlEt3 (0.9 M solution in toluene, 19.5 mL, 2.25 equiv) added
dropwise. The reaction was then allowed to come to room
temperature and stirred overnight (8−10 h). The resulting red-
brown suspension was cooled to −30 °C and the supernatant carefully
removed by cannula to give a yellow solid, which was washed with
degassed Et2O (the suspension was allowed to equilibrate to −30 °C
between washings). Analytically pure Ni(cod)2 (1.24 g, 4.51 mmol,
58%) was obtained as an air-sensitive yellow solid, dried in vacuo in
the Schlenk flask and stored under argon (preferentially inside a
glovebox) at low temperatures. In general, yields ∼60% Ni(cod)2 were
obtained on up to 10 g scale: 1H NMR (C6D6, 400 MHz) δ = 4.30 (s,
8 H), 2.08 (s, 16 H) ppm; 13C NMR (C6D6, 400 MHz) δ = 89.7
(CH), 30.9 (CH2) ppm.
2 Despite the complexity of the reaction mixtures, several
repeatedly occurring side products from the diene/alkyne
cooligomerization were isolated and fully elucidated.
This allowed for the establishment of a rapid qualitative
analysis of crude reaction mixtures.
3 In spite of intense efforts, the outcome of the diene/
alkyne cooligomerization was only minimally tunable by
variation of the reaction conditions. For example, the
ligand was incapable of significantly altering the outcome
of the reaction−in terms of both olefin geometry and
regioselective incorporation of isoprene or alkyne.
Ultimately, it was observed that the reaction is
contingent primarily upon the alkyne substrate.
4 With respect to butadiene, notable advances in substrate
scope were achieved. For the first time, both terminal
and α-branched alkynes were successfully incorporated
into the 4Z,7Z,10E-cyclodecatriene frameworks. Most
remarkably, full regiocontrol was obtained in the
incorporation of the unsymmetrical alkyne 16f, as
unambiguously established by the crystallographic
analysis of macrocycle 33.
5 Despite its propensity to undergo Cope rearrangement, a
crystal structure of 4,10-dimethyl-4Z,7Z,10E-cyclodeca-
diene 17d was obtained, as the triepoxide rac-38. This
directly confirmed the product structure, solid-state
conformation, and olefin geometry for the first time.
Although, ultimately, terpenes could not be produced from
isoprene due to several intractable factors, the cooligomeriza-
tion of alkynes with butadiene may still be useful for total
syntheses in a different context. Such studies, as well as the
utilization of the bulk chemical isoprene as a building block
for terpene synthesis, continue to be of interest in our
laboratory.
General Procedure for Ni0-Catalyzed Cooligomerization of
Butadiene and Alkynes 16. A 35 mL pressure vessel and stir bar
were flame-dried and brought into a glovebox. Freshly prepared Ni(cod)2
(79.0 mg, 0.29 mmol, 10 mol %) and PPh3 (75.0 mg, 0.29 mmol, 10 mol
%) were added to the reaction flask.103 The vessel was sealed with a
septum, brought out of the glovebox, connected to a Schlenk line and
cooled to −78 °C. Under an argon atmosphere, 1,3-butadiene (∼1.25−
10 mL, 14.4−115 mmol, 5−40 equiv) was condensed into the flask, and
the alkyne 16 (2.87 mmol) was added. Under a steady flow of argon, the
septum was rapidly exchanged with a screw cap. The reaction was allowed
to come to room temperature and stirred for 18 h. For safety reasons
the reactions were conducted behind an explosion-proof blast shield.
The reaction mixture was cooled to −78 °C, uncapped, and excess
1,3-butadiene was allowed to evaporate at room temperature. Sub-
sequently, the reaction was worked up according to one of the fol-
lowing two methods:
EXPERIMENTAL SECTION
■
General Methods. All reactions were carried out under an argon
atmosphere with dry solvents under anhydrous conditions, unless
otherwise noted. Dry tetrahydrofuran (THF), triethylamine (NEt3),
toluene, dichloromethane (DCM), and diethyl ether (Et2O) were
obtained by passing commercially available predried, oxygen-free
formulations through activated alumina columns. Reagents were
purchased at the highest commercial quality and used without further
purification unless otherwise stated. Dibromoformaldoxime was
synthesized according to a literature-known procedure.102 Reactions
were monitored by thin-layer chromatography (TLC), carried out on
0.25 mm silica gel plates (60F-254) using UV light as visualizing agent
(A) Wilke’s Standard Workup Procedure. The residue was taken up
in EtOAc (50 mL) and washed subsequently with 5 M HCl, 5%
834
dx.doi.org/10.1021/jo202314a | J. Org. Chem. 2012, 77, 825−842

The Journal of Organic Chemistry
Featured Article
aqueous H2O2, saturated aqueous sodium bicarbonate solution, water,
and finally brine. The organic phase was dried over magnesium sulfate,
filtered, and concentrated in vacuo. Purification by flash chromatog-
raphy then followed, if necessary and if the crude mixture was deemed
separable (i.e., categories C, D1, D2, and E104).
(B) Modified Workup Procedure. The residue was suspended in
hexanes, loaded directly onto a plug of silica gel, and eluted
sequentially with hexanes or appropriate hexanes/EtOAc mixtures.
The solvent was removed in vacuo, and the crude mixture thus
obtained was further purified by flash chromatography, if necessary
and if it was deemed separable (i.e., categories C, D1, D2, and
E106).
Isolation of the Linear Cooligomer 31 (Category C104). The
reaction was performed according to the standard procedure, however, with
altered stoichiometry: Homopropargyl alcohol 16c (806 mg,
0.87 mL, 11.5 mmol), 1,3-butadiene (∼5.0 mL, 57.3 mmol, 5.0 equiv),
Ni(cod)2 (106 mg, 0.39 mmol, 3.3 mol %), PPh3 (100 mg, 0.39 mmol, 3.3
mol %). Workup A afforded a pale yellow oil. TLC of the obtained crude
mixture showed four major components along with several minor side
products. This was confirmed by GC/MS analysis, which revealed a fairly
clean and selective reaction with full consumption of starting material giving
1,5-cyclodecadiene, Cope product rac-30, traces macrocycle 24c (the latter
two were inseparable and were tentatively assigned by 1H and 13C NMR of a
crude mixture), and one further, unknown compound. Subsequently, an
aliquot (20 mg, crude) of the crude mixture was purified by preparative TLC
(silica gel, EtOAc/hexanes, 1:1) yielding the main product, diol 31, in
analytically pure form. Its structure was fully elucidated by 2D NMR experi-
General Procedure for Ni0-Catalyzed Cooligomerization of
Isoprene and Alkynes 16. A 35 mL pressure vessel equipped with
a stir bar was flame-dried. In a glovebox freshly prepared Ni(cod)2
(79.0 mg, 0.29 mmol, 10 mol %) and PPh3 (75.0 mg, 0.29 mmol, 10 mol %)
were added to the reaction flask. The vessel was sealed with a septum,
brought out of the glovebox, and connected to a Schlenk line. Under
an argon atmosphere, isoprene (1.5−12 mL, 14.4−115 mmol, 5−40
equiv) and alkyne 16 (2.87 mmol) were added.103 Under a steady flow
of argon, the septum was rapidly exchanged with a screw cap, the
mixture heated to 60 °C and stirred for 18 h. The reaction mixture was
then cooled to room temperature and excess isoprene was allowed to
evaporate under a flow of nitrogen. The remaining crude material was
subsequently worked up by either method A or B (vide supra).
1
ments. Rf = 0.76 (silica gel, hexanes/EtOAc, 1:1); H NMR (500 MHz,
CDCl3) δ = 5.80 (ddt, J = 17.1, 10.4, 6.7 Hz, 1 H, H15), 5.71 (t, J = 7.3 Hz, 1
H, 8-H), 5.50−5.34 (m, 2 H, 10-H, 11-H), 4.99 (dd, J = 17.3, 1.3 Hz, 1 H, 16-
H-trans), 4.94 (dt, J = 10.1, 1.1 Hz, 1 H, 16-H-cis) 3.79−3.73 (m, 4 H, 1-H,
5-H), 2.95 (t, J = 6.7 Hz, 2 H, 9-H), 2.63 (t, J = 6.2 Hz, 2 H, 2-H), 2.37 (t, J =
5.9 Hz, 2 H, 6-H), 2.00 (m, 4 H, 12-H, 14-H), 1.44 (dt, J = 14.4, 7.4 Hz, 2 H,
14-H) ppm; 13C NMR (125 MHz, CDCl3) δ = 138.9 (C-15), 137.7 (C-8),
131.4 (C-11), 127.5 (C-10), 119.8 (C-7), 114.6 (C-16), 91.7 (C-3), 80.4 (C-
4), 61.5 (C-5), 61.4 (C-1), 40.5 (C-6), 33.9 (C-9), 33.4 (C-14), 32.1 (C-12),
28.8 (C-13), 24.1 (C-2) ppm; IR (ATR, neat) νmax 3335, 2923, 2853, 1640,
−1
1455, 1431, 1377, 1329, 1260, 1178, 1090, 1041, 967, 909, 845, 801, 699 cm ;
HRMS calcd for C16H24O2 249.1849 [M + H]+, found (ESI-TOF) 249.1852.
Synthesis of 10-Membered Ring Systems: Isolation and Char-
acterization of Primary Macrocycles, Secondary Derivatives,
and Undesired Side Products. Isolation of [2 + 2 + 2]
Cycloadducts from the Reaction of Alkyne 16e and Butadiene
(Category C104). The standard procedure was followed using
dimethyl acetylenedicarboxylate (16e, 1.62 g, 11.5 mmol) and 1,3-
butadiene (∼5 mL, 57.3 mmol, 5.0 equiv) to give a brown viscous oil.
GC/MS analysis of the obtained crude mixture allowed for the
tentative assignment of the above three homo- and hetero-[2 + 2 + 2]
cycloadducts 26−28 as the main products (for GC trace see the
Supporting Information).
Direct Characterization of 24d (Category E104). The standard
procedure was followed using 4-octyne (16d, 316 mg, 0.42 mL,
2.87 mmol) and 1,3-butadiene (∼10 mL, 115 mmol, 40 equiv) to give the
colorless oil 24d (518 mg, 2.37 mmol, 83%; this material contains 9% Cope
product meso-25 and 12% of unknown impurities according to GC/MS
1
analysis) applying workup procedure B: Rf = 0.72 (silica gel, hexanes); H
NMR (500 MHz, CDCl3, −50 °C) δ = 5.47 (t, J = 10.2 Hz, 1 H), 5.28
(ddd, J = 15.1, 10.4, 4.5 Hz, 1 H), 5.18 (m, 2 H), 3.22 (t, J = 12.5 Hz, 1 H),
2.92 (dd, J = 15.1, 10.8 Hz, 1 H), 2.57 (d, J = 14.6 Hz, 1 H), 2.22 (ddd, J =
24.6, 12.0, 5.8 Hz, 2 H), 2.11−1.95 (m, 3 H), 1.91 (dd, J = 23.0, 11.5 Hz, 1
H), 1.79 (m, 1 H), 1.63 (td, J = 11.9, 4.0 Hz, 1 H), 1.52 (q, J = 11.5 Hz, 1
H), 1.41−1.27 (m, 3 H), 1.24−1.17 (m, 1 H), 0.91−0.84 (m, 6 H) ppm;
13C NMR (125 MHz, CDCl3, −50 °C) δ = 134.5, 132.5, 132.4, 131.8,
129.4, 122.8, 38.7, 37.7, 35.6, 32.2, 31.4, 26.2, 22.4, 21.7, 14.9, 14.5 ppm; IR
(ATR, neat) νmax 3007, 2956, 2928, 2867, 1453, 1376, 1202, 1183, 1123,
Isolation of [2 + 2 + 2] Cycloadducts S1 and S2 from the
Reaction of Alkyne 16l and Butadiene (Category C104). The
standard procedure was followed using alkyne 16l (1.13 g, 11.5 mmol)
and 1,3-butadiene (∼5 mL, 57.3 mmol, 5.0 equiv) to give a yellow oil.
The main components were isolated in analytically pure form by
preparative TLC (silica gel, hexanes/EtOAc, 12:1). GC/MS analysis
and H NMR of the obtained products allowed for the unambiguous
assignment of the [2 + 2 + 2] cycloadducts S1 and S2 (GC trace given
in the Supporting Information).
1
−1
1088, 970, 950, 911, 842, 813, 738, 704 cm ; HRMS calcd for C16H26
219.2107 [M + H]+, found (ESI-TOF) 219.2107.
835
dx.doi.org/10.1021/jo202314a | J. Org. Chem. 2012, 77, 825−842

The Journal of Organic Chemistry
Featured Article
added. The reaction mixture was washed with H2O and brine, dried
over MgSO4, and concentrated in vacuo. Preparative TLC (silica gel,
hexanes/EtOAc 20:1) afforded the above dihydroisoxazole derivative
S4 (racemic mixture) as the only isolable regioisomer (colorless oil,
11.3 mg, 37.9 μmol, 18% estimated yield over two steps based on
Characterization of 24d in Form of Its Respective Cope
Product (meso-25, Category D2104). Macrocycle 24d (10.0 mg,
45.8 μmol) was heated neat to 120 °C for 1 h. The reaction afforded
the clear yellow oil, divinylcyclohexene meso-25 in quantitative yield
alkyne 16a):106 Rf = 0.50 (silica gel, hexanes/EtOAc, 20:1); H NMR
1
(600 MHz, CDCl3) δ = 5.77 (ddd, J = 10.8, 6.6, 6.0 Hz, 1 H, H-5),
5.34−5.24 (m, 2 H, H-4, H-8), 4.43 (d, J = 13.9 Hz, 1 H, H-10), 3.21
(t, J = 12.0 Hz, 1 H, H-6), 3.09 (t, J = 12.7 Hz, 1 H, H-1), 2.83 (t, J =
12.7 Hz, 1 H, H-9), 2.71−2.64 (m, 1 H, H-9), 2.44−2.31 (m, 3 H, H-
3, H-6, H-11), 2.12−2.07 (m, 1 H, H-3), 1.93 (t, J = 13.9 Hz, 1 H, H-
2), 1.59−1.53 (m, 1 H, H-2), 1.07−1.03 (m, 6 H, H-12, H-13) ppm;
13C NMR (150 MHz, CDCl3) δ = 150.4 (C-7), 146.8 (C-14), 129.7
(C-5), 128.1 (C-4), 114.6 (C-8), 89.6 (C-10), 49.8 (C-1), 34.5 (C-
11), 30.3 (C-2), 29.9 (C-9), 27.5 (C-6), 23.7 (C-3), 23.1 (C-12), 22.0
(C-13) ppm; IR (ATR, neat) νmax: 3011, 2959, 2924, 2866, 1611,
1571, 1466, 1447, 1381, 1356, 1294, 1251, 1130, 1100, 1082, 1045,
1009, 999, 963, 919, 878, 825, 799, 758, 705 cm−1; HRMS calcd for
C14H20BrNO 298.0801 [M + H]+, found (ESI-TOF) 298.0799.
1
(9.90 mg, 45.3 μmol, 99%): Rf = 0.66 (silica gel, hexanes); H NMR
(400 MHz, CDCl3) δ = 5.79 (ddd, J = 17.0, 10.6, 7.5 Hz, 2 H), 5.01
(ddd, J = 8.6, 2.0, 0.8 Hz, 2 H), 4.98 (d, J = 0.8 Hz, 2 H), 2.41 (dd, J =
11.8, 4.9 Hz, 2 H), 2.17−1.87 (m, 8 H), 1.44−1.32 (m, 4 H), 0.88 (t, J
= 7.3 Hz, 6 H) ppm. 13C NMR (125 MHz, CDCl3) δ = 140.6, 128.7,
114.4, 41.9, 35.2, 33.5, 21.7, 14.4 ppm; IR (ATR, neat) νmax 3075,
2957, 2929, 2900, 2871, 1639, 1465, 1455, 1434, 1377, 994, 967, 910,
105
•+
719 cm−1; GC/MS calcd for C16H26 218.2035 [M] , found (FID
)
218.2.
Direct Characterization of 24p in Crude Form (Category
D1104). The standard procedure was followed using alkyne 16p
(569 mg, 2.87 mmol) and 1,3-butadiene (∼10 mL, 115 mmol, 40 equiv).
In order to establish the constitution of the isolated component(s), the
following protocol was undertaken: After the standard workup
procedure A, GC/MS analysis of the crude mixture revealed a fairly
clean reaction, with almost all starting material consumed after 18 h.
Based on the GC/MS chromatogram, it was determined that the
reaction contained 10-membered ring 24p (tR = ∼6.0 min, broad) and
traces of Cope products (tR = ∼5.9 min). Column chromatography
(silica gel, silica gel, hexanes → hexanes/EtOAc 100:1) afforded the
main product of the reaction, 24p (576 mg, 1.89 mmol, 66%), as an
inseparable mixture of two components. 13C and APT NMR excluded
the presence of Cope product (no olefinic methylene groups with a
characteristic chemical shift of ∼115 ppm). Evidence for the formation
Characterization of 24o after Cope Rearrangement (Cat-
egory D2104). The standard procedure was followed using 1-octyne
(16o, 316 mg, 0.42 mL, 2.87 mmol) and 1,3-butadiene (∼1.25 mL,
14.4 mmol, 5.0 equiv). After workup A, the crude mixture containing
24o was distilled directly to afford the Cope product S3 (racemic
mixture) as a pale yellow oil (564 mg, 2.58 mmol, 70% over two
1
steps): Rf = 0.66 (silica gel, hexanes); H NMR (500 MHz, CDCl3)
1
δ = 5.85−5.76 (m, 2 H), 5.38−5.35 (m, 1 H), 5.04−4.98 (m, 4 H),
2.49−2.37 (m, 2 H), 2.22−2.19 (m, 1 H), 2.18−2.15 (m, 1 H), 2.06−
1.82 (m, 4 H), 1.43−1.22 (br m, 8 H), 0.88 (t, J = 6.9 Hz, 3 H) ppm;
13C NMR (150 MHz, CDCl3) δ = 140.5, 140.4, 136.5, 119.0, 114.5,
114.5, 41.7, 41.2, 37.9, 32.3, 31.9, 29.3, 29.2, 27.8, 22.8, 14.3 ppm; IR
(ATR, neat) νmax 3076, 2956, 2925, 2856, 1639, 1457, 1437, 1378,
994, 910, 819, 722 cm−1; GC/MS calcd for C16H26 218.2035 [M]•+,
found (FID105) 218.2.
of 10-membered ring was demonstrated by the resolution of the H
NMR at −54 °C: Upon cooling the sample, the broad signals of the
CH2 groups of the ring system became sharp due to the slowed down
molecular motion of the conformationally flexible macrocycle.
Observing two methylene carbons adjacent to an oxygen, two methyl
groups joined to sp2 carbons, and twelve olefinic carbons, the reaction
was determined to be a 1.6:1 mixture (cf. integrals of the two methyl
groups attached to olefinic carbons C-7 or C-8) of regioisomers, which
coeluted during column chromatography and GC analysis. Further
characterization of the ring system and partial assignment of
substructures of the two regioisomers was possible by full 2D NMR
analysis (COSY, HMQC, HMBC) at low temperature. Chromato-
grams and spectra are included in the Supporting Information. 24p
(mixture of two regiosiomers): Rf = 0.75 (silica gel, DCM/Et2O, 2:1);
1H NMR (500 MHz, CDCl3, −54 °C) δ = 5.48 (dt, J = 11.3, 10.4 Hz,
1 H), 5.12−5.34 (m, 3 H), 3.53 (t, J = 6.5 Hz, 2 H), 3.20−3.40 (m, 2
H), 2.90−3.03 (m, 1 H), 2.57 (q, J = 13.5 Hz, 3 H), 2.17−2.41 (m, 3
H), 1.85−2.14 (m, 4 H), 1.69 (s, 3 H), 1.52 (q, J = 11.5 Hz, 1 H), 0.86
(s, 9 H), 0.043 (br s, 6 H) ppm; 13C NMR (150 MHz, CDCl3, rt) δ =
132.1, 131.7, 130.1, 129.8, 128.0, 123.3, 62.5, 41.0, 40.3, 38.2, 34.9,
32.8, 26.5, 26.1 (two signals), 22.6, −5.1 (two signals) ppm; IR (ATR,
neat) νmax 3007, 2952, 2928, 2895, 2857, 1471, 1462, 1445, 1382,
1361, 1253, 1089, 1005, 971, 937, 913, 833, 812, 773, 731, 703, 661
cm−1; HRMS calcd for C19H34OSi 307.2452 [M + H]+, found (ESI-
TOF) 307.2449.
Characterization of 24a in Form of Its Dihydroisoxazole
Derivative S4 (Category D2104). The standard procedure was
followed using alkyne 16a (196 mg, 0.29 mL, 2.87 mmol) and 1,3-
butadiene (∼1.25 mL, 14.4 mmol, 5.0 equiv) to afford a crude mixture
(486 mg) containing 24a after workup A. An aliquot of this mixture
(35 mg) was brought up in EtOAc (2.0 mL), and NaHCO3 (100 mg,
1.19 mmol, excess) and dibromoformaldoxime (40 mg, 0.20 mmol,
approximately 1.0 equiv) were added sequentially. The reaction was
stirred at room temperature for 30 min and additional dibromo-
formaldoxime (40 mg, 0.20 mmol, approximately 1.0 equiv) was
836
dx.doi.org/10.1021/jo202314a | J. Org. Chem. 2012, 77, 825−842

The Journal of Organic Chemistry
Featured Article
Direct Characterization of 24q in Crude Form (Category
D1104). The standard procedure was followed with alkyne 16q
(736 mg, 2.87 mmol) and 1,3-butadiene (∼1.25 mL, 14.4 mmol, 5.0 equiv).
Different from the standard procedure the reaction mixture was stirred
for 96 h (due to the slower reaction rate for the sterically more
demanding alkyne) with monitoring by GC/MS, the final alkyne 16q
(tR = 4.62 min) to product 24q (tR = 6.20 min, broad) ratio being
∼65:35. The GC/MS analysis of the crude mixture revealed at least
31.0, 29.5, 25.9 ppm; IR (ATR, neat) νmax 3335, 3007, 2928, 2885,
2856, 1471, 1437, 1361, 1252, 1194, 1041, 869, 833, 773, 706,
666 cm−1; HRMS calcd for C12H18O 179.1430 [M + H]+, found (ESI-
TOF) 179.1429.
t
four peaks showing product mass (m/z = 307.2, [M − Bu]+).
Following workup A, the crude material was subjected to column
chromatography (silica gel, hexanes → hexanes/EtOAc 9:1) to remove
unreacted starting material, PPh3, and PPh3O. The obtained mixture
1
1
was analyzed by H and 13C NMR. H NMR revealed traces of Cope
product (characteristic olefinic methylene groups with a chemical
shift of ∼115 ppm) as well as the typical broadening associated
with the conformational flexibility of the desired 10-membered
ring system. The 13C NMR showed at least four methylene
carbons bonded to oxygen, indicating four major products which
incorporated alkyne 16q. Further, unambiguous identification and
assignment proved unsuccessful, but based on the obtained data it
is likely that the main components comprise the two regioisomers
(with regard to the alkyne incorporation) of the expected
macrocycle 24q and the follow-up Cope product. The obtained
spectral data and GC traces are provided in the Supporting
Information.
Isolation of the Main Product S6 from the Reaction of
Alkyne 16p with Isoprene (Category C104). The standard
procedure was followed with alkyne 16p (569 mg, 2.87 mmol) and
isoprene (1.5 mL, 14.4 mmol, 5.0 equiv). Workup A afforded a
complex product mixture showing >10 products by TLC, while GC
revealed the crude material to contain ∼20 components. Despite the
complexity of the crude mixture, isolation of the title compound was
achieved by column chromatography (silica gel, hexanes/EtOAc,
1
250:1) followed by H NMR analysis of single column fractions. The
constitution of the Cope product S6 was unambiguously established
by 2D NMR experiments (see the Supporting Information):106 1H
NMR (600 MHz, CDCl3) δ = 4.76 (s, 2 H, H-2, H-3), 4.64 (d, J = 9.2
Hz, 2 H, H-2, H-3), 3.61 (t, J = 7.5 Hz, 2 H, H-12) 2.51−2.44 (m, 2 H,
H-6, H-10), 2.33−2.27 (m, 1 H, H-11), 2.26−2.20 (m, 1 H, H-11),
2.17−2.06 (m, 4 H, H-5, H-9), 1.73 (s, 3 H, H-18), 1.72 (s, 3 H, H-
17), 1.65 (s, 3 H, H-16), 0.89 (s, 9 H, H-15), 0.05 (s, 6 H, H-13) ppm;
13C NMR (150 MHz, CDCl3) δ = 148.1 (C-1, C-4), 127.5 (C-8),
125.9 (C-7), 110.3 (C-2, C-3), 61.9 (C-12), 42.8 (C-5), 42.5 (C-10),
37.2 (C-11), 36.1 (C-9), 34.1 (C-6), 26.1 (C-15), 23.9 (C-18), 23.8
(C-17), 19.0 (C-16), 18.5 (C-14), −5.1 (C-13) ppm; IR (ATR, neat)
νmax 3084, 2954, 2928, 2857, 1643, 1462, 1375, 1361, 1253, 1087,
1005, 887, 834, 811, 775, 733, 662 cm−1; HRMS calcd for C21H38OSi
335.2765 [M + H]+, found (ESI-TOF) 335.2763.
Direct Characterization of 24f (Category E104). The standard
procedure was followed with Ni(cod)2 (395 mg, 1.44 mmol, 10 mol
%), PPh3 (377 mg, 1.44 mmol, 10 mol %), alkyne 16f (2.65 g,
14.4 mmol), and 1,3-butadiene (∼50 mL, 576 mmol, 40 equiv) to
afford crude 24f (after workup B), which was further purified by flash
chromatography (silica gel, hexanes → hexanes/EtOAc 100:1). After
removal of traces 1,5-cyclooctadiene by application of vacuum
overnight analytically pure 24f was obtained as a colorless oil
(980 mg, 3.35 mmol, 23%): Rf = 0.69 (silica gel, DCM/Et2O, 2:1); 1H
NMR (500 MHz, CDCl3, −54 °C) δ = 5.46 (t, J = 10.9 Hz, 1 H),
5.39−5.28 (m, 2 H), 5.24−5.15 (m, 2 H), 3.65 (td, J = 9.5, 5.0 Hz, 1
H), 3.56 (q, J = 8.9 Hz, 1 H), 3.25 (t, J = 12.5 Hz, 1 H), 2.79−2.63 (m,
2 H), 2.29−2.23 (m, 2 H), 2.21−2.14 (m, 1 H), 2.08−1.99 (m, 2 H),
1.94 (q, J = 11.6 Hz, 1 H), 1.54 (q, J = 11.6 Hz, 1H), 0.84 (s, 9 H),
0.02 (s, 6 H) ppm; 13C NMR (125 MHz, CDCl3, −54 °C) δ = 138.7,
131.4, 131.4, 129.4, 123.8, 123.6, 62.7, 41.9, 32.6, 31.4, 31.0, 30.4, 25.9,
18.5, −5.4 ppm; IR (ATR, neat) νmax 3007, 2928, 2894, 2856, 1471,
1462, 1435, 1387, 1360, 1253, 1091, 1005, 970, 932, 914, 833, 811,
772, 742, 706, 664 cm−1; HRMS calcd for C18H32OSi 293.2295
[M + H]+, found (ESI-TOF) 293.2289.
Isolation of a Main Component S7 from the Reaction of
Alkyne 16f with Isoprene (Category C104). Reaction performed as
in standard procedure with altered stoichiometry: Protected alcohol
(16f, 2.11 g, 11.5 mmol), isoprene (5.7 mL, 57.3 mmol, 5.0 equiv),
Ni(cod)2 (106 mg, 0.39 mmol, 3.3 mol %), PPh3 (100 mg, 0.39 mmol,
3.3 mol %).107 Workup A afforded a brown oil, which was judged to be
a mixture of >6 components by TLC and GC analysis. An aliquot
(20 mg) of the crude material was purified by preparative TLC (silica gel,
hexanes, plate run three times; then hexanes/EtOAc, 150:1, plate run
three times) yielding S7 as the only product that could be isolated in
analytically pure form: Rf = 0.21 (silica gel, 2% EtOAc in hexanes); 1H
NMR (600 MHz, CDCl3) δ = 6.04 (s, 1 H, H-5), 5.37 (s, 1 H, H-4),
5.02 (s, 1 H, H-4), 3.79−3.67 (m, 6 H, H-1, H-10, H-12), 2.63 (t, J =
7.0 Hz, 2 H, H-11), 2.57 (t, J = 7.4 Hz, 2 H, H-9), 2.35 (t, J = 6.9 Hz, 2
H, H-2), 0.90−0.88 (m, 36 H), 0.07 (s, 6 H), 0.05 (s, 6 H), 0.04 (s, 6
H) ppm; 13C NMR (150 MHz, CDCl3) δ = 142.5 (C-6), 136.7 (C-5),
118.5 (C-3), 118.0 (C-4), 94.1 (C-8), 81.3 (C-7), 63.0 (C-12), 62.2
(C-1), 62.0 (C-10), 43.3 (C-2), 38.8 (C-11), 26.8 (tBu), 26.7 (tBu),
24.3 (C-9), 18.5, −5.1 ppm; IR (ATR, neat) νmax 2955, 2926, 2855,
1724, 1463, 1380, 1361, 1253, 1096, 1005, 938, 834, 812, 775, 721,
667 cm−1; HRMS calcd for C30H60O3Si3 553.3923 [M + H]+, found
(ESI-TOF) 553.3912.
Characterization of 24f after Deprotection To Form 24c
(Category D2104). Protected alcohol 24f (760 mg, 2.60 mmol) was
dissolved in THF (15 mL, 0.17 M). TBAF in THF (1.0 M, 2.86 mL,
2.86 mmol, 1.1 equiv) was added dropwise over 15 min at room
temperature. The reaction was stirred at room temperature until it was
judged to be complete by TLC (ca. 30 min). The mixture was
neutralized with 1.0 N HCl and extracted with Et2O. The organic layer
was washed with water and brine, dried over MgSO4, and concentrated
in vacuo. The crude material was purified by column chromatography
(silica gel, hexanes → DCM/hexanes 2:1) to give 24c as a colorless oil
1
(448 mg, 2.51 mmol, 97%): Rf = 0.25 (silica gel, DCM); H NMR
(500 MHz, CDCl3, −50 °C) δ = 5.49−5.41 (m, 2 H), 5.35 (ddd, J =
15.3, 10.4, 4.7 Hz, 1 H), 5.27−5.18 (m, 2 H), 3.71−3.54 (m, 2 H),
3.28 (t, J = 12.6 Hz, 1 H), 2.85−2.68 (m, 2 H), 2.33−2.23 (br m, 3 H),
2.18 (br s, 1 H), 2.11−2.02 (m, 2 H), 1.95 (dd, J = 11.9, 11.5 Hz, 1
H), 1.56 (q, J = 11.6 Hz, 1 H) ppm; 13C NMR (125 MHz, CDCl3,
−50 °C) δ = 138.3, 131.2, 131.1, 129.6, 125.3, 124.0, 59.7, 41.4, 32.7,
837
dx.doi.org/10.1021/jo202314a | J. Org. Chem. 2012, 77, 825−842

The Journal of Organic Chemistry
Featured Article
Characterization of 17d after Cope Rearrangement To Form
rac-37 and meso-37′ (Category D2104). The standard procedure
with workup A was followed using alkyne 16d (316 mg, 0.42 mL,
2.87 mmol) and isoprene (∼1.5 mL, 14.4 mmol, 5.0 equiv). The obtained
crude material containing macrocycle 17d was distilled directly to
afford a 1:1 mixture of rac-37 and meso-37′ as determined by analysis
Divergent Derivatization of 24f and 24c. Ring Fusion To
Yield a 6,6-Bicyclic System (rac-32). Macrocycle 24f (50.0 mg,
0.17 mmol) was dissolved in glacial acetic acid (1 mL, 0.17 M), and
H2SO4 (1 drop) was added. The mixture was stirred at room
temperature for 18 h, diluted with H2O, neutralized with saturated
aqueous NaHCO3, and extracted with DCM. The organic layer was
dried over MgSO4 and concentrated in vacuo. Purification of the crude
material by preparative TLC (silica gel, DCM:hexanes, 3:2) afforded
the fused 6,6-bicyclic system rac-32 as a colorless oil (22.0 mg,
78.5 μmol, 46%) and uncyclized macrocycle 24ad as an oil (10.4 mg,
47.2 μmol, 28%).
1
of H, 13C, and APT NMR. In the following only the characteristic
proton NMR signals are listed (full spectral data provided in the
1
Supporting Information): Rf = 0.67 (silica gel, hexanes); H NMR
(500 MHz, CDCl3) δ = 6.02 (dd, J = 17.1, 11.4 Hz, 1 H, H-4 in rac-
37), 4.97 (s, 1 H), 4.94 (dd, J = 8.0, 1.7 Hz, 1 H), 4.80−4.75 (m, 3 H),
4.71 (d, J = 2.4 Hz, 1 H), 4.67 (s, 2 H), 1.74 (s, 6 H, H-17, H-18 in
rac-37′), 1.71 (s, 3H, H-17 in rac-37) ppm; 13C NMR (125 MHz,
CDCl3) δ = 148.3, 147.8, 144.0 (CH, C-4 in rac-37), 129.6, 129.5,
128.6, 112.2, 112.0, 110.2 (CH2, C-2, C-3 in meso-37′), 51.6, 44.5,
42.7, 38.5, 35.3, 35.1, 34.9, 33.6, 26.1, 23.9, 23.0, 21.8, 21.7 (two
signals), 14.5, 14.4 ppm; IR (ATR, neat) νmax 3082, 2957, 2930, 2870,
1639, 1454, 1374, 1006, 910, 888, 738 cm−1; GC/MS calcd for C18H30
rac-32: Rf = 0.10 (DCM/hexanes, 1:1); 1H NMR (500 MHz,
CDCl3) δ = 5.37 (dd, J = 2.0, 1.5 Hz, 1 H, H-8), 4.89 (dt, J = 11.0, 5.0
Hz, 1 H, H-4), 4.14 (t, J = 7.0 Hz, 2 H, H-12), 2.30−2.21 (m, 4 H, H-
5, H-9, H-11), 2.03 (s, 6 H, H-14, H-16), 1.85−1.67 (m, 5 H, H-3, H-
6, H-9, H-10), 1.57 (ddd, J = 24.0, 12.5, 4.5 Hz, H-3), 1.41−1.18 (m, 3
H, H-1, H-2) ppm; 13C NMR (125 MHz, CDCl3) δ = 171.2 (C-13),
170.7 (C-15), 130.8 (C-7), 121.7 (C-8), 75.6 (C-4), 63.1 (C-12), 36.9
(C-11), 36.0 (C-5), 33.2 (C-10), 31.5 (C-9), 25.9 (C-1), 25.6 (C-3),
24.1 (C-6), 23.8 (C-2), 21.5 (C-14), 21.1 (C-16) ppm; IR (ATR,
neat) νmax 2924, 2857, 1736, 1435, 1363, 1240, 1092, 1031, 973, 910
cm−1; HRMS calcd for C16H24O4 281.1747 [M + H]+, found (ESI-
TOF) 281.1742.
246.2348 [M] , found (FID105) 246.3.
•+
24ad: Rf = 0.38 (DCM/hexanes, 1:1); 1H NMR (500 MHz,
CDCl3) δ = 5.52−5.44 (m, 1 H), 5.43−5.33 (m, 2 H), 5.32−5.19 (m,
2 H), 4.13 (t, J = 7.0 Hz, 2 H), 3.30 (br s, 1 H), 2.76 (br s, 2 H), 2.35−
2.10 (m, 4 H), 2.04 (s, 3 H) ppm; 13C NMR (125 MHz, CDCl3) δ =
171.2, 131.4, 131.3, 131.2, 129.8, 124.5, 124.1, 63.5, 38.0, 32.9, 31.0,
26.4, 21.1 ppm; IR (ATR, neat) νmax 3007, 2957, 2925, 2855, 1741,
1436, 1381, 1364, 1236, 1036, 971, 913, 842, 793, 708 cm−1; HRMS
calcd for C14H20O2 221.1536 [M + H]+, found (ESI-TOF) 221.1536.
Characterization of 17d after Exhaustive Epoxidation
Yielding Triepoxide rac-38 (Category D2104). The standard
procedure was followed with alkyne 16d (316 mg, 0.42 mL, 2.87 mmol)
and isoprene (∼1.5 mL, 14.4 mmol, 5.0 equiv). After workup A,
an aliquot of the crude mixture containing 17d (500 mg, crude) was
brought up in DCM (20 mL) and 70% m-CPBA (3.00 g, 12.2 mmol,
∼6.0 equiv) was added. The reaction was allowed to stir overnight at
room temperature. Excess m-CPBA was quenched with saturated
aqueous NaS2O3, and the mixture was extracted with DCM. The
organic layer was washed with half-saturated aqueous NaHCO3, H2O,
brine, and dried over MgSO4. The crude material was purified by
column chromatography (silica gel, hexanes/EtOAc, 9:1) to afford
triepoxide rac-38 (105 mg, 0.36 mmol; fully characterized) along with
a second diastereomer (characterized by X-ray, see the Supporting
Information). Crystals suitable for X-ray diffraction were obtained by
slow evaporation of solution of rac-38 in Et2O layered with hexanes: Rf
= 0.25 (silica gel, 15% EtOAc in hexanes); 1H NMR (400 MHz,
CDCl3) δ = 2.81 (d, J = 8.4 Hz, 1 H), 2.70 (d, J = 14.9 Hz, 1 H), 2.65
(dt, J = 8.0, 2.4 Hz, 1 H), 2.45 (d, J = 15.6 Hz, 1 H), 2.19−2.11 (m, 2
H), 2.02 (ddd, J = 14.1, 10.5, 5.2 Hz, 1 H), 1.72−1.34 (m, 11 H), 1.33
(s, 3 H), 1.26 (s, 3 H), 1.01−0.97 (m, 6 H) ppm; 13C NMR (100
MHz, CDCl3) δ = 69.9, 65.9, 65.4, 61.7, 61.4, 58.6, 44.7, 37.1, 33.9,
32.2, 30.3, 24.2, 22.8, 18.2, 17.9, 17.1, 14.6, 14.5 ppm; IR (ATR, neat)
νmax 2962, 2932, 2872, 1458, 1383, 1289, 1255, 1158, 1135, 1109,
1066, 1045, 1006, 958, 912, 872, 842, 798, 762, 731, 670 cm−1; HRMS
calcd for C18H30O3 295.2268 [M + H]+, found (ESI-TOF) 295.2265.
Preparation of the Crystalline Derivative 33. Et3N (0.23 mL,
1.68 mmol, 6.0 equiv) and DMAP (8.6 mg, 0.07 mmol, 25 mol %)
were added to a solution of 24c (50 mg, 0.28 mmol) in DCM (3
mL, 0.09 M). 4-Bromobenzyl isocyanate (221 mg, 1.12 mmol, 4.0
equiv) was added, and the resulting suspension was stirred for 18 h
at room temperature. The mixture was acidified with 1 N HCl and
extracted with DCM and EtOAc. The organic layers were washed
with 1 N HCl, H2O, and brine, dried over MgSO4, and con-
centrated in vacuo. Purification of the crude product by preparative
TLC (silica gel, DCM/hexanes, 1:1) afforded 33 (61.0 mg, 0.11
mmol, 38%) as a white foam. Single crystals suitable for X-ray
crystallographic analysis were obtained by slow evaporation of a
solution of 33 in Et2O/hexanes (1:1): Rf = 0.41 (silica gel, DCM/
1
hexanes, 1:1); H NMR (500 MHz, CDCl3) δ = 10.87 (br s, 1 H),
7.55 (d, J = 7.0 Hz, 2 H), 7.43 (s, 4 H), 7.07 (d, J = 7.5 Hz, 2 H),
5.37−5.26 (m, 2 H), 5.24−5.17 (m, 2 H), 5.13 (t, 4.5 Hz, 1 H),
4.21 (br s, 2 H), 3.19 (br s, 2 H), 2.68 (br s, 2 H), 2.22 (t, J =
5.0 Hz, 2 H), 2.02 (br s, 4 H) ppm; 13C NMR (125 MHz, CDCl3)
δ = 155.6, 151.3, 136.8, 136.0, 132.4, 132.1, 131.2, 130.6, 125.3,
124.3, 122.6, 121.6, 116.8, 66.3, 37.9, 33.0, 30.5, 26.3 ppm; IR
838
dx.doi.org/10.1021/jo202314a | J. Org. Chem. 2012, 77, 825−842

The Journal of Organic Chemistry
Featured Article
(ATR, neat) νmax 3220, 3104, 3002, 2924, 2855, 1720, 1688, 1598,
1541, 1487, 1439, 1397, 1378, 1312, 1300, 1270, 1238, 1183, 1133,
1115, 1103, 1072, 1016, 1004, 993, 969, 925, 904, 833, 774, 743,
705, 688 cm−1; HRMS calcd for C26H26Br2N2O3 573.0383
[M + H]+, found (ESI-TOF) 573.0395.
doxime (115 mg, 0.56 mmol, 1.0 equiv) and NaHCO3 (94.0
mg,1.12 mmol, 2.0 equiv) were added in intervals of 1 h until TLC
analysis indicated full consumption of starting material.108 The
reaction was quenched with H2O and extracted with EtOAc. The
organic layer was washed with H2O and brine, dried over MgSO4,
filtered, and concentrated in vacuo. Column chromatography (silica
gel, DCM/Et2O, 100:1→10:1) afforded the desired monocycload-
duct rac-45 (35 mg, 21%) and its undesired regioisomer rac-46 (98
mg, 58%).
1
rac-45: Rf = 0.48 (silica gel, DCM/Et2O, 10:1); H NMR (600
Hydrogenation of 24c To Afford 34 and rac-35. [Ir(cod)-
(PCy3)(py)]PF6 (4.50 mg, 5.59 μmol, 10 mol %) was added to a solution
of macrocycle 24c (10.0 mg, 56.1 μmol) in DCM. The reaction vessels was
placed in a hydrogenation apparatus, charged with H2 (150 psi), and stirred
at room temperature for 18 h. The crude mixture was filtered through a
silica plug (eluting with DCM) and the organic phase evaporated to
dryness. Preparative TLC (silica gel, DCM) afforded analytically pure 34
(6.1 mg, 33.8 μmol, 60%) and rac-35 (3.1 mg, 17.0 μmol, 30%).
MHz, CDCl3) δ = 5.68 (td, J = 10.7, 6.4 Hz, 1 H, H-5), 5.44 (td, J =
11.0, 5.5 Hz, 1 H, H-4), 5.11 (dd, J = 11.4, 5.7 Hz, 1 H-8), 4.30 (ddd,
J = 13.9, 11.5, 2.4 Hz, 1 H, H-1), 3.78−3.70 (m, 2 H, H-12), 3.29 (t,
J = 11.9 Hz, 1 H-6), 3.24 (ddd, J = 13.8, 5.4, 2.4 Hz, 1 H, H-10),
2.87 (ddd, J = 14.9, 11.4, 5.6 Hz, 1 H, H-9), 2.59 (ddd, J = 14.9, 5.8,
2.6 Hz, 1 H, H-9), 2.53−2.44 (m, 1 H, H-3), 2.39−2.27 (m, 3 H, H-
6, H-11), 2.14 (ddd, J = 13.3, 8.5, 4.1 Hz, 1 H, H-3), 2.08 (tt, J =
13.8, 2.4 Hz, 1 H, H-2), 1.88 (dddd, J = 13.8, 10.8, 4.2, 3 Hz, 1 H,
H-2) ppm; 13C (150 MHz, CDCl3) δ = 145.5 (C-13), 140.5 (C-7),
129.6 (C-4), 126.8 (C-5), 120.2 (C-8), 83.3 (C-1), 60.5 (C-12), 56.7
(C-10), 40.0 (C-11), 32.1 (C-2), 27.7 (C-6), 26.7 (C-9), 22.8 (C-3); IR
(ATR, neat) νmax 3410, 3014, 2957, 2922, 2853, 1715, 1644, 1571, 1466,
1377, 1289, 1180, 1118, 1107, 1084, 1045, 979, 916, 887, 802, 760, 720
cm−1; HRMS calcd for C13H18BrNO2 300.0594 [M + H]+, found (ESI-
TOF) 300.0602.
1
34: Rf = 0.43 (silica gel, DCM); H NMR (600 MHz, CDCl3) δ =
5.74 (app q, J = 9.3 Hz, 1 H, H-5), 5.29 (app dd, J = 9.4, 8.9 Hz, 1 H, H-
4), 5.08 (t, J = 8.4 Hz, 1 H, H-8), 3.72 (d, J = 5.7 Hz, 2 H, H-12), 2.78 (d,
J = 8.4 Hz, 2 H, H-6), 2.38−2.30 (m, 6 H, H-3, H-9, H-11), 1.56−1.50
(m, 2 H), 1.27−1.22 (m, 4 H) ppm; 13C NMR (150 MHz, CDCl3) δ =
136.1 (C-7), 130.7 (C-4), 127.7 (C-8), 127.5 (C-5), 60.4 (C-12),
40.8 (C-11), 30.4, 30.1, 28.3 (C-9), 27.9 (C-6), 27.4 (C-3), 20.6
ppm; IR (ATR, neat) νmax 3335, 3007, 2958, 2921, 2852, 1465,
1443, 1377, 1044, 1025, 935, 877, 864, 796, 761, 710 cm−1; HRMS
calcd for C12H20O 181.1587 [M + H]+, found (ESI-TOF) 181.1592.
1
rac-46: Rf = 0.38 (silica gel, DCM/Et2O, 10:1); H NMR (600
MHz, CDCl3) δ = 5.78 (td, J = 10.8, 6.3 Hz, 1 H, H-5), 5.38−5.29 (m,
2 H, H-4, H-8), 4.43 (dt, J = 13.5, 2.8 Hz, 1 H, H-10), 3.75 (dd, J =
8.1, 3.6 Hz, 2 H, H-12), 3.26 (t, J = 11.9 Hz, 1 H, H-6), 3.08 (t, J =
12.0 Hz, 1 H, H-1), 2.86 (ddd, J = 15.0, 11.1, 3.6 Hz, 1 H, H-9), 2.70
(ddd, J = 14.9, 5.2, 2.0 Hz, 1 H, H-9), 2.43−2.31 (m, 4 H, H-3, H-6,
H-11), 2.15−2.09 (m, 1 H, H-3), 1.95 (t, J = 13.9 Hz, 1 H, H-2), 1.56
(m, 1 H, H-2) ppm; 13C NMR (150 MHz, CDCl3) δ = 146.9 (C-13),
140.6 (C-7), 129.0 (C-4), 128.4 (C-5), 119.8 (C-8), 89.3 (C-10), 60.6
(C-12), 49.9 (C-1), 40.0 (C-11), 30.1 (C-2), 29.9 (C-9), 27.9 (C-6),
23.8 (C-3) ppm; IR (ATR, neat) νmax: 3390, 3010, 2922, 2862, 1719,
1657, 1572, 1466, 1379, 1305, 1251, 1128, 1072, 1043, 1007, 927, 879,
798, 772, 720 cm−1; HRMS calcd for C13H18BrNO2 300.0594
[M + H]+, found (ESI-TOF) 300.0597.
1
rac-35: Rf = 0.43 (silica gel, DCM); H NMR (600 MHz, CDCl3)
δ = 5.47−5.38 (m, 2 H), 3.75−3.68 (m, 2 H), 2.35 (br s, 1 H), 2.14
(br s, 2 H), 1.76 (br d, J = 6.6 Hz, 1 H), 1.62−1.37 (m, 12 H), 1.24
(m, 2 H); 13C NMR (150 MHz, CDCl3) δ = 131.2, 128.1, 61.6, 37.4,
32.9, 30.3, 29.9, 27.9, 27.5, 26.1, 25.7, 20.9 ppm; IR (ATR, neat) νmax
3333, 2957, 2922, 2853, 1463, 1377, 1048, 888, 789, 722 cm−1; HRMS
calcd for C12H22O 183.1743 [M + H]+, found (ESI-TOF) 183.1752.
Cope Rearrangement to Divinylcyclohexene rac-36. Macro-
cycle 24f (20.0 mg, 68.4 μmol) was heated neat to 100 °C in a sealed
tube for 3 h. After purification of the crude mixture by column
chromatography (silica gel, hexanes/EtOAc, 99:1) rac-36 was obtained
1
as a clear, pale yellow oil (16.0 mg, 54.7 μmol, 80%): H NMR (500
MHz, CDCl3) δ = 5.84−5.74 (m, 1 H), 5.41 (s, 1 H), 5.03−4.98 (m, 3
H), 3.66 (t, J = 7.0 Hz, 2 H), 2.48−2.36 (m, 2 H), 2.23−2.10 (m, 4
H), 2.01−1.89 (m, 2 H), 0.89 (s, 9 H), 0.05 (s, 6 H) ppm; 13C NMR
(125 MHz, CDCl3) δ = 140.3 (two signals), 133.5, 121.2, 114.6, 114.5,
62.4, 41.6, 41.2, 41.1, 32.8, 29.5, 26.1, 18.5, −5.1 ppm; IR (ATR, neat)
νmax: 3076, 2955, 2928, 2889, 2857, 1639, 1472, 1463, 1436, 1387,
1361, 1254, 1094, 995, 911, 833, 812, 774, 717, 661 cm−1; HRMS
calcd for C18H32SiO 293.2295 [M + H]+, found (ESI-TOF) 293.2287.
Dicycloadducts rac-39 and rac-40. Monocycloadduct rac-45
(30.1 mg, 0.10 mmol) was dissolved in EtOAc (1 mL, 0.1 M).
NaHCO3 (33.6 mg, 0.40 mmol, 4.0 equiv) was added to the reaction
followed by dibromoformaldoxime (41.6 mg, 0.20 mmol, 2.0 equiv).
Repeated additions of base (16.8 mg, 0.20 mmol, 2.0 equiv) and
dibromoformaldoxime (20.8 mg, 0.10 mmol, 1.0 equiv) were
continued over 22 h until the reaction was judged to be complete
by TLC.108 The mixture was diluted with H2O and extracted with
EtOAc. The organic was washed with brine, dried over MgSO4, and
concentrated in vacuo. Preparative TLC (silica gel, DCM/Et2O, 9:1)
Delayed Installation of C1 Units onto 24c. Monocycload-
ducts rac-45 and rac-46. Macrocycle 24c (100 mg, 0.56 mmol)
was dissolved in EtOAc (5.6 mL, 0.1 M). NaHCO3 (282 mg, 3.36
mmol, 6.0 equiv) followed by dibromoformaldoxime (115 mg, 0.56
mmol, 1.0 equiv) were added at room temperature and the reaction
was stirred under ambient atmosphere. Additional dibromoformal-
1
afforded a 1:1 mixture (by H NMR) of the two cycloadducts rac-39
and rac-40 (20.0 mg, 47.4 μmol, 47%). For full structural elucidation
the crude mixture was further purified by preparative TLC (silica gel,
DCM:Et2O, 9:1): The seemingly single PTLC spot was arbitrarily split
839
dx.doi.org/10.1021/jo202314a | J. Org. Chem. 2012, 77, 825−842

The Journal of Organic Chemistry
Featured Article
in half affording enriched samples of the two regioisomers rac-39 and
rac-40 which allowed for full assignment by 2D NMR.
NMR (150 MHz, CDCl3) δ = 145.8 (C-13), 145.4 (C-14), 135.9 (C-
7), 123.7 (C-8), 88.3 (C-10), 82.7 (C-5), 61.1 (C-12), 53.2 (C-4),
49.6 (C-1), 38.9 (C-11), 30.5 (C-2), 29.6 (C-9), 28.4 (C-6), 21.6 (C-
3) ppm; IR (ATR, neat) νmax 3404, 2922, 2855, 1728, 1570, 1464,
1447, 1378, 1325, 1303, 1263, 1210, 1167, 1126, 1084, 1043, 1010,
964, 907, 884, 855, 829, 801, 729 cm−1; HRMS calcd for
C14H18Br2N2O3 420.9757 [M + H]+, found (ESI-TOF) 420.9771.
rac-42: Rf = 0.40 (silica gel, DCM/Et2O, 9:1); 1H NMR (600 MHz,
CDCl3) δ = 5.45 (dd, J = 11.1, 5.7 Hz, 1 H, H-8), 4.50−4.42 (m, 2 H,
H-4, H-10), 3.87−3.75 (m, 2 H, H-12), 3.07 (t, J = 7.3 Hz, 1 H, H-5),
2.97 (t, J = 13.0 Hz, 1 H, H-1), 2.85−2.70 (m, 3 H, H-6, H-9), 2.50−
2.44 (m, 1 H, H-11), 2.41 (q, J = 7.7 Hz, 1 H, H-11), 2.27−2.21 (m, 1
H, H-3), 2.14−2.06 (m, 2 H, H-2, H-3), 1.86 (d, J = 15.3 Hz, 1 H, H-
6), 1.72 (m, 1 H, H-2) ppm; 13C NMR (150 MHz, CDCl3) δ = 145.7
(C-14), 145.2 (C-13), 138.9 (C-7), 122.2 (C-8), 88.3 (C-10), 87.1 (C-
4), 60.6 (C-12), 52.7 (C-5), 50.5 (C-1), 38.9 (C-11), 29.6 (C-9), 28.3
(C-2), 25.2 (C-6), 23.6 (C-3) ppm; IR (ATR, neat) νmax 3421, 2922,
2854, 1730, 1571, 1464, 1378, 1301, 1254, 1127, 1088, 1047, 966, 948,
886, 853, 797, 728 cm−1; HRMS calcd for C14H18Br2N2O3 420.9757
[M + H]+, found (ESI-TOF) 420.9774.
rac-39: Rf = 0.41 (silica gel, DCM/Et2O, 9:1); 1H NMR (600 MHz,
CDCl3) δ = 5.21 (t, J = 8.5 Hz, 1 H, H-8), 4.51 (t, J = 9.1 Hz, 1 H, H-
4), 4.19 (t, J = 12.8 Hz, 1 H, H-1), 3.84−3.74 (m, 2 H, H-12), 3.26 (d,
J = 13.3 Hz, 1 H, H-10), 3.01 (t, J = 6.9 Hz, 1 H, H-5), 2.78 (dd, J =
15.3, 6.0 Hz, 1 H, H-6), 2.70 (m, 2 H, H-9), 2.45 (dt, J = 13.1, 4.0 Hz,
1 H, H-11), 2.37 (dt, J = 14.9, 7.4 Hz, 1 H, H-11), 2.30 (dd, J = 14.0,
5.3 Hz, 1 H, H-3), 2.14 (m, 2 H, H-2, H-3), 2.05 (m, 1 H, H-2), 1.84
(d, J = 15.6 Hz, 1 H, H-6) ppm; 13C NMR (150 MHz, CDCl3) δ =
145.6 (C-14), 145.0 (C-13), 139.6 (C-7), 122.1 (C-8), 87.3 (C-4),
82.1 (C-1), 60.4 (C-12), 56.1 (C-10), 52.7 (C-5), 39.0 (C-11), 30.7
(C-2), 26.5 (C-9), 24.9 (C-6), 22.6 (C-3) ppm; IR (ATR, neat) νmax
3415, 2926, 2866, 1717, 1571, 1467, 1448, 1392, 1306, 1288, 1250,
1214, 1177, 1107, 1085, 1043, 1002, 952, 908, 872, 854, 800,
731 cm−1; HRMS calcd for C14H18Br2N2O3 420.9757 [M + H]+, found
(ESI-TOF) 420.9752.
rac-40: Rf = 0.37 (silica gel, DCM/Et2O, 9:1); 1H NMR (600 MHz,
CDCl3) δ = 5.26 (dd, J = 10.9, 6.0 Hz, 1 H, H-8), 4.70 (t, J = 9.8 Hz, 1
H, H-5), 4.23 (t, J = 12.7 Hz, 1 H, H-1), 3.83−3.72 (m, 2 H, H-12),
3.24 (d, J = 13.8 Hz, 1 H, H-10), 2.88 (t, J = 12.7 Hz, 1 H, H-6), 2.83
(m, 1 H, H-4), 2.74−2.68 (m, 2 H, H-9), 2.52 (d, J = 14.2 Hz, 1 H, H-
6), 2.41−2.31 (m, 2 H-11), 2.12 (t, J = 14.5 Hz, 1 H), 1.98 (t, J = 15.1
Hz, 1 H), 1.66 (dd, J = 16.2, 3.6 Hz, 1 H) ppm; 13C NMR (150 MHz,
CDCl3) δ = 145.7, 145.6, 136.5 (C-7), 123.7 (C-8), 82.5 (C-5), 81.8
(C-1), 61.0 (C-12), 55.8 (C-10), 53.5 (C-4), 39.0 (C-11), 32.7, 28.3
(C-6), 26.5 (C-9), 20.7 ppm; IR (ATR, neat) νmax 3432, 2923, 2855,
1718, 1570, 1467, 1448, 1379, 1287, 1270, 1253, 1214, 1176, 1108,
1085, 1044, 976, 952, 909, 877, 853, 801, 733 cm−1; HRMS calcd for
C14H18Br2N2O3 420.9757 [M + H]+, found (ESI-TOF) 420.9738.
ASSOCIATED CONTENT
* Supporting Information
■
S
Copies of relevant 1D and 2D NMR spectra, gas chromato-
grams, and details of the crystal structure determinations. This
material is available free of charge via the Internet at http://
pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
*E-mail: pbaran@scripps.edu.
■
Author Contributions
†These authors contributed equally to this work.
ACKNOWLEDGMENTS
■
We thank Dr. D.-H. Huang and Dr. L. Pasternack for assistance
with NMR spectroscopy and Prof. V. Fokin for providing
access to GC/MS and glovebox. We also acknowledge Prof.
A. Rheingold and Dr. C. Moore for X-ray crystallographic
assistance. We thank N. Ihme and J. Lorenz (University of
Bochum) for providing the unpublished work of the Heimbach
group in the form of 13 Ph.D. dissertations. We are grateful to
the German Academic Exchange Service, DAAD (postdoctoral
fellowship to D.C.G.G.), the NSF (predoctoral fellowship to
D.H.), and Amgen (predoctoral fellowship to D.H.). Financial
support for this work was provided by Leo Pharmaceuticals and
the NIH/NIGMS (GM-097444).
Dicycloadducts rac-41 and rac-42. Monocycloadduct rac-46
(91.0 mg, 0.30 mmol) was dissolved in EtOAc (3 mL, 0.1 M).
NaHCO3 (101 mg, 1.20 mmol, 4.0 equiv) was added to the reaction
followed by dibromoformaldoxime (122 mg, 0.60 mmol, 2.0 equiv).
Repeated additions of base (51.0 mg, 0.60 mmol, 2.0 equiv) and
dibromoformaldoxime (61.0 mg, 0.30 mmol, 1.0 equiv) were
continued over 22 h until the reaction was judged to be complete
by TLC.108 The mixture was diluted with H2O and extracted with
EtOAc. The organic was washed with brine, dried over MgSO4, and
concentrated in vacuo. Purification of the crude mixture by column
chromatography (silica gel, DCM:Et2O, 50:1 → 2:1) afforded a 1:1
REFERENCES
■
(1) Maimone, T. J.; Baran, P. S. Nat. Chem. Biol. 2007, 3, 396−407.
(2) Breitmaier, E. Terpenes: Flavors, Fragrances, Pharmaca, Pher-
omones; Wiley-VCH: Weinheim, 2006; pp 1−9, 24−50.
(3) Sacchettini, J. C.; Poulter, C. D. Science 1997, 277, 1788−1789.
(4) Hosler, D.; Burkett, S. L.; Tarkanyan, M. J. Science 1999, 284,
1988−1991.
1
mixture (by H NMR) of the two cycloadducts rac-41 and rac-42
(52.0 mg, 0.12 mmol, 41%). For full structural elucidation an aliquot
of the crude mixture was further purified by preparative TLC (silica
gel, DCM:Et2O, 9:1): The seemingly single PTLC spot was arbitrarily
split in half affording enriched samples of the two regioisomers rac-41
and rac-42 which allowed for full assignment by 2D NMR.
(5) Weitz, H. M.; Loser, E. Isoprene. In Ullmann’s Encyclopedia of
Industrial Chemistry; Wiley-VCH: Weinheim, 2002; pp 1−20.
(6) For a review that focuses on the syntheses of germacrane
sesquiterpenes, see: Minnaard, A. J.; Wijnberg, J. B. P. A.; de Groot, A.
Tetrahedron 1999, 55, 2115−2146.
rac-41: Rf = 0.36 (silica gel, DCM/Et2O, 9:1); 1H NMR (600 MHz,
CDCl3) δ = 5.53 (dd, J = 11.0, 5.5 Hz, 1 H, H-8), 4.74 (t, J = 10.0 Hz,
1 H, H-5), 4.45 (d, J = 13.4 Hz, 1 H, H-10), 3.84−3.74 (m, 2 H, H-
12), 3.06 (t, J = 12.8 Hz, 1 H, H-1), 2.91−2.68 (m, 4 H, H-4, H-6, H-
9), 2.54 (d, J = 14.2 Hz, 1 H, H-6), 2.43−2.33 (m, 2 H, H-11), 2.03−
1.83 (m, 2 H, H-2, H-3), 1.70−1.57 (m, 2 H, H-2, H-3) ppm; 13C
(7) Kellogg, B. A.; Poulter, C. D. Curr. Opin. Chem. Biol. 1997, 1,
570−578.
(8) Ishihara, Y.; Baran, P. S. Synlett 2010, 12, 1733−1745.
(9) Chen, K.; Baran, P. S. Nature 2009, 459, 824−828.
(10) Chen, K.; Ishihara, Y.; Moron Galan, M.; Baran, P. S.
́
́
Tetrahedron 2010, 66, 4738−4744.
840
dx.doi.org/10.1021/jo202314a | J. Org. Chem. 2012, 77, 825−842

The Journal of Organic Chemistry
Featured Article
(11) Spurgeon, S. L.; Porter, J. W. Conversion of acetyl coenzyme A
to isopentenyl pyrophosphate. In Biosynthesis of Isoprenoid Compounds;
John Wiley & Sons: New York, 1981; Vol. 1, pp 1−93.
(12) Heimbach, P. Angew. Chem., Int. Ed. 1973, 12, 975−989.
(13) Hu, T.; Corey, E. J. Org. Lett. 2002, 4, 2441−2443 and
references cited therein.
(49) Buchholz, H. A. Ph.D. Thesis, University of Bochum, 1971.
(50) Bussemeier, B. Ph.D. Thesis, University of Bochum, 1973.
̈
(51) Thomel, F. Ph.D. Thesis, University of Bochum, 1970.
̈
(52) The majority of these studies are only published in detail in the
form of several Ph.D. dissertations from the Heimbach group. They
were graciously made available by The University of Bochum. These
documents can be obtained by emailing the corresponding author
(P.S.B.). Furthermore, many of the published studies are not easily
available and are in German; however, the primary literature is
comprehensively referenced in this publication.
(14) For selected examples, see refs 15−31.
(15) Wilke, G.; Bogdanovic, B.; Hardt, P.; Heimbach, P.; Keim, W.;
Kroner, M.; Oberkirch, W.; Tanaka, K.; Steinrucke, E.; Walter, D.;
̈
̈
Zimmermann, H. Angew. Chem., Int. Ed. 1966, 5, 151−164.
(16) Neilan, J. P.; Laine, R. M.; Cortese, N.; Heck, R. F. J. Org. Chem.
1976, 41, 3455−3460.
(53) Heimbach, P.; Wilke, G. Liebigs Ann. Chem 1969, 727, 183−193.
(54) For examples of intramolecular Ni0-catalyzed [4 + 4]
cycloaddition in total synthesis, see refs 55 and 56.
(17) Keim, W.; Roper, M.; Schieren, M. J. Mol. Catal. 1983, 20, 139−
̈
(55) Wender, P. A.; Snapper, M. L. Tetrahedron Lett. 1987, 28,
151.
(18) Roper, M.; He, R.; Schieren, M. J. Mol. Catal. 1985, 31, 335−
2221−2224.
̈
(56) Wender, P. A.; Ihle, N. C.; Correia, C. R. D. J. Am. Chem. Soc.
1988, 110, 5904−5906.
343.
(19) Jolly, P. W. Angew. Chem., Int. Ed. 1985, 24, 283−295.
(20) Goliaszewski, A.; Schwartz, J. Tetrahedron Lett. 1985, 5779−
5789.
(21) Hoberg, H.; Minato, M. J. Organomet. Chem. 1991, 406, C25−
C28.
(22) Maddock, S. M.; Finn, M. G. Organometallics 2000, 19, 2684−
2689.
(23) Misono, A.; Uchida, Y.; Hidai, M.; Ohsawa, Y. Bull. Chem. Soc.
(57) For an example of Ni0-catalyzed isoprene oligomerization,
affording complex mixtures of terpenes, see: Bogatian, M. V.; Bogatian,
G.; Udrea, S.; Corbu, A. C.; Chiraleu, F.; Maganu, M.; Deleanu, C. Rev.
Roum. Chem. 2006, 51, 735−741.
(58) Baker, R.; Cook, A. H.; Crimmin, M. J. J. Chem. Soc., Chem.
Commun. 1975, 727−728.
(59) Bogdanovic, B.; Kroner, M.; Wilke, G. Liebigs Ann. Chem. 1966,
̈
699, 1−23.
Jpn. 1966, 39, 2425−2429.
(60) Oenbrink, G.; Schiffer, T. Cyclododecatriene, Cyclooctadiene,
and 4-Vinylcyclohexene. In Ullmann’s Encyclopedia of Industrial
Chemistry; Wiley-VCH: Weinheim, 2009; pp 1−3.
(61) A similar, although mechanistically distinct, phenomenon
regarding the oligomerization of ethylene is called the “nickel effect:”
Fischer, K.; Jonas, K.; Misbach, P.; Stabba, R.; Wilke, G. Angew. Chem.,
Int. Ed. 1973, 12, 943−953.
(24) Buchkremer, J. Ph.D. Thesis, University of Bochum, 1973.
(25) tom Dieck, H.; Dietrich, J. Angew. Chem., Int. Ed. 1985, 24,
781−783.
(26) Kopp, W. Ph.D. Thesis, University of Bochum, 1970.
(27) Wilke, G.; Kroner, M. Angew. Chem. 1959, 71, 574.
̈
(28) Selbeck, H. Ph.D. Thesis, University of Bochum, 1972.
(29) Wilke, G.; Briel, H. Makromol. Chem. 1963, 69, 18−40.
(30) Yasuda, H.; Nakamura, A. Angew. Chem., Int. Ed. 1987, 26, 723−
(62) Ozawa, F. Group 10 (Ni, Pd, Pt) Metal Compounds. In
Synthesis of Organometallic Compounds; Komiya, S., Ed.; John Wiley &
Sons Ltd.: New York, 1997; pp 249−303.
(63) For an early, extensive study on the reduction of Ni(acac)2 with
organoaluminum reagents, see: Misbach, P. Ph.D. Thesis, University of
Bochum, 1969.
(64) For a discussion of the catalytic cycle as originally proposed by
Wilke and co-workers, see ref 42.
(65) For detailed investigations on the complexity of conformational
isomerism in π-allyl nickel systems, see: Schenkluhn, H. Ph.D. Thesis,
University of Bochum, 1971.
742.
(31) Morikawa, H; Kitazume, S. Ind. Eng. Chem. Prod. Res. Dev. 1979,
18, 254−258.
(32) Itoh, K.; Masuda, K.; Fukahori, T.; Nakano, K.; Aoki, K.;
Nagashima, H. Organometallics 1994, 13, 1020−1029.
(33) Wilke, G. Angew. Chem., Int. Ed. 1988, 27, 185−206.
(34) Heimbach, P. J. Synthetic Org. Chem 1973, 31, 300−312.
(35) Wilke, G.; Eckerle, A. Cyclooligomerizations and Cyclo-co-
oligomerizations of 1,3-Dienes. In Applied Homogeneous Catalysis with
Organometallic Compounds: A Comprehensive Handbook in Two
Volumes; Cornils, B., Herrmann, W., Eds.; Wiley-VCH: Weinheim,
1996; Vol. 1, pp 368−382.
(66) van Leeuwen, P. W. N. M.; Roobeek, C. F. Tetrahedron 1980,
37, 1973−1983.
(67) For crystallographic and IR studies on related butadiene-derived
systems, see refs 68 and 69.
(36) Heimbach, P.; Hey, H.-J. Angew. Chem., Int. Ed. 1970, 9, 528−
529.
(37) Hey, H.-J. Ph.D. Thesis, University of Bochum, 1969.
(38) Wiese, W. Ph.D. Thesis, University of Bochum, 1969.
(39) Kiji, J.; Masuy, K.; Furukawa, J. Bull. Chem. Soc. Jpn. 1971, 44,
1956−1961.
(68) Taube, R.; Gehrke, J.-P.; Bohme, P.; Kottnitz, J. J. Organomet.
̈
̈
Chem. 1990, 395, 341−358.
(69) Brown, J. M.; Golding, B. T.; Smith, M. J. J. Chem. Soc. D, Chem.
Commun. 1971, 1240−1241.
(40) Barnett, B.; Bussemeier, B; Heimbach, P.; Jolly, P. W.; Kruger,
(70) For theoretical workon hapticity in butadiene dimerization, see:
Tobisch, S.; Ziegler, T. J. Am. Chem. Soc. 2002, 124, 13290−13301.
(71) In order to simplify the comparison of our 4Z,7Z,10E-
macrocycles to germacrene natural products, they are consistently
numbered in the aforementioned fashion, resulting in numbering and
E/Z descriptors which differ from the preferred IUPAC nomenclature.
(72) There are examples of (4E,10Z), (4Z,10E), (4Z,10Z)
germacrene natural products; however, they are in the minority. For
several examples, see: Adio, A. M. Tetrahedron 2009, 65, 1533−1552.
(73) A different metal may also change the outcome, for example it is
known that titanium catalysts produce primarily 1,5,9-cis,trans,trans-
cyclododecatriene, whereas nickel and chromium afford the respective
all-trans isomer: Weber, H.; Ring, W.; Hochmuth, U.; Franke, W.
Liebigs Ann. Chem. 1965, 681, 10−20.
̈
̈
C.; Tkatchenko, I.; Wilke, G. Tetrahedron Lett. 1972, 15, 1457−1460.
(41) For Ni0-catalyzed construction of 5-membered rings from 1,5-
cyclooctadiene, see: Nussel, H.-G. Ph.D. Thesis, University of
̈
Bochum, 1970.
(42) Benn, R.; Bussemeier, B.; Holle, S.; Jolly, P. W.; Mynott, R.;
̈
Tkatchenko, I.; Wilke, G. J. Organomet. Chem. 1985, 279, 63−86.
(43) Heimbach, P.; Wilke, G. Liebigs Ann. Chem 1969, 727, 183−193.
(44) Brenner, W.; Heimbach, P.; Wilke, G. Liebigs Ann. Chem. 1969,
727, 194−207.
(45) Brenner, W.; Heimbach, P.; Ploner, K.-J.; Thomel, F. Liebigs
̈
Ann. Chem. 1973, 1882−1892.
(46) Bogdanovic, B.; Heimbach, P.; Kroner, M.; Wilke, G.; Brandt, J.
̈
Liebigs Ann. Chem. 1969, 727, 143−160.
(47) Miyake, A.; Kondo, H.; Nishino, M.; Tokizane, S. Special
Lectures, XXIII International Congress of Pure and Applied Chemistry;
Butterworth: London, 1971; v 6, p 201.
(74) Schreiber, S. L.; Santini, C. J. Am. Chem. Soc. 1984, 106, 4038−
4039.
(75) Corey, E. J.; Hamanaka, E. J. Am. Chem. Soc. 1967, 89, 2758−
2759.
(48) Meyer, R.-V. Ph.D. Thesis, University of Bochum, 1973.
841
dx.doi.org/10.1021/jo202314a | J. Org. Chem. 2012, 77, 825−842

The Journal of Organic Chemistry
Featured Article
(76) Fleck, W. Ph.D. Thesis, University of Bochum, 1971.
(77) Brille, F.; Heimbach, P.; Kluth, J.; Schenklunh, H. Angew. Chem.,
Int. Ed. 1979, 18, 400−401.
(106) In addition, severalpartially inseparablefurther compo-
nents were obtained that could not be characterized due to a lack of
material.
(107) Alternatively, the reaction was run with a dropwise addition of
16f or with a 40:1 diene/alkyne ratio. Although there was a noticeable
increase in macrocycle formation by GC/MS, the mixture was still too
complex for isolation.
(78) For a comprehensive overview, see the review included in ref 72.
(79) For NMR studies, see: Faraldos, J. A.; Wu, S.; Chappell, J.;
Coates, R. M. Tetrahedron 2007, 63, 7733−7742.
(80) For implications in synthesis, see, for example: Kodama, M.;
Yokoo, S.; Matsuki, Y.; Ito, S. Tetrahedron Lett. 1979, 20, 1687−1690.
(81) Tashkhodzhaev, B; Abduazimov, B. K. Chem. Nat. Compd. 1997,
33, 382−388.
(82) Heimbach, P. Angew. Chem., Int. Ed. 1966, 5, 961.
(83) Heimbach, P.; Brenner, W. Angew. Chem., Int. Ed. 1966, 5, 961−
962.
(108) The dipole, generated in situ, dimerizes and must be
replenished for the reaction to go to completion.
(84) There are 64 theoretical isomers, half of which are degenerate. If
one assumes exclusive head-to-tail coupling of isoprene, this number
reduces to 16 possible isomers. On the other hand, if the 4Z,7Z,10E-
olefin geometry is obtained selectively, there are only four possible
isomers.
(85) No yield was given for the Cope rearrangements.
(86) The assignment of the cis-relationship of the substituents at the
5 and 10 positions is based on previous results from extensive studies
in the Heimbach laboratory. See ref 51 and references cited therein.
(87) Obtained from Ingenieurburo SFS Zurich, Switzerland by
̈
̈
G. Wilke.
(88) Electron-poor alkynes have been incorporated into 10-
membered rings in a related stoichiometric experiment at low
temperature: Bussemeier, B.; Jolly, P. W.; Wilke, G. J. Am. Chem.
̈
Soc. 1974, 94, 4726−4727.
(89) Meriwether, L. S.; Colthup, E. C.; Kennerly, G. W.; Reusch, R. N.
J. Org. Chem. 1961, 26, 5155−5163.
(90) Mori, T.; Nakamura, T.; Kimura, M. Org. Lett. 2011, 13, 2266−
2269.
(91) Baker, R.; Crimmin, M. J. J. Chem. Soc., Perkin Trans. 1 1979,
1264−1267.
(92) Despite complete characterization, unambiguous assignment of
the major component from the 1.6:1 mixture was not possible. On the
basis of the X-ray analysis obtained for 33, the likely major/minor
structures are proposed in the Experimental Section.
(93) The fact that the 5,10-dimethyl regioisomer was not observed is
in full agreement with previous results from the Heimbach group; see
ref 51.
(94) The third triepoxide product was not fully characterized.
(95) See, e.g. Maimone, T. J.; Shi, J.; Ashida, S.; Baran, P. S. J. Am.
Chem. Soc. 2009, 131, 17066−17067.
(96) Curran, D. P. J. Am. Chem. Soc. 1983, 105, 5826−5833.
(97) Cheng, G.; Wang, X.; Chu, R.; Shao, C.; Xu, J.; Hu, Y. J. Org.
Chem. 2011, 76, 2694−2700.
(98) Kociolek, M. G.; Kalbarczyk, P. K. Synth. Commun. 2004, 34,
4387−4394.
(99) For benzonitrile cross-couplings, see: Miller, J. A.; Dankwardt, J.
W. Tetrahedron Lett. 2002, 44, 1907−1910. Preliminary studies in our
group have shown that vinyl-nitriles can also be cross-coupled under
Dankwardt’s conditions.
(100) The dicycloadduct can be accessed in one step by utilizing
excess dipole, but this results in the inability to separate the
regioisomers.
(101) Analytical samples were separated by preparative TLC (see the
Experimental Section).
(102) Rohloff, J. C.; Robinson, J. III; Gardner, J. O. Tetrahedron Lett.
1992, 33, 3113−3116.
(103) When screening other conditions (e.g., ligands other than
PPh3, additives, cosolvents, ...), any solids were added in the glovebox
and liquids were added through the septum under argon prior to the
addition of alkyne.
(104) For assignment of the categories of reaction outcome refer to
Figure 3.
(105) The compound did not ionize under ESI-TOF conditions.
842
dx.doi.org/10.1021/jo202314a | J. Org. Chem. 2012, 77, 825−842