Development of a concise synthesis of (+)-ingenol

Article abstract of DOI:10.1021/ja501881p

The complex diterpenoid (+)-ingenol possesses a uniquely challenging scaffold and constitutes the core of a recently approved anti-cancer drug. This full account details the development of a short synthesis of 1 that takes place in two separate phases (cyclase and oxidase) as loosely modeled after terpene biosynthesis. Initial model studies establishing the viability of a Pauson-Khand approach to building up the carbon framework are recounted. Extensive studies that led to the development of a 7-step cyclase phase to transform (+)-3-carene into a suitable tigliane-type core are also presented. A variety of competitive pinacol rearrangements and cyclization reactions were overcome to develop a 7-step oxidase phase producing (+)-ingenol. The pivotal pinacol rearrangement is further examined through DFT calculations, and implications for the biosynthesis of (+)-ingenol are discussed.

Full text of DOI:10.1021/ja501881p

Article
pubs.acs.org/JACS
Development of a Concise Synthesis of (+)-Ingenol
Steven J. McKerrall, Lars Jørgensen,† Christian A. Kuttruff,† Felix Ungeheuer, 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: The complex diterpenoid (+)-ingenol possesses a
uniquely challenging scaffold and constitutes the core of a recently
approved anti-cancer drug. This full account details the development
of a short synthesis of 1 that takes place in two separate phases
(cyclase and oxidase) as loosely modeled after terpene biosynthesis.
Initial model studies establishing the viability of a Pauson−Khand
approach to building up the carbon framework are recounted.
Extensive studies that led to the development of a 7-step cyclase
phase to transform (+)-3-carene into a suitable tigliane-type core are also presented. A variety of competitive pinacol
rearrangements and cyclization reactions were overcome to develop a 7-step oxidase phase producing (+)-ingenol. The pivotal
pinacol rearrangement is further examined through DFT calculations, and implications for the biosynthesis of (+)-ingenol are
discussed.
INTRODUCTION
■
For decades, architecturally complex terpenoid natural products
have pushed the limits of both the strategies and methods
employed in organic synthesis. As one of the most diverse
classes of secondary metabolites, terpenes continue to provide
significant challenges through their extensive oxidation and
varied, unique skeletons.1 Terpenes have historically driven
improvements in the synthesis of polycyclic molecules,2
fostered the development of retrosynthetic analysis,3 and led
to numerous broadly useful methods for the formation of
carbon−carbon bonds.4 In addition, the important biological
function and use of terpenoids in medicines necessitate
improved routes for their preparation and for the development
of more medicinally valuable analogues.5 This full account
traces the development of a concise total synthesis of the
complex terpenoid (+)-ingenol (1, Figure 1).6
The Euphorbia genus of plants contains a variety of heavily
oxidized diterpenoid natural products.7 Among the most
Figure 1. Diterpenoids ingenol (1), 20-deoxyingenol (2), and ingenol
notable members of this family of natural products is the
diterpenoid ingenol (1). Ingenol (1) was isolated in 1966 by
Hecker and co-workers from Euphorbia ingens, from which its
name is derived.8 1 was initially isolated on the basis of irritant
activity, and its structure was not elucidated until 1967, when
an X-ray crystal structure of the triacetate was obtained.9 The
structure of 1 contains a number of challenging features found
in many oxidized terpene natural products, including a
congested stereogenic triol moiety, a neighboring all-carbon
quaternary stereocenter, and a dimethylcyclopropane. Unique
to 1, however, is the bicyclo[4.4.1]undecane core bearing a
trans-intrabridgehead stereochemical configuration, also known
as in/out stereochemistry.10,11 The configuration of the
bridgehead hydrogen imparts significant angular strain on the
bicyclic framework, as evidenced by the substantially elongated
bond angles found in the X-ray crystal structure obtained for
20-deoxyingenol (2).
mebutate (3).
In addition to their structural complexity, ingenane
diterpenes possess important biological activity. Derivatives of
1 are reported to possess activity ranging from anti-cancer12a to
anti-HIV;12b however, the most medicinally important
derivative of 1 is ingenol mebutate (3, commercially available
as Picato, developed by LEO Pharma). Ingenol mebutate (3)
was approved by the FDA in early 2012 as a first-in-class topical
treatment for actinic keratosis, a pre-cancerous skin condition.13
Isolated from E. peplus, 3 was discovered through a screen of
medicinal plants used by indigenous populations in Australia.14
In addition to its FDA-approved indications, 3 has undergone
Received: February 26, 2014
Published: April 8, 2014
© 2014 American Chemical Society
5799
dx.doi.org/10.1021/ja501881p | J. Am. Chem. Soc. 2014, 136, 5799−5810

Journal of the American Chemical Society
Article
clinical testing for the topical treatment of both basal cell
carcinoma and squamous cell carcinoma.15 Interestingly, the
mechanism of action is not well understood. A dual-phase
mechanism has been proposed consisting of an initial ill-defined
necrotic cell-killing mechanism, followed by immune stim-
ulation mediated by activation of protein kinase C, ultimately
leading to apoptosis.16 Currently, 3 is obtained by a tedious
To this end, an expedient synthesis was sought that would be
amenable to both the preparation of large quantities of 1 and
the diversion of a late-stage intermediate to novel analogues of
3.22
Our insights into the retrosynthetic analysis of 1 were greatly
influenced by the biosynthesis of Euphorbia diterpenes
(Scheme 2). The biosynthesis is thought to begin with the
formation of casbene (15) and continue with stepwise
cyclizations to form the carbon skeletons of the lathyrane
(16), tigliane (17), ingenane (18), and daphnane (19) natural
products.7,23 It should be noted that only the biogenesis of
casbene (15) has been characterized in any detail, and
subsequent biosynthetic steps are speculative.24 The proposal
of intermediate cyclizations is best supported by the existence
of oxidized natural products (20−23) with each of the
intermediate scaffolds. The level of oxidation that is installed
during the various cyclase steps also remains a mystery, and it is
likely that some cyclase and oxidase steps intertwine in the
biosynthesis of Euphorbia diterpenes. Nevertheless, the general
wisdom of first forming an unfunctionalized carbon skeleton
and then diversifying to various molecules through oxidation
and rearrangement is preserved.
isolation protocol, in a low yield of ∼1.1 mg/kg.17
A
semisynthesis of 3 has also been developed from ingenol (1),
which can be isolated in a greater yield of ∼250 mg/kg.18 While
these methods have proved capable of delivering the necessary
quantities for clinical testing and development, the isolation
procedures are inefficient, difficult, and costly. In addition, a
bioengineering approach is severely hampered by the poorly
understood, complex biosynthesis described below. Most
importantly, from the perspective of developing new medicines,
available semisynthetic modifications of 1 are limited, and a
rigorous structure−activity relationship has remained elusive.
These challenges led LEO Pharma to initiate a collaboration
with our laboratory with the aim of developing a blueprint for
the chemical synthesis of 1 that would be amenable to large-
scale preparation.
The unusual structural features of ingenol (1), combined
with its biological promise, have motivated significant interest
from synthetic chemists. Several decades of effort culminated in
three total syntheses of 1 (Scheme 1),19 one total synthesis of
RESULTS AND DISCUSSION
■
Retrosynthetic Analysis. Our retrosynthetic analysis
began with the realization that the biosynthetic step connecting
the tigliane (17) and ingenane (18) skeletons was a greatly
simplifying disconnection. The rearrangement of 17 to 18
forms the all-carbon quaternary stereocenter, installs the
bicyclic skeleton, and secures the in/out stereochemistry.
While the biosynthesis of 1 remains largely speculative, the
rearrangement is likely a pinacol-type 1,2 migration since
tigliane natural products generally contain the necessary tertiary
alcohol seen in phorbol (23, C-10). Furthermore, application of
the isoprene rule to 1 indicates a structural rearrangement.7 In
further support of this connection, a retro-pinacol rearrange-
ment of 1 has been observed under mild conditions to give
tigliane-type compound 24 (Scheme 3).25 Additionally,
activation of the secondary alcohol in 25 leads to a similar
1,2-shift to give 26.9 In the forward direction, Cha has shown
that epoxy-alcohol 27 rearranges to 28 upon treatment with
AlMe3 to form the carbocyclic skeleton of 1.26 Paquette also
attempted a photo-pinacol rearrangement of 29, but those
efforts were thwarted by competitive rearrangements to give
products such as 31 instead.27 The formation of 24 from 1
supports the biosynthetic relationship but was troubling from
our vantage point because it implies that the pinacol
rearrangement is energetically unfavorable and a tigliane
skeleton (24, 26) is the thermodynamically favored ring
system. In the case of 27, the epoxide strain compensates for
the in/out bicycle, but in our design there is no such driving
force. Despite these clear challenges, we viewed the
disconnection as powerfully simplifying and designed our
synthesis around a biomimetic vinylogous pinacol rearrange-
ment linking the tigliane ring system to the ingenane ring
system.
Scheme 1. Previous Approaches to the Total Synthesis of
Ingenol (1)
In addition to mimicking a biosynthetic reaction to form the
in/out stereochemistry, we also sought to borrow the general
two-phase strategy that nature employs in terpenoid biosyn-
thesis.28 In this design, abiotic carbon−carbon bond-forming
reactions would deliver a key carbon skeleton, and subsequent
substrate-controlled oxidation and rearrangement would deliver
1. Thus, 1 would be formed from 32, bearing minimal oxidation
13-oxyingenol,20 and numerous approaches to the tricyclic
skeleton.21 Each of these syntheses is a landmark achievement,
but they require a large number of chemical manipulations
(37−46 steps) to prepare 1. In order to access fully synthetic
analogues of ingenol mebutate (3), we also required an
approach that would be amenable to late-stage diversification.
5800
dx.doi.org/10.1021/ja501881p | J. Am. Chem. Soc. 2014, 136, 5799−5810

Journal of the American Chemical Society
Article
a
Scheme 2. Unified Biosynthesis of Euphorbia Diterpenoids
a
Abbreviations: DMAPP = dimethylallyl pyrophosphate, FPPS = farnesyl pyrophosphate synthase, GGPPS = geranyl geranyl pyrophosphate
synthase.
Scheme 3. Pinacol and Retropinacol Rearrangements in the Ingenane Scaffold
Scheme 4. Retrosynthetic Analysis of Ingenol (1)
(Scheme 4). Ingenane 32 would be formed by a vinylogous
pinacol rearrangement, via carbocation 33, from tigliane core
34. The tigliane skeleton in 34 would be secured through a
Pauson−Khand (P-K) cyclization29 of 35, which would in turn
be derived from inexpensive (+)-3-carene (36).30 In addition to
producing 1, this route would be amenable to late-stage
diversification of the key hydroxyl groups to form novel
synthetic analogues of ingenol mebutate (3).
5801
dx.doi.org/10.1021/ja501881p | J. Am. Chem. Soc. 2014, 136, 5799−5810

Journal of the American Chemical Society
Article
a
Scheme 5. Synthesis and Evaluation of Ene-yne Pauson−Khand Model Compounds
a
Reagents and conditions: (a) LiHMDS, 4-pentenal, THF, −78 °C, 71%, 2:1 dr; (b) TMS-acetylene, n-BuLi, CeCl3, THF, −78 °C; (c) K2CO3,
MeOH, rt, 67% over 2 steps, 6:1 dr; (d) TBSCl, imidazole, DMF, 60 °C, 79%; (e) Co2(CO)8 (1.0 equiv), NMO, CH2Cl2, 0 °C to rt, 90%; (f) air,
CHCl3, rt, ∼50%; (g) H2O2, NaOH, MeOH, rt, 80%. Abbreviations: LiHMDS = lithium hexamethyldisilazide, THF = tetrahydrofuran, TBSCl = tert-
butyldimethylsilyl chloride, DMF = N,N-dimethylformamide, NMO = N-methylmorpholine N-oxide, Bz* = 2,4-dinitrobenzoate.
Model Pauson−Khand Studies. In order to quickly
establish the viability of a P-K approach to tigliane-type
compound 34, studies began on a more easily accessible model
system lacking the C-ring methyl group. Whereas no routes
were known at the time to the requisite methyl ketone (SI-6),
ketone 37 could be easily prepared from (+)-3-carene (36)
according to literature procedures (Scheme 5).31b Kinetic
deprotonation of 37 with LiHMDS followed by treatment with
4-pentenal gave aldol adduct 38 in 71% yield and a 2:1 mixture
of alcohol epimers. Acetylide addition was accomplished in 2
steps using the TMS-protected cerium acetylide followed by
deprotection with K2CO3 and MeOH to give diol 39 in 67%
yield as a 6:1 mixture of diastereomers. Unfortunately,
cyclization of 39, or a variety of protected derivatives, failed
to deliver 40 under several conditions attempted. In general, no
reaction was observed, followed by eventual decomposition.
The single exception, TBS-protected substrate 41formed
by treatment of 39 with TBSCl in 79% yieldwas reactive
under classical P-K conditions. Thus, treatment of 41 with
stoichiometric Co2(CO)8, followed by NMO, produced an
unusual enone 42 which contains a trans 8-membered ring.
This product presumably arises from a head-to-tail reaction of
the alkene and alkyne rather than the typically observed head-
to-head reaction. In contrast, diol 39 or a variety of related
derivatives simply underwent decomplexation upon treatment
of the in situ formed cobalt−alkyne complex with NMO.
Although this reactivity has been reported previously, the origin
of this selectivity remains unknown.32 Enone 42 was readily
auto-oxidized in the presence of air and light to produce
epoxide 43, and an X-ray crystal structure of derivative 43′
confirmed the unusual ring system and the initial head-to-tail
reaction by inference.
Scheme 6. Previous Examples of Allene-yne Pauson−Khand
Cyclizations
ring, this subtle change would also install an additional olefin in
the cyclase phase product, which could be more easily
transformed to the requisite diol than a saturated counterpart.
The synthesis of the allene-yne model system again begins
with ketone 37, which was deprotonated with LiHMDS then
treated with chiral aldehyde 48 to provide adduct 49 in 62%
yield as a single observable diastereomer (Scheme 7). Two-step
ethynylation using the organocerium reagent derived from 50
and deprotection provided diol 51 in 69% yield as a 5:1 mixture
of diastereomers. Whereas diol 51 decomposed quickly under
standard P-K conditions, the bis-TBS-protected substrate 52
successfully underwent cyclization upon treatment with
catalytic quantities of [RhCl(CO)2]2 under a CO atmosphere
to give dienone 53 in 33% yield as the only observed
regioisomer. With this result in hand, it was clear that an allene-
yne P-K reaction could, in principle, deliver the key precursor
to ingenol (1).
Cyclase Phase. With a proof of concept for constructing a
viable tigliane precursor, the next challenge encountered was
installation of the missing C-11 methyl group.31 To this end, it
was discovered that (+)-3-carene (36) could be regioselectively
chlorinated with NCS and catalytic DMAP (Scheme 8).
Subsequent ozonolysis of the exocyclic olefin delivered 54 in
48% yield over 2 steps. The requisite methyl ketone (SI-6)
could be synthesized as a single diastereomer by treatment with
lithium naphthalenide (LiNap) followed by methyl iodide;
however, it was discovered that isolation of the methyl ketone
was operationally challenging. Gratifyingly, sequential treat-
The failure of 39 to undergo the desired ene-yne P-K
cyclization led us to reconsider our strategy to the tigliane ring
system in 34. Several variants of the P-K reaction have been
detailed.29 In particular, reports from the Brummond and
Mukai laboratories detailing rhodium-catalyzed allene-yne P-K
reactions (Scheme 6) were extremely promising because they
showed examples in which 7- and even 8-membered rings could
be formed in good yield.33 Motivated by these reports, we
revised our strategy to target an allene-yne intermediate. In
addition to aiding in the formation of the central 7-membered
5802
dx.doi.org/10.1021/ja501881p | J. Am. Chem. Soc. 2014, 136, 5799−5810

Journal of the American Chemical Society
Article
a
provided a small amount of product, but the yield could not be
improved. The cyclase phase was completed by treatment of 58
with methylmagnesium bromide to provide 59 as a single
diastereomer in 80% yield with 18% recovered 58. With all of
the carbon atoms in 1 installed in the form of precursor
skeleton 59, the core carbon skeleton was complete, and only
oxidation and skeletal rearrangement remained.
Scheme 7. Successful Allene-yne Model System
Exploring the Pinacol Rearrangement. The first
rearrangements targeted were directly on diene 59. Despite
many attempts, the desired rearrangement to 60 could not be
accomplished (Scheme 9). It was found that under a variety of
a
Scheme 9. Attempted Pinacol Rearrangement of 59
a
Reagents and conditions: (a) LiHMDS, 48, THF, −78 °C, 62%; (b)
TMS-acetylene, n-BuLi, CeCl3, THF, −78 °C; (c) TBAF, THF, 0 °C,
69% over 2 steps, 5:1 dr; (d) TBSOTf, Et3N, CH2Cl2, 0 °C to rt, 85%;
(e) [RhCl(CO)2]2 (0.1 equiv), CO, PhMe, 110 °C, 33%.
Abbreviations: TBAF = tetrabutylammonium fluoride, TBSOTf =
tert-butyldimethylsilyl triflate.
ment of chloroketone 54 with LiNap and MeI, followed by the
addition of LiHMDS and aldehyde 48, delivered aldol product
55 as a single diastereomer in 44% yield in a one-pot operation.
Ethynylation was accomplished by treatment with ethynylmag-
nesium bromide to provide 56 in 81% yield as a 10:1 mixture of
diastereomers. One-pot protection by successive treatment with
TBSOTf then TMSOTf provided 57 in 71% yield. Bis-silyl
ether 57 crystallized upon cooling, and X-ray crystallographic
analysis confirmed the stereochemistry of the four stereocenters
formed in the sequence. P-K cyclization with catalytic
[RhCl(CO)2]2 under CO atmosphere at 0.005 M concen-
tration delivered dienone 58 in 72% yield. A summary of
conditions examined is shown in Scheme 8. It was found that
concentration has a dramatic impact on yield, with the yield
improving from 37% to 72% upon a 10-fold dilution. Increased
temperature and the use of other commonly employed catalysts
provided inferior yields. Interestingly, iridium catalysis also
a
Reagents and conditions: (a) CeCl3, SiO2, CH2Cl2, rt, 92%; (b)
OsO4 (0.1 equiv), NMO, t-BuOH, H2O, rt, 30%.
conditions for activating the tertiary alcohol, only elimination to
form 61 was observed. Treatment of 59 with CeCl3 and SiO2
gave 61 in 92% yield. Fulvene 61 possesses oxidation in
positions that could be useful for preparing 1, so attempts were
made to initiate the rearrangement from it. The fulvene in 61
could be dihydroxylated regioselectively using catalytic OsO4 to
a
Scheme 8. Synthesis of Cyclase Phase End-Point 59
a
Reagents and conditions: (a) NCS, DMAP, CH2Cl2, rt; (b) O3, CH2Cl2, MeOH, −78 °C, then thiourea, rt, 48% over 2 steps; (c) LiNap, MeI,
HMPA, THF, −78 °C, then LiHMDS, 48, −78 °C, 44%; (d) ethynylmagnesium bromide, THF, −78 to 0 °C, 81%, 10:1 dr; (e) TBSOTf, Et3N,
CH2Cl2, 0 °C, then TMOTf, Et3N, 71%; (f) [RhCl(CO)2]2 (0.1 equiv), CO, xylenes, 140 °C, 72%; (g) methylmagnesium bromide, THF, −78 to 0
b
c
°C, 80% + 18% 58. Standard conditions: catalyst (0.1 equiv), CO atmosphere, xylenes, temperature specified. Reaction conducted in 1,2-
d
dichlorobenzene. Added dppp as a ligand. Abbreviations: NCS = N-chlorosuccinimide, DMAP = N,N-dimethylaminopyridine, HMPA =
hexamethylphosphoramide, TMSOTf = trimethylsilyl triflate, dppp = 1,2-bis(diphenylphosphino)propane.
5803
dx.doi.org/10.1021/ja501881p | J. Am. Chem. Soc. 2014, 136, 5799−5810

Journal of the American Chemical Society
Article
observed. Dihydroxylation conducted at high temperature
produced ketol 64 in 45% yield instead of the desired diol.
Surprisingly, even at partial conversion, only a mixture of ketol
64 and diene 59 was observed, implying that either the
intermediate osmate decomposes to 64 directly or the over-
oxidation is extremely rapid.34 Reduction of the ketol 64 gave
the undesired alcohol epimer, and activation of the tertiary
alcohol with TFA produced spirocyclic diketone 65 in 62%
yield as a single diastereomer. This product arises through a
competitive pinacol rearrangement of the C-4 tertiary alcohol
in 64, which secures the observed stereochemical configuration
of the newly formed quaternary carbon.
Scheme 10. Catalytic Dihydroxylation and Pinacol
Rearrangement of 59
a
Successful dihydroxylation was ultimately achieved through a
stoichiometric reaction with OsO4 followed by reductive
cleavage of the resulting osmate ester with Na2SO3 to deliver
the diol 66 (Scheme 11). Treatment of 66 with TFA produces
two major spirocyclic products, 68 and 69which differ only
in the stereochemistry at the all-carbon quaternary stereo-
centerin 26% and 37% yield from 59, respectively. Spiro
ketone 68 could potentially arise from a pinacol rearrangement
reminiscent of the formation of 65; however, the additional
presence of the epimer at the quaternary stereocenter (69)
suggests an alternative mechanism. The likely mechanism
involves an initial Grob-type fragmentation to produce enol
aldehyde 67, followed by cyclization to produce a mixture of 68
and 69. The use of numerous different conditions provided the
same two products with only minor variations in the ratio. The
single exception was found when heating 66 in HFIP produced
epimeric spirocycle 70 and angularly fused 71 in 47% and 35%
yield from 59, respectively. Derivatization of 70 and 71 as
dinitrobenzoates provided crystalline 70′ and 71′, which
allowed for X-ray crystallographic analysis to confirm the
assigned structures. Spiro compound 70 likely arises through
the same Grob-type fragmentation that produces 69, while 71
arises from a direct rearrangement of 66 with migration of the
undesired C-9/C-10 bond. The exact origin of 71 is unclear,
but molecular modeling suggests that an unusual conformation
of 66 is required for migration of the C-9/C-10 bond.
a
Reagents and conditions: (a) OsO4 (0.1 equiv), NMO, t-BuOH,
H2O, 80 °C, 45%; (b) TFA, CH2Cl2, rt, 62%. TFA = trifluoroacetic
acid.
give 62 in 30% yield. Subsequent dihydroxylation failed under
either catalytic or stoichiometric conditions. Again, attempted
rearrangements of 62 to 60 were unsuccessful under a variety of
conditions. Several explanations for the lack of desired reactivity
can be made. For instance, the highly delocalized carbocation
63formed by ionization of 59could be too stable to give
rise to the strained in/out bicycle. Alternatively, simple
molecular modeling indicates that the presence of the diene
forces the migrating bond to be nearly planar to the π-system
and provides extremely poor orbital overlap for migration.
Regardless of the explanation, it was reasoned that the C-4/C-5
olefin was severely limiting the desired reactivity, and removing
it would provide a better opportunity to test the pinacol
rearrangement.
Ultimately, the C-4/C-5 olefin in 59 needed to be converted
to a 1,2-diol, so dihydroxylation of this substrate was attempted
(Scheme 10). It was quickly found that the trisubstituted olefins
in 59 were extremely hindered, and, under most typical
conditions for OsO4-catalyzed dihydroxylation, no turnover was
Spirocyclic rearrangements of diol 66 occur with the cleavage
of the C-4/C-5 bond, so it was reasoned that a cyclic protecting
group that is stable to acidic conditions would prevent
a
Scheme 11. Synthesis and Rearrangements of Diol 66
a
Reagents and conditions: (a) OsO4 (1.0 equiv), pyridine, rt, then saturated aq Na2SO3, THF; (b) TFA, CH2Cl2, rt, 26% 68 + 37% 69; (c) HFIP, 50
°C, 47% 70 + 35% 71. HFIP = hexafluoroisopropanol. The dinitrobenzoate is omitted from the crystal structures shown for 70′ and 71′ for clarity.
See Supporting Information for full images.
5804
dx.doi.org/10.1021/ja501881p | J. Am. Chem. Soc. 2014, 136, 5799−5810

Journal of the American Chemical Society
Article
a
Scheme 12. Synthesis and Successful Rearrangement of Carbonate 72
a
Reagents and conditions: (a) CDI, DMAP, CH2Cl2, rt, 64% from 59; (b) MsCl, Et3N, CH2Cl2, 0 °C, 77%; (c) TBAF, THF, 60 °C, 45%; (d) TFA,
CH2Cl2, rt, 73%; (e) BF3·Et2O, CH2Cl2, −78 °C to −40 °C, then Et3N, MeOH, 80%; (f) HF, CH3CN, 50 °C, 89%. Abbreviations: CDI = N,N-
carbonyldiimidazole, MsCl = methanesulfonyl chloride.
spirocyclization. Additionally, molecular modeling suggested
that the orientation of the migrating bond with respect to the π-
system is improved with a cyclic protecting group on the diol.
Treatment of diol 66 with CDI in the presence of catalytic
quantities of DMAP gives carbonate 72 in 64% yield from 59
(Scheme 12). Gratifyingly, the carbonate protecting group
proved stable to acidic conditions, but subjection to a variety of
activating conditions produced only elimination or no reaction.
Diene 73 was obtained in 77% yield by treatment of 72 with
MsCl and Et3N. To probe whether the bulky silyl protecting
groups were preventing reactivity, 72 was deprotected with
TBAF to provide triol 74 in 45% yield. Treatment of 74 with
TFA instead produced bicyclic ether 75 in 73% yield. Extensive
screening of various Lewis acids, protic acids, and activating
agents on 66, 72, and 74 was unsuccessful. However,
experimenting with the quenching conditions yielded very
promising results. When 72 was treated with BF3·Et2O in
CH2Cl2 and quenched at −78 °C, only starting material was
recovered. When the reaction was instead warmed to 0 °C, a
mixture of eliminated byproducts was obtained. When the
reaction was quenched with saturated aqueous NaHCO3 at
−78 °C, however, a low yield of the desired product 76 was
obtained. A thorough examination revealed that the temper-
ature of the reaction quench was essential to generating the
desired product, with 80% yield of 76 obtained when the
reaction was quenched with 1:1 MeOH/Et3N at −40 °C. While
a comprehensive investigation of Lewis acids under these
specific conditions has not been undertaken, successful pinacol
rearrangement to product 76 has only been observed with the
use of BF3·Et2O. Further experiments also illustrated that 76 is
a kinetic product of the reaction. Treatment of 76 with BF3·
Et2O at 0 °C produced retro-pinacol rearrangement to 77,
while treatment with aqueous HF produced previously
observed ether 75. Success in this instance ultimately hinged
on the recognition that the vinylogous pinacol rearrangement is
reversible, and so the desired, less stable product must be
trapped out as a kinetic product. Further mechanistic analysis is
provided below through DFT calculations.
SeO2, followed by in situ acylation with Ac2O to provide 78 in
59% yield as a single diastereomer (Scheme 13). The observed
selectivity is consistent with a previously observed preference
for secondary over primary positions in SeO2-mediated allylic
oxidation.35 TBS deprotection was accomplished, without
retro-pinacol rearrangement, by treatment with aqueous HF
to give 79 in 90% yield. X-ray crystallographic analysis of the
a
Scheme 13. Completing the Total Synthesis of Ingenol (1)
a
Reagents and conditions: (a) SeO2, dioxane, 80 °C, then Ac2O,
With the in/out bicyclic framework secured, completing the
synthesis required two allylic oxidations and elimination of the
TBS ether. The first allylic oxidation was accomplished with
DMAP, rt, 59%; (b) HF, CH3CN, 50 °C, 90%; (c) Martin sulfurane,
CHCl3, 60 °C, then 10% aq NaOH, THF, rt; (d) SeO2, HCO2H,
dioxane, 80 °C, 76%.
5805
dx.doi.org/10.1021/ja501881p | J. Am. Chem. Soc. 2014, 136, 5799−5810

Journal of the American Chemical Society
Article
Table 1. Computed Gibbs Free Energy (ΔG), Activation Energy (ΔG⧧), and Relative Activation Energy (ΔΔG⧧) of Synthetic
a
Pinacol Rearrangements
a
All energies are reported in kcal/mol. Energies in bold are at the M06-2X/6-31+G(d,p)//B3LYP/6-31G(d) level, while energies in italics are at the
B3LYP/6-31G(d) level. Free energies are at 298.15 K and 1 atm.
corresponding p-bromobenzoate 79′ established the in/out and
acetate stereochemistry. Alcohol elimination was effected by
treatment with Martin sulfurane, followed by basic work-up to
remove the protecting groups, to give 20-deoxyingenol (2) in
81% yield. The final allylic oxidation was accomplished in a
fashion analogous to the Wood synthesis.19c Using Shibuya’s
conditions36 of SeO2 in the presence of formic acid, allylic
oxidation was performed, without over-oxidation, to provide
ingenol (1) in 76% yield. Synthetic 20-deoxyingenol (2) and
ingenol (1) were spectroscopically identical to naturally
obtained material. Thus, the total synthesis of ingenol was
completed in 14 linear steps from (+)-3-carene to give
(+)-ingenol in 1.2% overall yield (73% per step).
DFT Examination of Pinacol Rearrangement. In order
to gain additional insight into the mechanism of the apparently
contra-thermodynamic pinacol rearrangement, DFT calcula-
tions were performed using Gaussian 09 at the B3LYP/6-
31G(d) level, with single-point energies calculated at the M06-
2X/6-31+G(d,p) level (Table 1).37 Initial calculations
suggested that the reaction proceeds in a stepwise manner
through an allylic carbocation.38 When the rearrangement of
carbocation 80 was examined, it was found that the reaction to
form silyl oxonium 81 is both thermodynamically (ΔG = −7.0
kcal/mol) and kinetically favorable (ΔG⧧ = 4.9 kcal/mol). In
order to probe the structural features that are important for
reactivity, various simplified derivatives were also examined.
The rearrangement of cation 82, lacking the TMS ether, is less
thermodynamically favorable and, more importantly, requires
∼2 kcal/mol more activation energy than 80. More
dramatically, the rearrangement of cation 84, lacking the
carbonate protecting group, is thermodynamically unfavorable
(ΔG = 1.7 kcal/mol) and requires ∼9 kcal/mol additional
activation energy relative to 80. This effect appears to be
predominantly conformational in nature, as the rearrangement
of 84′in which the diol is in a conformation similar to the
carbonate 80was nearly identical in activation energy to the
rearrangement of 80. It is ironic that the two major groups that
seem to be responsible for successful reactivity were added
primarily to modulate undesired reactivity, and their substantial
effect on the pinacol rearrangement was unexpected.
Further mechanistic insight was gained by examining the
potential energy surface (PES) for the reaction of 80 and 82.
An intrinsic reaction coordinate (IRC) calculation on TS 80−
81 reveals that the rearrangement of 80 is highly asynchronous,
and the bond-breaking and bond-forming events are separated
by a near plateau on the PES (Figure 2A). On the other hand,
an IRC calculation on TS 82−83 shows that, in the absence of
the TMS group, the asynchronicity disappears, and the PES is
only slightly flat near the transition state (Figure 2B). The same
asynchronicity is observed on other compounds bearing a TMS
5806
dx.doi.org/10.1021/ja501881p | J. Am. Chem. Soc. 2014, 136, 5799−5810

Journal of the American Chemical Society
Article
Figure 2. Mechanistic analysis of the pinacol rearrangement. (A) Intrinsic reaction coordinate (IRC) diagram of the rearrangement of 80 to 81
(energies uncorrected for ZPE). (B) IRC diagram for the rearrangement of 82 to 83 (energies uncorrected for ZPE). (C) Simplified mechanism of
the vinylogous pinacol rearrangement. Energies are given in kcal/mol, and the reaction coordinate is in units of amu1/2bohr. Energies in bold are at
the M06-2X/6-31+G(d,p)//B3LYP/6-31G(d) level, while energies in italics are at the B3LYP/6-31G(d) level. Free energies are at 298.15 K and 1
atm. Reactions indicated by dashed arrows were not examined computationally.
group and not found on those without one.39 This mechanistic
shift is attributed to a stabilizing effect of the silicon atom on
the positive charge in 81 and the developing positive charge in
the transition state. The stabilizing effect seen here is akin to
the commonly encountered β-silicon effect, exemplified by
resonance structure 81′.40 An analogous mechanism mediated
by a proton (H+) is shown in Figure 2C. Initial loss of water
produces carbocation 80, which rearranges exergonically to 81
over a low barrier. Loss of a TMS grouppossibly in the form
of TMSF in the case of BF3·Et2Oleads to 76 as the initially
formed kinetic product. Retro-pinacol rearrangement is
initiated by protonation of the ketone to form 83. The barrier
for the reverse rearrangement from 83 to 82 is predicted to be
∼4 kcal/mol higher than the barrier for 80 to 81. Thus,
products 73 and 75 only arise under thermodynamic conditions
when sufficient energy is available to cross the higher barrier for
retro-pinacol rearrangement in the absence of the silyl group.
This mechanistic picture is greatly simplifiedrelying on a
proton instead of BF3 and neglecting solvent effectsbut it
provides an approximate explanation of the factors governing
the selectivity observed in the pivotal pinacol rearrangement.
Pinacol Rearrangement in the Biosynthesis of Ingenol
(1). With an approximate picture of the mechanism of the
pinacol rearrangement in our synthesis, we sought to expand
the DFT calculations to reactions with relevance to the
biosynthesis of ingenol (1). The simplest putative precursor of
ingenol is 86, which is an oxidized tigliane derivative with the
requisite tertiary alcohol and allylic carbocation (Table 2).
Surprisingly, the rearrangement of 86 to 87 is highly
endergonic (ΔG = 12.6 kcal/mol), and the activation barrier
is relatively high (ΔG⧧ = 17.8 kcal/mol). This exemplifies the
substantial strain of the ingenane ring system. The unfavorable
energetics of the pinacol rearrangement must be overcome in
order for the in/out bicycle to form in nature. Given that
substituents on the tertiary alcohol seem to have a beneficial
effect on the pinacol rearrangement in our synthesis, it was
hypothesized that Lewis base activation of the tertiary alcohol
would stabilize the transition state and reduce the activation
energy. To this end, coordination of a simple imidazole or
methylamine theozyme to the tertiary alcohol was found to
reduce the activation energy of the pinacol rearrangement by
∼5 kcal/mol.41 These results suggest that a viable mechanism
for enzyme catalysis is a Lewis-basic active site amino acid
(histidine or lysine), which would coordinate the tertiary
alcohol and stabilize the developing positive charge in the
transition state. Additionally, it is possible that additional
peripheral oxidation is present in the precursor tigliane
molecule prior to the pinacol rearrangement. Indeed, oxidation
in the A-ring significantly reduces the activation barrier for
pinacol rearrangement and destabilizes the carbocation
precursor. Carbocation 92 has a ∼8 kcal/mol lower energy
barrier to rearrangement than 86. This result is also consistent
with observed reactivity. Compound 76 undergoes retro-
pinacol rearrangement upon treatment with aqueous HF,
while 78which has only an additional acetate on the A-ring
deprotects cleanly without rearrangement under identical
5807
dx.doi.org/10.1021/ja501881p | J. Am. Chem. Soc. 2014, 136, 5799−5810

Journal of the American Chemical Society
Article
Table 2. Computed Gibbs Free Energy (ΔG), Activation Energy (ΔG⧧), and Relative Activation Energy (ΔΔG⧧) of Biosynthetic
a
Pinacol Rearrangements
a
All energies are reported in kcal/mol. Energies in bold are at the M06-2X/6-31+G(d,p)//B3LYP/6-31G(d) level, while energies in italics are at the
B3LYP/6-31G(d) level. Free energies are at 298.15 K and 1 atm.
conditions. These results suggest that electron-withdrawing
substituents around the allylic carbocation destabilize the
tigliane precursor relative to the rearrangement product, so a
pinacol rearrangement after peripheral p450 oxidation would be
more favorable. Overall, two distinct mechanisms have been
proposed that could act synergistically to render the
biosynthetic pinacol rearrangement more kinetically and
thermodynamically favorable.
order to discover a kinetically controlled, vinylogous pinacol
rearrangement that furnishes the in/out bicyclic core in a
further 3 steps. The oxidase phase was completed through the
use of stereo- and regioselective allylic oxidations to deliver
(+)-ingenol (1) in only 14 steps from (+)-3-carene (36). The
high level of redox economy is an especially notable feature of
the synthesis; however, of the 14 steps, 5 are concession steps,
and thus even further refinement to this route can be
envisioned.42 In addition to providing a concise route to 1,
the synthesis adds validity to the biosynthetic proposal of a
pinacol rearrangement to form the in/out bicyclic skeleton of 1.
Finally, DFT calculations were employed to further understand
factors influencing the pinacol rearrangement in the context of
both the chemical synthesis and the biosynthesis of 1. The
synthesis of ingenol (1) presented herein allows access to
novel, fully synthetic analogues of the anti-cancer drug ingenol
mebutate (Picato, 3) and provides a blueprint for commercial
production. Further work on improving the scalability of the
synthesis and the preparation of new analogues of 3 is ongoing
in collaboration with LEO Pharma and will be reported in due
course.
CONCLUSION
■
Ingenol (1) has tested the creativity of synthetic chemists for
nearly three decades with its unique architecture and
stereochemical challenge. It is a fabulous case study for a
terpene that attracted enormous attention initially based solely
on its challenging structural features and later because of its
proven medicinal properties. Those path-pointing studies,
outlined in dozens of research articles and theses, formed a
solid foundation for the development of a concise synthesis of
(+)-ingenol (1), as outlined in this full account. Armed with a
retrosynthesis featuring a biomimetic yet speculative pinacol
rearrangement, we developed a two-phase approach featuring
an allene-yne P-K reaction to establish a tigliane-type precursor
in only 7 steps. Competitive rearrangements were tamed in
5808
dx.doi.org/10.1021/ja501881p | J. Am. Chem. Soc. 2014, 136, 5799−5810

Journal of the American Chemical Society
Article
(12) (a) Ogbourne, S. M.; Suhrbier, A.; Jones, B.; Cozzi, S.; Boyle, G.
M.; Morris, M.; McAlpine, D.; Johns, J.; Scott, T. M.; Sutherland, K.
P.; Gardner, J. M.; Le, T. T. T.; Lenarczyk, A.; Aylward, J. H.; Parsons,
P. G. Cancer Res. 2004, 64, 2833−2839. (b) Fujiwara, M.; Fujiwara,
M.; Ijichi, K.; Tokuhisa, K.; Katsuura, K.; Shigeta, S.; Konno, K.;
Wang, G. Y.; Uemura, D.; T Yokota, T.; Baba, M. Antimicrob. Agents
Chemother. 1996, 40, 271−273.
ASSOCIATED CONTENT
■
S
* Supporting Information
Experimental details, spectra, X-ray crystallographic data,
computational details, and computed structures and energies.
This material is available free of charge via the Internet at
http://pubs.acs.org.
(13) U.S. Food and Drug Administration, 2012 Novel New Drugs
Summary (FDA, Washington, DC, 2013); www.fda.gov/downloads/
Drugs/DevelopmentApprovalProcess/DrugInnovation/UCM337830.
pdf.
AUTHOR INFORMATION
■
Corresponding Author
pbaran@scripps.edu
(14) For a review of the development of ingenol mebutate (3), see:
Vasas, A.; Red
Chem. 2012, 5115−5130.
(15) (a) Siller, G.; Rosen, R.; Freeman, M.; Welburn, P.; Katsamas, J.;
Ogbourne, S. M. Aust. J. Dermatol. 2010, 51, 99−105. (b)
ClinicalTrials.gov, NCT00329121, http://clinicaltrials.gov/show/
NCT00329121.
́ ́
ei, D.; Csupor, D.; Molnar, J.; Hohmann, J. Eur. J. Org.
Author Contributions
†L.J. and C.A.K. contributed equally to this work.
Notes
The authors declare no competing financial interest.
(16) (a) Rosen, R. H.; Gupta, A. K.; Tyring, S. K. J. Am. Acad.
Dermatol. 2012, 66, 486−493. (b) Ersvaer, E.; Kittang, A. O.;
Hampson, P.; Sand, K.; Gjertsen, B. T.; Lord, J. M.; Bruserud, Ø
Toxins 2010, 2, 174−194. (c) Serova, M.; Ghoul, A.; Benhadji, K. A.;
Faivre, S.; Le Tourneau, C.; Cvitkovic, E.; Lokiec, F.; Lord, J.;
Ogbourne, S. M.; Calvo, F.; Raymond, E. Mol. Cancer Ther. 2008, 7,
915−922.
ACKNOWLEDGMENTS
■
Financial support for this work was provided by LEO Pharma.
The Carlsberg Foundation and Danish Council for Independ-
ent Research (Technology and Production Sciences) provided
postdoctoral fellowships for L.J., and the Alexander von
Humboldt Foundation provided a postdoctoral fellowship for
C.A.K. We are grateful to J. Felding and H. Bladh (LEO
Pharma) for helpful discussions. We thank LEO Pharma for
generous donations of natural samples of ingenol and 20-
deoxyingenol and the High-Performance Computing Center
(TSRI) for computational services. We thank Prof. A Rheingold
and C. Moore (UCSD) for X-ray crystallographic measure-
ments and G. Siuzdak (TSRI) for mass spectrometry assistance.
A provisional patent application has been filed and is available
under patent application no. U.S. 61/829,861.
́
(17) Hohmann, J.; Evanics, F.; Berta, L.; Bartok, T. Planta Med.
2000, 66, 291−294.
(18) (a) Liang, X.; Grue-Sørensen, G.; Petersen, A. K.; Hogberg, T.
̈
Synlett 2012, 23, 2647−2652. (b) Appendino, G.; Tron, G. C.;
Cravatto, G.; Palmisano, G.; Jakupovic, J. Nat. Prod. Rep. 1999, 62,
76−79.
(19) (a) Winkler, J. D.; Rouse, M. B.; Greaney, M. F.; Harrison, S. J.;
Jeon, Y. T. J. Am. Chem. Soc. 2002, 124, 9726−9728. (b) Tanino, K.;
Onuki, K.; Asano, K.; Miyashita, M.; Nakamura, T.; Takahashi, Y.;
Kuwajima, I. J. Am. Chem. Soc. 2003, 125, 1498−1500. (c) Nickel, A.;
Maruyama, T.; Tang, H.; Murphy, P. D.; Greene, B.; Yusuff, N.;
Wood, J. L. J. Am. Chem. Soc. 2004, 126, 16300−16301.
(20) Ohyoshi, T.; Ohyoshi, T.; Funakubo, S.; Miyazawa, Y.; Niida,
K.; Hayakawa, I.; Kigoshi, H. Angew. Chem., Int. Ed. 2012, 51, 4972−
4975.
(21) For thorough reviews of the approaches to and syntheses of
ingenol, see: (a) Kim, S.; Winkler, J. D. Chem. Soc. Rev. 1997, 26, 387−
399. (b) Kuwajima, I.; Tanino, K. Chem. Rev. 2005, 105, 4661−4670.
(c) Cha, J. K.; Epstein, O. L. Tetrahedron 2006, 62, 1329−1343.
(22) (a) Kuttruff, C. A.; Eastgate, M. D.; Baran, P. S. Nat. Prod. Rep.
2014, 31, 419−432. (b) Szpilman, A. M.; Carreira, E. M. Angew.
Chem., Int. Ed. 2010, 49, 9592−9628. (c) Newhouse, T.; Baran, P. S.;
Hoffmann, R. W. Chem. Soc. Rev. 2009, 38, 3010−3021. (d) Wender,
P. A.; Verma, V. A.; Paxton, T. J.; Pillow, T. H. Acc. Chem. Res. 2008,
41, 40−49.
REFERENCES
■
(1) Maimone, T. J.; Baran, P. S. Nat. Chem. Biol. 2007, 3, 396−407.
(2) (a) Velluz, L.; Valls, J.; Nomine,
41, 181−200. (b) Yoder, R. A.; Johnston, J. N. Chem. Rev. 2005, 105,
4730−4756.
(3) (a) Corey, E. J.; Ohno, M.; Vatakencherry, P. A.; Mitra, R. B. J.
Am. Chem. Soc. 1961, 83, 1251−1253. (b) Corey, E. J.; Ohno, M.;
Mitra, R. B.; Vatakencherry, P. A. J. Am. Chem. Soc. 1964, 86, 478−
485. (c) Corey, E. J.; Cheng, X. The Logic of Chemical Sythesis; John
Wiley & Sons: New York, 1995.
(4) (a) Rapson, W. S.; Robinson, R. J. Chem. Soc. 1935, 1285−1288.
(b) Lo, J. C.; Yabe, Y.; Baran, P. S. J. Am. Chem. Soc. 2014, 136, 1304−
1307. (c) McCarroll, A. J.; Walton, J. C. Angew. Chem., Int. Ed. 2001,
40, 2224−2248.
́
G. Angew. Chem., Int. Ed. 1965,
(23) (a) Adolf, W.; Hecker, E. Isr. J. Chem. 1977, 16, 75−83.
(b) Schmidt, R. J. Bot. J. Linn. Soc. 1987, 94, 221−230.
(24) Kirby, J.; Nishimoto, M.; Park, J. G.; Withers, S. T.; Nowroozi,
F.; Behrendt, D.; Garcia Rutledge, E. J.; Fortman, J. L.; Johnson, H. E.;
Anderson, J. V.; Keasling, J. D. Phytochemistry 2010, 71, 1466−1473.
(25) Appendino, G.; Tron, G. C.; Cravotto, G.; Palmisano, G.;
Annunziata, R.; Baj, G.; Surico, N. Eur. J. Org. Chem. 1999, 3413−
3420.
(5) (a) Raskin, I.; Ribnicky, D. M.; Komarnytsky, S.; Ilic, N.; Poulev,
A.; Borisjuk, N.; Brinker, A.; Moreno, D. A.; Ripoll, C.; Yakoby, N.;
O’Neal, J. M.; Cornwell, T.; Pastor, I.; Fridlender, B. Trends in
Biotechnol. 2002, 20, 522−531. (b) Koehn, F. E.; Carter, G. T. Nat.
Rev. Drug Discuss. 2005, 4, 206−220. (c) Yang, H.; Dou, Q. P. Curr.
Drug Targets 2010, 11, 733−744. (d) Gershenzon, J.; Dudareva, N.
Nat. Chem. Biol. 2007, 3, 408−414.
(6) (a) Jørgensen, L.; McKerrall, S. J.; Kuttruff, C. A.; Ungeheuer, F.;
Felding, J.; Baran, P. S. Science 2013, 341, 878−882. (b) Appendino,
G. Angew. Chem., Int. Ed. 2014, 53, 927−929.
(26) Epstein, O. L.; Cha, J. K. Angew. Chem., Int. Ed. 2005, 44, 121−
123.
(27) (a) Paquette, L. A.; Zhao, Z.; Gallou, F.; Liu, J. J. Am. Chem. Soc.
2000, 122, 1540−1541. (b) Paquette, L. A.; Gallou, F.; Zhao, Z.;
Young, D. G.; Liu, J.; Yang, J.; Friedrich, D. J. Am. Chem. Soc. 2000,
122, 9610−9620.
(28) (a) Ishihara, Y.; Baran, P. S. Synlett 2010, 12, 1733−1745.
(b) Chen, K.; Baran, P. S. Nature 2009, 459, 824−828. (c) Mendoza,
A.; Ishihara, Y.; Baran, P. S. Nat. Chem. 2012, 4, 21−25. (d) Cherney,
E. C.; Green, J. C.; Baran, P. S. Angew. Chem., Int. Ed. 2013, 52, 9019−
9022.
(7) Hecker, E. Pure Appl. Chem. 1977, 49, 1423−1431.
(8) Hecker, E. Cancer Res. 1968, 28, 2338−2348.
(9) Zeichmeister, K.; Brandl, F.; Hoppe, W.; Hecker, E.; Opferkuch,
H. J.; Adolf, W. Tetrahedron Lett. 1970, 47, 4075−4078.
(10) Only one additional natural product type is known to contain
in/out stereochemistry, and it contains a macrocyclic (14-membered)
ring in the bicycle skeleton. See ref 11.
(11) For a review of in/out isomerism, see: Alder, R. W.; East, S. P.
Chem. Rev. 1996, 96, 2097−2111.
5809
dx.doi.org/10.1021/ja501881p | J. Am. Chem. Soc. 2014, 136, 5799−5810

Journal of the American Chemical Society
Article
(29) (a) Blanco-Urgoiti, J.; Anorbe, L.; Per
́
ez-Serrano, L.;
̃
NOTE ADDED AFTER ASAP PUBLICATION
■
Domínguez, G.; Per
́
ez-Castells, J. Chem. Soc. Rev. 2004, 33, 32−42.
The structures for compounds 29 and 30 in Scheme 3 were
corrected, and the reagents used to give diol 39 were corrected
in the text discussion of Scheme 5. The corrected version was
re-posted on April 16, 2014.
(b) Shibata, T. Adv. Synth. Catal. 2006, 348, 2328−2336. For an
example of a Pauson−Khand reaction in a scalable total synthesis, see:
(c) Su, S.; Rodriguez, R. A.; Baran, P. S. J. Am. Chem. Soc. 2011, 133,
13922−13925.
(30) At the time of writing, (+)-3-carene (90% chemical purity) is
commercially available from Sigma-Aldrich for $86/L (115576-1L).
(31) The requisite methyl ketone (SI-6) has a noteworthy history.
The enantiomer was examined (as a mixture of diastereomers) in 1967
by H. C. Brown and A. Suzuki,who surprisingly found it is the
thermodynamically disfavored diastereomer: (a) Brown, H. C.; Suzuki,
A. J. Am. Chem. Soc. 1967, 89, 1933−1941. Later, Paqeutte and co-
workers attempted to synthesize SI-6 for use in phorbol (23) and
ingenol (1) total syntheses but were unable to form it with acceptable
regio- or stereocontrol. These struggles are detailed at length in the
following and were very inspirational to our work: (b) Paquette, L. A.;
Ross, R. J.; Shi, Y. J. Org. Chem. 1990, 55, 1589−1598. We examined
several different approaches before success was achieved as described.
(32) Lovely, C. J.; Seshadri, H.; Wayland, B. R.; Cordes, A. W. Org.
Lett. 2001, 3, 2607−2610.
(33) (a) Mukai, C.; Nomura, I.; Yamanishi, K.; Hanaoka, M. Org.
Lett. 2002, 4, 1755−1758. (b) Brummond, K. M.; Chen, H.; Fisher, K.
D.; Kerekes, A. D.; Rickards, B.; Sill, P. C.; Geib, S. J. Org. Lett. 2002,
4, 1931−1934. (c) Mukai, C.; Nomura, I.; Kitagaki, S. J. Org. Chem.
2003, 68, 1376−1385. (d) Brummond, K. M.; Chen, D.; Davis, M. M.
J. Org. Chem. 2008, 73, 5064−5068.
(34) For a mechanistic study on over-oxidation in OsO4-catalyzed
dihydroxylations, see: Lohray, B. B.; Bhushan, V.; Kumar, R. K. J. Org.
Chem. 1994, 59, 1375−1380.
(35) (a) Hoekstra, W. J.; Fairlamb, I. J. S.; Giroux, S.. Selenium(IV)
Oxide. e-EROS Encyclopedia of Reagents for Organic Synthesis; John
Wiley
& Sons, Ltd.: New York, 2007 (DOI: 10.1002/
047084289X.rs008.pub2). (b) Nakamura, A.; Nakata, M. Synthesis
2013, 45, 1421−1451.
(36) Shibuya, K. Synth. Commun. 1994, 24, 2923−2941.
(37) (a) Becke, A. D. J. Chem. Phys. 1993, 98, 5648−5652. (b) Zhao,
Y.; Truhlar, D. G. Theor. Chem. Acc. 2008, 120, 215−241. Gaussian 09
was utilized for all calculations: (c) Frisch, M. J.; et al. Gaussian09,
Revision D.01; Gaussian, Inc.: Pittsburgh, PA, 2009 (full reference in
the Supporting Information). ; Structures were visualized with
́
CYLView: (d) Legault, C. Y. CYLview, 1.0b; Universite de Sherbrooke:
Quebec, Canada, 2009; http://www.cylview.org. A detailed procedure
for calculations and a full list of references for computational methods
employed is provided in the Supporting Information. Computed
energies at the mPW1PW91/6-31+G(d,p)//B3LYP/6-31G(d) level
and additional computed structures and energies relevant to the
biosynthesis of 1 are also presented in the Supporting Information.
(38) Protonated 72 was not found to be a stationary point. Instead,
geometry optimization of protonated 72 led to carbocation 80 and
H2O, indicating the activation barrier for ionization is very low or zero
and the reaction downhill in energy. Additionally, in a related
compound series, both epimers of the tertiary alcohol provide product,
implying a cationic mechanism.
(39) Additional support for an asynchronous reaction can be found
in the plot of RMS gradient along the IRC (see Supporting
Information).
(40) Wierschke, S. G.; Chandrasekhar, J.; Jorgensen, W. L. J. Am.
Chem. Soc. 1985, 107, 1496−1500.
(41) For a review on the use of theozymes (theoretical enzyme
mimics made up of model functional groups) in computational
chemistry, see: Tantillo, D. J.; Chen, J.; Houk, K. N. Curr. Opin. Chem.
Biol. 1998, 2, 743−750.
(42) For detailed discussions of redox economy and step economy,
see: (a) Burns, N. Z.; Baran, P. S.; Hoffmann, R. W. Angew. Chem., Int.
Ed. 2009, 48, 2854−2867. (b) Gaich, T.; Baran, P. S. J. Org. Chem.
2010, 75, 4657−4673.
5810
dx.doi.org/10.1021/ja501881p | J. Am. Chem. Soc. 2014, 136, 5799−5810