14-Step synthesis of (+)-ingenol from (+)-3-carene

Article abstract of DOI:10.1126/science.1241606

Ingenol is a diterpenoid with unique architecture and has derivatives possessing important anticancer activity, including the recently Food and Drug Administration-approved Picato, a first-in-class drug for the treatment of the precancerous skin condition actinic keratosis. Currently, that compound is sourced inefficiently from Euphorbia peplus. Here, we detail an efficient, highly stereocontrolled synthesis of (+)-ingenol proceeding in only 14 steps from inexpensive (+)-3-carene and using a two-phase design. This synthesis will allow for the creation of fully synthetic analogs of bioactive ingenanes to address pharmacological limitations and provides a strategic blueprint for chemical production. These results validate two-phase terpene total synthesis as not only an academic curiosity but also a viable alternative to isolation or bioengineering for the efficient preparation of polyoxygenated terpenoids at the limits of chemical complexity.

Full text of DOI:10.1126/science.1241606

1
4-Step Synthesis of (+)-Ingenol from (+)-3-Carene
Lars Jørgensen et al.
Science 341, 878 (2013);
DOI: 10.1126/science.1241606
This copy is for your personal, non-commercial use only.
If you wish to distribute this article to others, you can order high-quality copies for your
colleagues, clients, or customers by clicking here.
Permission to republish or repurpose articles or portions of articles can be obtained by
following the guidelines here.
The following resources related to this article are available online at
www.sciencemag.org (this information is current as of May 29, 2014 ):
Updated information and services, including high-resolution figures, can be found in the online
version of this article at:
http://www.sciencemag.org/content/341/6148/878.full.html
Supporting Online Material can be found at:
http://www.sciencemag.org/content/suppl/2013/07/31/science.1241606.DC1.html
This article cites 34 articles, 6 of which can be accessed free:
http://www.sciencemag.org/content/341/6148/878.full.html#ref-list-1
This article appears in the following subject collections:
Chemistry
http://www.sciencemag.org/cgi/collection/chemistry
Science (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by the
American Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. Copyright
2
013 by the American Association for the Advancement of Science; all rights reserved. The title Science is a
registered trademark of AAAS.

REPORTS
depth where upward migrating Fe(II) was first importance of one-electron–transfer reactions for
detected. Given the calculated fluxes of Mn(III) the Mn cycle (Fig. 2A) and explicitly consider dis-
and Fe(II) and the 1:1 stoichiometry in reaction solved Mn(III) and Fe(III) (Fig. 2B). Second,
17. R. E. Trouwborst, B. G. Clement, B. M. Tebo, B. T. Glazer,
G. W. Luther III, Science 313, 1955–1957 (2006).
8. E. Yakushev, S. Pakhomova, K. Sørenson, J. Skei,
1
Mar. Chem. 117, 59–70 (2009).
19. O. Dellwig, B. Schnetger, H. J. Brumsack, H. P. Grossart,
5, we estimated that up to 31% of Mn(III) can because Mn(III) can act as either an electron accep-
L. Umlauf, J. Mar. Syst. 90, 31–41 (2012).
be reduced by Fe(II) at Station CE (Table 1).
tor or an electron donor, the reduction-oxidation
capacity of the soluble Mn pool in sediments has
been underestimated. For (hemi)pelagic sediments,
2
0. B. Schnetger, O. Dellwig, J. Mar. Syst. 90, 23–30 (2012).
21. A. S. Madison, B. M. Tebo, G. W. Luther 3rd, Talanta 84,
74–381 (2011).
Mn3þ-L þ Fe2þ → Mn þ Fe3þ-L ð5Þ
2þ
3
At stations Cabot (13%) and 16 (0.5%), this it appears that oxidation of Mn(II) by O
reduction pathway is less important. Although important pathway to soluble Mn(III). Soluble
these results imply that abiotic Mn(III) reduction Mn(III) intermediates are also produced during
is an
22. P. Anschutz, B. Sundby, L. Lefrancois, G. W. Luther III,
2
A. Mucci, Geochim. Cosmochim. Acta 64, 2751–2763 (2000).
3. H. Lin, N. H. Szeinbaum, T. J. Dichristina, M. Taillefert,
2
Geochim. Cosmochim. Acta 99, 179–192 (2012).
24. B. P. Boudreau, A. Mucci, B. Sundby, G. W. Luther,
is important for Mn(III) removal, it is not the dissimilatory MnO
most important anoxic removal pathway. Never- mineralization and during abiotic reduction of
theless, it does have important implications for MnO by reductants such as Fe(II) and H S.
2
reduction upon organic-matter
N. Silverberg, J. Mar. Res. 56, 1259–1284 (1998).
5. S. Mulsow, B. P. Boudreau, J. N. Smith, Limnol.
Oceanogr. 43, l–9 (1998).
6. S. Katsev, G. Chaillou, B. Sundby, A. Mucci,
Limnol. Oceanogr. 52, 2555–2568 (2007).
7. B. Sundby, N. Silverberg, R. Chesselet, Geochim.
Cosmochim. Acta 45, 293–307 (1981).
8. G. W. Luther III, B. Sundby, B. L. Lewis, P. J. Brendel,
N. Silverberg, Geochim. Cosmochim. Acta 61,
4043–4052 (1997).
2
2
2
2
2
2
the Fe cycle, as the downward flux of Mn(III) Soluble Mn(III) is likely to be ubiquitous in sedi-
is sufficient to oxidize the entire upward flux of ment porewaters, where it would facilitate key
Fe(II). This pathway, along with several others electron-transfer processes in global biogeochem-
(
29, 30), also leads to soluble and/or colloidal ical cycles.
organic Fe(III) production as shown in Fig. 1 and
figs. S3 and S4 [Fe(III) = Fetotal – Fe(II)]. The
reduction of Mn(III) by Mn-reducing bacteria, such
as S. putrefaciens (14, 23), likely accounts for the
consumption of the remaining (69 to 99.5%)
downward flux of Mn(III). Prior studies found
a linear correlation between Mn(III) reduction and
carbon oxidation (14, 23), so this pathway could
have implications for organic-matter mineral-
ization. For these reasons, we have revised the
current sedimentary redox paradigm (5) to in-
clude the above Mn(III) production and removal
pathways (Fig. 2). We have also added reactions
References and Notes
1
2
. B. L. Lewis, W. M. Landing, Deep-Sea Res. 38,
S773–S803 (1991).
29. G. W. Luther III, J. E. Kostka, T. M. Church,
B. Sulzberger, W. Stumm, Mar. Chem. 40, 81–103 (1992).
30. M. Taillefert, A. B. Bono, G. W. Luther, Environ. Sci.
Technol. 34, 2169–2177 (2000).
31. B. G. Clement, G. W. Luther III, B. M. Tebo,
Geochim. Cosmochim. Acta 73, 1878–1889 (2009).
32. N. P. Revsbech et al., Limnol. Oceanogr. Methods 7,
371–381 (2009).
. B. M. Tebo, Deep-Sea Res. A, Oceanogr. Res. Pap. 38,
S883–S905 (1991).
3. G. W. Luther III, Geomicrobiol. J. 22, 195–203 (2005).
. G. W. Luther, Aquat. Geochem. 16, 395–420 (2010).
4
5
. P. Froelich et al., Geochim. Cosmochim. Acta 43,
1075–1090 (1979).
6. S. Emerson et al., Geochim. Cosmochim. Acta 46,
1073–1079 (1982).
Acknowledgments: This work was funded by grants from
the Chemical Oceanography Division of the National Foundation
to G.W.L. (OCE-1155385 and OCE-1031272) and B.M.T.
7
8
. R. C. Aller, Philos. Trans. R. Soc. Lond. A 331, 51–68 (1990).
. T. J. Shaw, J. M. Gieskes, R. A. Jahnke, Geochim.
Cosmochim. Acta 54, 1233–1246 (1990).
9. K. S. Johnson et al., Science 257, 1242–1245 (1992).
0. L. N. Neretin, C. Pohl, G. Jost, T. Leipe, F. Pollehne,
Mar. Chem. 82, 125–143 (2003).
(
OCE-1031200 and OCE-1154307), and from the Natural Sciences
and Engineering Research Council of Canada (NSERC) through
Discovery and Ship-time grants to A.M. and B.S. Raw data are
presented in the supplementary materials.
1
1
1
and 4, as well as an unknown generalized Mn(III)
microbial reduction pathway (see supplemen-
tary materials), to an established biogeochemical
model of sediment diagenesis (24). The model re-
sults (Fig. 1, A and H) reproduce the position and
shape of the soluble Mn(III) peak.
1. A. Schippers, L. N. Neretin, G. Lavik, T. Leipe, F. Pollehne,
Geochim. Cosmochim. Acta 69, 2241–2252 (2005).
12. W. B. Homoky et al., Geochim. Cosmochim. Acta 75,
032–5048 (2011).
3. O. W. Duckworth, G. Sposito, Environ. Sci. Technol. 39,
037–6044 (2005).
Supplementary Materials
www.sciencemag.org/cgi/content/full/341/6148/875/DC1
Materials and Methods
5
1
6
Supplementary Text
Our findings have broad implications for un-
derstanding Mn cycling in sediments. First, the
current sedimentary redox model (5) should be
revised to accommodate soluble Mn(III) and Fe(III)
species. A proposed revision should highlight the
14. J. E. Kostka, G. W. Luther, K. H. Nealson, Geochim.
Figs. S1 to S7
Tables S1 and S2
References (33–45)
Cosmochim. Acta 59, 885–894 (1995).
1
5. K. J. Klewicki, J. J. Morgan, Environ. Sci. Technol. 32,
916–2922 (1998).
2
16. D. L. Parker, G. Sposito, B. M. Tebo, Geochim. Cosmochim.
3 June 2013; accepted 29 July 2013
10.1126/science.1241396
Acta 68, 4809–4820 (2004).
(
artemisinin) (1). Despite the promise of complex
1
4-Step Synthesis of (+)-Ingenol
from (+)-3-Carene
1
terpenoids as drug molecules, their utility has been
hampered by a number of challenges to devel-
opment (2). In particular, many plant-derived
terpenoid natural products, such as Taxol and
artemisinin, suffer from a combination of low iso-
lation yields, inconsistent isolation procedures,
and nonrenewable natural sources (3, 4). In the
case of Taxol, plant cell culture technology over-
came low-yielding isolation (3), whereas in the
1
1
1
Lars Jørgensen, * Steven J. McKerrall, * Christian A. Kuttruff, Felix Ungeheuer,
Jakob Felding, Phil S. Baran †
2
1
Ingenol is a diterpenoid with unique architecture and has derivatives possessing important
anticancer activity, including the recently Food and Drug Administration–approved Picato, a
first-in-class drug for the treatment of the precancerous skin condition actinic keratosis. Currently, case of artemisinin, collaborations between scien-
that compound is sourced inefficiently from Euphorbia peplus. Here, we detail an efficient,
highly stereocontrolled synthesis of (+)-ingenol proceeding in only 14 steps from inexpensive
tists in genetic engineering and chemical syn-
thesis hold promise for providing a consistent
supply (4, 5). These seminal works in the bio-
logical production of natural products have led
(+)-3-carene and using a two-phase design. This synthesis will allow for the creation of fully
synthetic analogs of bioactive ingenanes to address pharmacological limitations and provides a
strategic blueprint for chemical production. These results validate two-phase terpene total synthesis to a widely held presumption that bioengineering
as not only an academic curiosity but also a viable alternative to isolation or bioengineering for
the efficient preparation of polyoxygenated terpenoids at the limits of chemical complexity.
1
Department of Chemistry, The Scripps Research Institute,
1
0550 North Torrey Pines Road, La Jolla, CA 92037, USA.
2
External Discovery, LEO Pharma A/S, Industriparken 55,
tructurally complex, polyoxygenated ter- used in a myriad of different therapeutic areas 2750 Ballerup, Denmark.
penoids and their derivatives constitute a such as oncology (Taxol, Bristol-Myers Squibb),
*These authors contributed equally to this work.
S
78
medicinally vital class of natural products immunology (prednisone), and infectious diseases †Corresponding author. E-mail: pbaran@scripps.edu
8
23 AUGUST 2013 VOL 341 SCIENCE www.sciencemag.org

REPORTS
is superior to organic synthesis in the production leaves of Artemisia annua (21). In addition, en-
In light of the challenge of the in,out sys-
of all complex terpenoids (6, 7). Further, it has gineering ingenol biosynthesis faces considerable tem in 1, inspiration was sought from the bio-
even been claimed that metabolic engineering is challenges because the biosynthetic pathway is synthesis of ingenane terpenoids (Fig. 2A).
favored in the production of “natural products, nontrivial and largely unknown [vide infra, and Their biosyntheses begin with the production
particularly active pharmaceutical ingredients (22)]. Indeed, only the production of casbene (4, of geranylgeranyl pyrophosphate (3) and its cy-
(APIs), some of which are too complex to be Fig. 2A) has been expressed in nonnative hosts, clization to casbene (4) by casbene synthase.
chemically synthesized” in a meaningful way (7). with a low titer of 31 mg/l (23). The above con- Beyond 4, the biosynthetic pathway remains spec-
We repudiate these claims and present a case siderations led LEO Pharma, developer of Picato, ulative; however, the presence of a variety of
study in which chemical synthesis provides an to initiate a partnership with our group to design intermediately cyclized natural products suggests
approach to a structurally complex and medic- and implement a total synthesis of ingenol that discrete cyclizations to lathyrane, tigliane, and
inally relevant terpenoid that is both amenable would simultaneously be amenable to the crea- lastly ingenane skeletons (22). To the best of
to analog synthesis and concise enough to finally tion of analogs and, more importantly, short enough our knowledge, enzymes responsible for these
make chemical production possible.
to make commercial synthesis possible.
later transformations for use in bioengineering
A particularly chemically and biologically
The structure and biological activity of ingenol have not been identified. The conversion from
important terpene is ingenol (1, Fig. 1), a plant- (1) have motivated tremendous interest from syn- a tigliane skeleton, such as phorbol (5, Fig. 2B),
derived diterpenoid first isolated by Hecker in thetic organic chemists over the past 30 years (24). to an ingenane skeleton should proceed through
1968 from Euphorbia ingens (8). Its structure During that time, chemists have completed three a 1,2-pinacol rearrangement, as shown with pu-
was elucidated through x-ray crystallography (9), total syntheses (25–27), one formal synthesis (28), tative intermediate 6. Cha has demonstrated this
showing a unique in,out-[4.4.1]bicycloundecane and numerous approaches to the bicyclic core (24). type of rearrangement by leveraging the strain
core, which is considerably more strained than Although these landmark efforts achieved elegant of an epoxide (29) to overcome thermodynamic
the typical out,out configuration (10). In addi- syntheses of 1, those routes required between 37 forces (vide infra). On the other hand, a vinyl-
tion to its intriguing structure, ingenol esters also and 45 steps, precluding their use in accessing ogous retropinacol rearrangement of ingenol (1)
possess important anticancer activity in animal synthetic analogs in meaningful quantities. As such, has also been observed (30), suggesting that
models and anti-HIV activity in vitro (11–13), de- we required a much more concise, direct approach. the strain of the in,out system thermodynamically
riving from their interaction with protein kinase
C (14). In particular, ingenol mebutate [Picato
(LEO Pharma), 2] was recently Food and Drug
Administration–approved as a first-in-class treat-
ment for actinic keratosis (15), a precancerous
skin condition, and has completed phase II clin-
ical trials for the topical treatment of basal cell
carcinomas (16). Although ingenol mebutate (2)
is highly suited for topical applications, it is de-
graded through rapid ester migration and hydrol-
ysis (17). These characteristics preclude systemic
administration in the treatment of a broader range
of cancers and are not easily addressed by avail-
able synthetic modifications of the natural product.
Currently the supply of ingenol mebutate (2)
is limited to direct isolation (the current com-
mercial source) and semisynthesis from ingenol
(
2
17, 18). Direct isolation yields only 1.1 mg of
per kg of E. peplus (19), thus requiring an im-
mense amount of plant material to achieve large-
scale production. The semisynthesis requires
isolation of ingenol, which, although available
in a more abundant yield of 275 mg of 1 per kg
of dried seeds of E. lathryis (20), still presents a
dearth of material. For comparison, the isolation
of artemisinin—where limited supply drives price
instability—provides around 5 g per kg of dried
Fig. 2. Biosynthetic inspiration and retrosynthetic analysis. (A) Biosynthesis and bioengineering
Fig. 1. Ingenol (1) and ingenol mebutate (2). of ingenane diterpenoids. Ang is Z-2-methyl-2-butenoate. (B) Two-phase approach to ingenol.
www.sciencemag.org SCIENCE VOL 341 23 AUGUST 2013
879

REPORTS
favors the reverse reaction. Nevertheless, we viewed of C-20 oxidation in his total synthesis of 1 (27). To
The synthesis begins with the transformation of
the vinylogous pinacol rearrangement, however conclude the oxidase phase, the remaining two al- 9 into a suitable intermediate for the installation of
challenging, as a key structurally simplifying dis- cohols (C-4 and C-5) and the in,out-bicyclic sys- the C-11 methyl group and the coupling with 10
connection, reducing the highly strained, bridged tem in 7 arise from a stereoselective dihydroxylation (Fig. 3). Chlorination of 9with N-chlorosuccinimide
ring system into an angularly fused ring system. of the corresponding olefin and a vinylogous pin- (NCS), followed by ozonolysis, gave chloroketone
A two-phase design consisting of a cyclase phase acol rearrangement, respectively. In the cyclase 11 in 48% yield over two steps (22.5-g scale). The
and an oxidase phase was chosen (31). The cyclase phase, intermediate diene 8 was traced back to the requisite methyl group could be installed through
phase would establish the precursor tigliane carbon commodity chemical (+)-3-carene (9, $10.20/mol), a reductive alkylation; however, the corresponding
skeleton, and the oxidase phase would install the four in which the dimethyl cyclopropane would serve methyl ketone proved configurationally unstable
hydroxyl groups and rearrange the tigliane skeleton as a stereochemistry-controlling element, ethynyl and formed azeotropes with a variety of solvents,
to an ingenane skeleton (Fig. 2B). The retrosynthetic magnesium bromide, and aldehyde 10. Our hy- complicating its isolation. It was found that a one-pot
plan thus begins with the strategic removal of oxi- pothesis was that this route would provide rapid, procedure for reductive alkylation and aldol reac-
dized functionality in the oxidase phase. The alco- scalable access to diene 8, which would serve as a tion provided superior isolated yield and eliminated
hols at C-3 and C-20 were removed in anticipation key intermediate for both efficient preparation of handling of the challenging methyl ketone. Thus,
that allylic oxidations would enable their installation ingenol and divergent access to ingenol analogs in 11 was treated with lithium naphthalenide followed
(32). Indeed, Wood has demonstrated the feasibility sufficient quantity for drug development.
by hexamethylphosphoramide (HMPA) and methyl
Fig. 3. Reaction sequence for the total synthesis of ingenol (1). Re- 12 hours, 72%. Reaction 7. Methylmagnesium bromide (4.1 equiv.), THF, –78°
agents and conditions are as follows: Reaction 1. NCS [3.0 equivalents (equiv.)], to 0°C, 30 min, 80% + 18% 15. Reaction 8. OsO (1.5 equiv.), pyridine, r.t.,
N,N-dimethylaminopyridine (DMAP) (0.1 equiv.), CH Cl , room temperature 12 hours; 1:1 aqueous Na SO /THF, r.t., 12 hours. Reaction 9. CDI (5.0 equiv.),
r.t.), 3 hours. Reaction 2. O , CH Cl /MeOH; thiourea (1.6 equiv.), 2 hours, 48% DMAP (0.1 equiv.), CH Cl , r.t., 8 hours, 64% over 2 steps. Reaction 10.
4
2
2
2
3
(
3
2
2
2 2
over two steps. Reaction 3. Lithium naphthalenide (~2 equiv.), MeI (10 equiv.), BF •Et O (10.0 equiv.), CH Cl , –78° to –40°C, 30 min, then Et N/MeOH, 80%.
3
2
2
2
3
HMPA/tetrahydrofuran (THF), –78°C to r.t., 1.5 hours; LiHMDS (1.25 equiv.), Reaction 11. SeO
0 (2.0 equiv.), –78°C, 4 hours, 44%. Reaction 4. Ethynylmagnesium bromide DMAP (0.1 equiv.), pyridine (40.0 equiv.), r.t., 1 hour, 59%. Reaction 12. HF
5.0 equiv.), THF, –78° to –15°C, 3 hours, 81%. Reaction 5. TBSOTf (2.0 equiv.), (60.0 equiv.), CH CN, 50°C, 10 hours, 90%. Reaction 13. Martin’s sulfurane
2 2
(5.0 equiv.), dioxane, 80°C, 14 hours; Ac O (20.0 equiv.),
1
(
Et
3
3
N (4.0 equiv.), CH
2
Cl
2
, 0°C, 30 min; TMSOTf (2.0 equiv.), Et
3
N (4.0 equiv.), (6.0 equiv.), CHCl
3
, 80°C, 3 hours, then NaOH, THF, 1 hour, 81%. Reaction 14.
0
°C, 1.5 hours, 71%. Reaction 6. [RhCl(CO)
2
]
2
(0.1 equiv.), p-xylene, 140°C, SeO
2
(10.0 equiv.), 2:1 dioxane/formic acid, 80°C, 1.25 hours, then NaOH, 76%.
880
23 AUGUST 2013 VOL 341 SCIENCE www.sciencemag.org

REPORTS
iodide. The excess methyl iodide was removed by mally oxidized core structure and using powerful elimination and global deprotection were achieved
distillation, and subsequent addition of lithium bis carbon–carbon bond–forming methodologies. by treatment with Martin’s sulfurane, followed
trimethylsilyl)amide (LiHMDS) and 10—derived The oxidase phase began with the dihydrox- by basic hydrolysis, to provide crystalline 20-
(
from 2-iodoxybenzoic acid (IBX) oxidation of ylation of 8. Unfortunately, catalytic dihydrox- deoxyingenol (20) in 80% yield (23-mg scale).
the corresponding alcohol (33)—delivered aldol ylation was ineffective on 8, and only starting Installation of the final C-20 alcohol was ac-
product 12 as a single diastereomer in 44% yield material or overoxidation product was obtained complished by using Shibuya’s conditions for
(
6.4-g scale).
Treatment of 12 with ethynyl magnesium ation of the desired olefin could be realized through conversion, without overoxidation, than Wood’s
bromide provided diol 13 in 81% yield as a 10:1 the use of stoichiometric OsO , followed by re- conditions (27)—to deliver 1 in 76% yield (10-mg
under numerous conditions used. Dihydroxyl- allylic oxidation (35)—which produced higher
4
mixture of diastereomers (3.0-g scale). This ini- ductive hydrolysis, to deliver the requisite diol 16 scale). Synthetic 1 and 20 were spectroscopically
tial four-step sequence rapidly assembles 13, the as a single stereoisomer. The crude diol 16 was identical to naturally obtained material. Thus, the
substrate for the key Pauson-Khand cyclization, by protected with N,N-carbonyldiimidazole (CDI) to oxidase phase consists of seven steps, which in-
setting four contiguous stereocenters and forming deliver carbonate 17 in 68% yield over two steps stall the congested tetraol moiety and use a bio-
three carbon–carbon bonds. Under known condi- (100-mg scale). With 17 in hand, conditions were mimetic rearrangement to generate the all-carbon
tions (34) for the Pauson-Khand cyclization, 13 developed for the pivotal vinylogous pinacol rear- quaternary stereocenter and in,out stereochem-
decomposed rapidly. The alcohols were identi- rangement. Initial results were disappointing be- istry. The use of late-stage oxidations to introduce
fied as problematic functionalities and were cause 17 was unreactive or underwent elimination the four alcohols, particularly those that involve
protected by sequential, one-pot treatment with under conditions used. However, it was discovered functionalization of activated C–H bonds, greatly
tert-butyldimethylsilyl triflate (TBSOTf) and tri- that reaction temperature plays a crucial role. The increases the step economy of the synthesis (36).
methylsilyl triflate (TMSOTf) to deliver 14 in 72% desired pinacol rearrangement could ultimately
yield (2.7-g scale). Unexpectedly, 14 crystallized be effected by treating 17 with BF •Et O (where 1.2% overall yield (73% average per step) from
upon cooling, and x-ray crystallographic analysis Et indicates an ethyl group) in CH Cl at –78° to 9. This compares favorably with the 1.1 mg/kg
unambiguously confirmed the desired stereochemistry. –40°C, followed by quenching with Et N and [0.0011% weight/weight (w/w)] isolation yield
A synthesis of 1 was achieved in 14 steps and
3
2
2
2
3
Aided by the pioneering studies of Brummond and MeOH (Me, methyl), to set the strained in,out of 2 (18) and the 275 mg/kg (0.028% w/w) iso-
others on allenic Pauson-Khand reactions (34), stereochemistry and produce ingenane 7 in 80% lation yield of 1 (19). Salient features of the
cyclization of 14 was effected by treatment with 10 yield (191-mg scale). These results, in light of current route include (i) a four-step sequence
mole percent (mol %) [RhCl(CO) ] under a CO previous reports of the retropinacol rearrangement transforming 9 into stereochemically complex
2
2
atmosphere to provide dienone 15 in 72% yield (30) of ingenol (1), suggest competing kinetic and 13, (ii) a catalytic allenic Pauson-Khand reac-
1.5-g scale). The use of degassed and anhydrous thermodynamic factors. Efforts to understand the tion producing the core carbon skeleton, (iii) a
solvent under high dilution (0.005 M) conditions dynamics of this reaction and its implications for seemingly improbable (30) vinylogous pinacol
(
proved essential to the high yield of this reaction.
Installation of the C-2 methyl group was
ingenol (1) biosynthesis are ongoing.
rearrangement setting the requisite in,out stereo-
The C-3 alcohol was installed by allylic oxi- chemistry, and (iv) sequential, chemoselective
achieved by treatment of 15 with methylmag- dation with SeO , followed by in situ protection oxidation reactions installing the alcohol tetrad.
2
nesium bromide to give 8, the cyclase phase as the acetate, to deliver 18 as a single diaster- A notable feature of the synthesis is excellent
end point, in 80% yield with an additional 18% eomer in 59% yield (128-mg scale). Removal redox economy, evidenced by the absence of
recovered 15 (1.3-g scale). By using the de- of the TBS group with HF gave alcohol 19 in nonstrategic redox manipulations (37). Addi-
scribed route, 1.08 g of 8 was prepared in a sin- 90% yield (82-mg scale). An x-ray crystal struc- tionally, the approach provides gram quantities
gle sequence. The cyclase phase proceeded in ture, obtained by conversion of 19 to the para- of key intermediate 8, which serves as a point
only seven steps from 9 to generate five stereo- bromobenzoate 21, unambiguously confirmed of divergence for the synthesis of analogs of 2
centers and form seven carbon–carbon bonds, the stereochemistry of both the alcohol triad and (vide infra). Such studies are ongoing, and the
illustrating the strategic value of targeting a mini- the in,out bicyclic system (Fig. 4A). Alcohol current synthesis is being scaled up in conjunc-
tion with LEO Pharma to produce large quanti-
ties of intermediate 8 for use in analog preparation.
This synthesis of 1 is not without limitations.
In particular, the oxidase phase relies on the use of
stoichiometric quantities of toxic oxidants (OsO
4
and SeO ) to install the alcohol functionality. A first
2
step to address these issues is a catalytic dihydrox-
ylation of dienone 15 (Fig. 4B), which uses 5 mol %
4
OsO in the presence of citric acid, to deliver diol
22 in 88% yield (180-mg scale). The use of these
conditions on the dihydroxylation of 8 gave decom-
position, however, and retailoring the synthesis to
use 22 requires a change in protecting group strat-
egy. These challenges point to limitations in existing
chemical methods for oxidation (31), and addi-
tional efforts to address the scalability of the oxi-
dase phase are ongoing.
The oxidase phase shown is designed to yield
1; however, by using the predictable oxidase re-
activity exemplified above, intermediate 8 can
be diverted to a variety of ingenanes of varying
Fig. 4. Stereochemical confirmation and preliminary catalytic dihydroxylation. (A) X-ray oxidation and substitution patterns to address
crystal structure of 21. Bz* is p-bromobenzoate. (B) Catalytic dihydroxylation of 15. OsO
N-methylmorpholine N-oxide (NMO) (2.0 equiv.), citric acid (2.0 equiv.), 1:1 t-BuOH/H O, r.t., 24 hours, 88%. of 2. By altering the oxidase phase, we can
4
(0.05 equiv.), the pharmacological challenges to oral dosing
2
www.sciencemag.org SCIENCE VOL 341 23 AUGUST 2013
881

REPORTS
produce analogs of 2 differing in the alcohol
9. K. Zechmeister et al., Tetrahedron Lett. 11, 4075–4078 (1970).
0. R. W. Alder, S. P. East, Chem. Rev. 96, 2097–2112 (1996).
1. S. M. Ogbourne et al., Cancer Res. 64, 2833–2839 (2004).
2. M. Fujiwara et al., Antimicrob. Agents Chemother. 40,
271–273 (1996).
31. K. Chen, P. S. Baran, Nature 459, 824–828 (2009).
32. A. Nakamura, M. Nakada, Synthesis 45, 1421–1451 (2013).
33. T. Konegawa, Y. Ohtsuka, H. Ikeda, T. Sugai, H. Ohta,
Synlett 1997, 1297–1299 (1997).
34. K. M. Brummond et al., Org. Lett. 4, 1931–1934 (2002).
35. K. Shibuya, Synth. Commun. 24, 2923–2941 (1994).
36. P. A. Wender, V. A. Verma, T. J. Paxton, T. H. Pillow,
Acc. Chem. Res. 41, 40–49 (2008).
1
1
1
functionalities (C-3, C-4, C-5, and C-20) to better
understand the structure-activity relationship of
ingenanes.
The synthesis of ingenol (1) presented herein
illustrates the power of two-phase logic to deliver
an efficient, concise synthesis even in architec-
turally complex settings. The usefulness of the
two-phase approach will undoubtedly continue
to expand as new methods for C–C bond for-
mation and C–H oxidation are developed. Fur-
thermore, this report provides a strong rebuttal
to the presumption that chemical synthesis is ill-
equipped to deal with the preparation of struc-
turally complex terpenoid drug molecules. Rather,
in this instance, total chemical synthesis holds
promise as the best method to both prepare ingenol
mebutate (2) and enable the development of
therapeutic analogs with broader utility in the
treatment of human diseases.
13. A. Vasas, D. Rédei, D. Csupor, J. Molnár, J. Hohmann,
Eur. J. Org. Chem. 2012, 5115–5130 (2012).
1
4. C. M. Hasler, G. Acs, P. M. Blumberg, Cancer Res. 52,
02–208 (1992).
2
37. N. Z. Burns, P. S. Baran, R. W. Hoffmann, Angew. Chem.
1
5. U.S. Food and Drug Administration, 2012 Novel New
Drugs Summary (FDA, Washington, DC, 2013); www.fda.
gov/downloads/Drugs/DevelopmentApprovalProcess/
DrugInnovation/UCM337830.pdf.
Int. Ed. 48, 2854–2867 (2009).
Acknowledgments: Financial support for this work was
provided by LEO Pharma. The Carlsberg Foundation and the
Danish Council for Independent 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
A. Rheingold and C. E. Moore (University of California, San
Diego) for x-ray crystallographic analysis and G. Siuzdak
(the Scripps Research Institute) for mass spectrometry
assistance. We thank H. Bladh (LEO Pharma) for valuable
discussions and LEO Pharma for generous donations of
natural 20-deoxyingenol and ingenol. Metrical parameters
for the structures of compounds 14, 20, and 21 are available
free of charge from the Cambridge Crystallographic Data
Centre under reference numbers CCDC-943074, 943076,
and 943075, respectively. A provisional patent application
has been filed and is available under patent application
no. U.S. 61/829,861.
1
1
6. G. Siller et al., Australas. J. Dermatol. 51, 99–105 (2010).
7. “European public assessment report: Picato” [European
Medicines Agency (EMA) publication 650464, London,
2
012]; www.ema.europa.eu/docs/en_GB/document_
library/EPAR_-_Public_assessment_report/human/
02275/WC500135329.pdf.
0
18. X. Liang, G. Grue-Sørensen, A. K. Petersen, T. Högberg,
Synlett 23, 2647–2652 (2012).
19. J. Hohmann, F. Evanics, L. Berta, T. Bartók, Planta Med.
6, 291–294 (2000).
6
20. G. Appendino, G. C. Tron, G. Cravotto, G. Palmisano,
J. Jakupovic, J. Nat. Prod. 62, 76–79 (1999).
21. V. Hale, J. D. Keasling, N. Renninger, T. T. Diagana,
Am. J. Trop. Med. Hyg. 77 (suppl.), 198–202 (2007).
References and Notes
. F. E. Koehn, G. T. Carter, Nat. Rev. Drug Discov. 4,
1
22. R. J. Schmidt, J. Linn. Soc. Bot. 94, 221–230 (1987).
23. J. Kirby et al., Phytochemistry 71, 1466–1473 (2010).
206–220 (2005).
2
3
. I. Raskin et al., Trends Biotechnol. 20, 522–531 (2002).
. P. G. Mountford, in Green Chemistry in the
Pharmaceutical Industry, P. J. Dunn, A. S. Wells,
M. T. Williams, Eds. (Wiley, Weinheim, Germany, 2010),
pp. 145–160.
2
2
4. I. Kuwajima, K. Tanino, Chem. Rev. 105, 4661–4670 (2005).
5. J. D. Winkler, M. B. Rouse, M. F. Greaney, S. J. Harrison,
Y. T. Jeon, J. Am. Chem. Soc. 124, 9726–9728 (2002).
6. K. Tanino et al., J. Am. Chem. Soc. 125, 1498–1500 (2003).
7. A. Nickel et al., J. Am. Chem. Soc. 126, 16300–16301
Supplementary Materials
www.sciencemag.org/cgi/content/full/science.1241606/DC1
Materials and Methods
Supplementary Text
Figs. S1 to S4
Tables S1 to S7
Data
2
2
4. C. J. Paddon et al., Nature 496, 528–532 (2013).
(2004).
5. C. Zhu, S. P. Cook, J. Am. Chem. Soc. 134, 13577–13579
2
2
8. K. Watanabe et al., J. Org. Chem. 69, 7802–7808 (2004).
9. O. L. Epstein, J. K. Cha, Angew. Chem. Int. Ed. 44,
(
2012).
. J. D. Keasling, A. Mendoza, P. S. Baran, Nature 492,
88–189 (2012).
6
121–123 (2004).
7 June 2013; accepted 18 July 2013
Published online 1 August 2013;
10.1126/science.1241606
1
7
. J. D. Keasling, Science 330, 1355–1358 (2010).
30. G. Appendino et al., Eur. J. Org. Chem. 1999,
3413–3420 (1999).
8. E. Hecker, Cancer Res. 28, 2338–2349 (1968).
neutron, and electron) and remains unexamined
with other characterization methods. We show that
multidimensional solid-state nuclear magnetic res-
onance (SSNMR) combined with molecular simu-
lations can be used to generate three-dimensional
(3D) maps of the apportionment of functional groups
(Fig. 1) within and between the pores of MTV-
MOF-5. These MTV-MOFs adopt MOF-5 structure
containing BDC (1,4 benzenedicarboxylate) linkers
with different functional groups–B (BDC-NH2), E
Mapping of Functional Groups in
Metal-Organic Frameworks
1,2
3,4
1,4
4
1,4
Xueqian Kong, * Hexiang Deng, *† Fangyong Yan, * Jihan Kim, ‡ Joseph A. Swisher,
1,3,4
3,4,5
1,2
Berend Smit,
Omar M. Yaghi,
§ Jeffrey A. Reimer §
We determined the heterogeneous mesoscale spatial apportionment of functional groups in a series
of multivariate metal-organic frameworks (MTV-MOF-5) containing BDC (1,4-benzenedicarboxylate)
linkers with different functional groups—B (BDC-NH
2
), E (BDC-NO
2
), F [(BDC-(CH
3
)
2
], H [BDC-(OC
3
H
5
)
2
], (BDC-NO2), F [(BDC-(CH3)2], H [BDC-(OC3H5)2],
and I [BDC-(OC ]—using solid-state nuclear magnetic resonance measurements combined
7
H
7
)
2
and I [BDC-(OC7H7)2]. This method can be ap-
with molecular simulations. Our analysis reveals that these methods discern between random (EF), plied to other ordered or disordered systems to
alternating (EI and EHI), and various cluster (BF) forms of functional group apportionments.
This combined synthetic, characterization, and computational approach predicts the adsorptive
1
properties of crystalline MTV-MOF systems. This methodology, developed in the context of
ordered frameworks, is a first step in resolving the more general problem of spatial disorder in
other ordered materials, including mesoporous materials, functionalized polymers, and defect
distributions within crystalline solids.
Department of Chemical and Biomolecular Engineering, Uni-
2
versity of California, Berkeley, CA 94720, USA. Environmental
Energy Technologies Division, Lawrence Berkeley National Lab-
oratory, Berkeley, CA 94720, USA. Department of Chemistry,
University of California, Berkeley, CA 94720, USA. Materials
3
4
Sciences Division, Lawrence Berkeley National Laboratory,
5
Berkeley, CA 94720, USA. NanoCentury KAIST Institute and
strategy for optimizing the properties of interior (2). This heterogeneity can enhanced se-
Graduate School of Energy, Environment, Water, and Sustain-
ability (World Class University), Daejeon 305-701, Republic of
Korea.
synthetic crystalline materials is to delib- lectivity for carbon dioxide (CO
erately introduce heterogeneity by increas- hydrogen uptake relative to mixtures of pure ma-
ing the number of constituents (1). This approach terials. On a fundamental level, these materials
2
) capture and
A
*
These authors contributed equally to this work.
can be compromised by the dual challenges of present characterization challenges, in that there †Present address: College of Chemistry and Molecular Sci-
phase separation, lack of order, or both. Multi- are no experimental techniques that can elucidate ences, Wuhan University Luojiashan, Wuhan, 430072 China.
‡
Present address: Department of Chemical and Biomolecular
Engineering, Korea Advanced Institute of Science and Tech-
nology, 291 Daehak-ro Yuseong-gu, Daejeon, 305-701 Korea.
variate metal-organic frameworks (MTV-MOFs) the intermingling of functional groups that char-
can meet this challenge by incorporating multiple acterizes the heterogeneity within the crystalline
linkers that bear different functional groups into MTV-MOFs. For example, this “heterogeneity prob-
§
Corresponding author. E-mail: yaghi@berkeley.edu (O.M.Y.);
the same crystal, which creates a heterogeneous lem” is intractable for diffraction methods (x-ray, reimer@berkeley.edu (J.A.R.)
882 23 AUGUST 2013 VOL 341 SCIENCE www.sciencemag.org