Total synthesis of the antibiotic BE-43472B

Article abstract of DOI:10.1002/anie.201301591

Total control: The antibiotic BE-43472B with a unique bisanthraquinone structure has been synthesized in a completely stereocontrolled manner. The key steps are 1) a pinacol rearrangement to install the angular naphthyl group, 2) a diastereoselective methylation of a lactol derivative, and 3) the late-stage installation of the labile hydroxy group through an epoxide. Copyright

Full text of DOI:10.1002/anie.201301591

Angewandte
Chemie
Communications
Natural Product Synthesis
Total control: The antibiotic BE-43472B
with a unique bisanthraquinone structure
has been synthesized in a completely
stereocontrolled manner. The key steps
are 1) a pinacol rearrangement to install
the angular naphthyl group, 2) a diaste-
reoselective methylation of a lactol deriv-
ative, and 3) the late-stage installation of
the labile hydroxy group through an
epoxide.
Y. Yamashita, Y. Hirano, A. Takada,
H. Takikawa, K. Suzuki*
&&&& — &&&&
Total Synthesis of the Antibiotic
BE-43472B
Angew. Chem. Int. Ed. 2013, 52, 1 – 5
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1
These are not the final page numbers!

.
Angewandte
Communications
DOI: 10.1002/anie.201301591
Natural Product Synthesis
Total Synthesis of the Antibiotic BE-43472B**
Yu Yamashita, Yoichi Hirano, Akiomi Takada, Hiroshi Takikawa, and Keisuke Suzuki*
Dedicated to Professor Shigeo Katsumura
The antitumor antibiotic BE-43472B (1) was first isolated in
1996 by researchers at Banyu from Streptomyces sp. A43472
(Figure 1),[1] and recently rediscovered by Rowley and co-
framework, and 3) stereocontrolled installation of the C3
hydroxy group.
The last issue seemed particularly challenging given the
pronounced lability of 1, which underwent elimination to give
the dehydrated product 2 even when kept standing in
[D6]DMSO.[2a] Thus, we envisioned that the C3 hydroxy
group should be introduced at a late synthetic stage.
Scheme 1 outlines the retrosynthetic analysis of 1: The
ꢁ
immediate precursor 3 was envisaged, having the C2 C3
alkene for the late-stage installation of the C3 hydroxy group,
and an isoxazole as a 1,3-diketone surrogate.[6] The next
precursor 4 was planned by assuming the construction of the
anthraquinone moiety by the regioselective Diels–Alder
reaction of the chloroquinone moiety in 4 with siloxydiene
5, and installation of the 13-methyl group on the tetrahydro-
Figure 1. Structures of BE-43472B (1) and its dehydration artifact 2.
workers from a Caribbean ascidian (Ecteinascidia turbi-
nata).[2] In addition to antitumor activity, the latter study
uncovered a significant bactericidal activity of 1 against drug-
resistant pathogens, such as MSSA, MRSA, and VRE. The
basic structure of 1 is composed of two nonidentical
anthraquinone moieties connected by a highly hindered
carbon–carbon bond. Also notable is the stereochemical
complexity associated with the five contiguous stereogenic
centers. Important biological activities as well as the uniquely
complex molecular architecture of 1 have stimulated consid-
erable synthetic interests, and Nicolaou and co-workers
reported the first total synthesis of 1, which was accomplished
in 21 steps in 1.0% yield by exploiting a cascade sequence
including a Diels–Alder reaction.[3]
Herein, we describe the second total synthesis of
(ꢀ)-1 through an isoxazole-based polyketide assembly strat-
egy[4] and a 1,2-rearrangement to install the angular substitu-
ent.[5] The three major challenges in planning a synthetic
ꢁ
ꢁ
route were: 1) C C bond (C4a C7’) formation at the highly
hindered angular position, 2) construction of the octacyclic
[*] Y. Yamashita, Y. Hirano, Dr. A. Takada, Dr. H. Takikawa,
Prof. Dr. K. Suzuki
Department of Chemistry, Tokyo Institute of Technology
2-12-1 O-okayama, Meguro-ku, Tokyo 152-8551 (Japan)
E-mail: ksuzuki@chem.titech.ac.jp
[**] We are grateful to Banyu Pharmaceutical Co., Ltd. for kindly
providing an authentic sample of 1. We thank Sachiyo Kubo for X-ray
analysis. This work was supported by a Grant-in-Aid for Specially
Promoted Research (No. 23000006) from Japan Society for the
Promotion of Science (JSPS) (Japan). Y.Y. thanks the JSPS for
a Research Fellowship for Young Scientists.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201301591.
Scheme 1. Retrosynthetic analysis of BE-43472B (1).
2
www.angewandte.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 1 – 5
These are not the final page numbers!

Angewandte
Chemie
furan ring by exploiting the g-lactone moiety in 4. Further
disconnection at the central spiroacetal suggested keto ester 6
as the key intermediate, which would allow installation of the
angular naphthyl group through a isoxazole-directed regio-
selective pinacol-type 1,2-shift[5a] of diol 7. Final disconnec-
tion of 7 suggested a-ketol 8 and naphthyl bromide 9 as the
starting materials.
Our synthesis started with the generation of a naphthyl-
lithium species from bromide 11 (tBuLi, THF, ꢁ788C), which
was treated with ketone 10[4e] to give the single cis-diol 12 in
quantitative yield (Scheme 2).[7] The key pinacol 1,2-shift of
diol 12 was nicely promoted by TfOH (25 mol%), along with
concomitant removal of the MOM group to give the desired
ketone 13 in 96% yield. This transformation proceeded in
a regio- and stereospecific manner owing to the excellent
cation-stabilizing ability of the isoxazole moiety in 12.[5] Upon
heating ketone 13 in toluene in the presence of CSA (1108C,
1 h), lactone 14 was obtained in 98% yield via the hemiacetal
intermediate followed by ester exchange. For the installation
of the 13-methyl group, lactone 14 was reduced (DIBAL-H,
CH2Cl2, ꢁ1008C) and acetylated to give acetate 15 in 84%
combined yield (b/a = 11:1). The completely stereoselective
methylation[8] of 15 was achieved under carefully optimized
conditions with three key requirements: 1) presence of
water,[9] 2) use of TMSOTf as a Lewis acid, and 3) the precise
ratio of the reagents. Thus, treatment of acetate 15 with
Me3Al (7.5 equiv), H2O (4.5 equiv), and TMSOTf (10 equiv)
afforded 16 with complete diastereoselectivity in 92%
yield.[10] The stereostructure of 16 was unambiguously deter-
mined by X-ray crystallography.[11]
After having constructed the A–G rings in a fully
stereocontrolled manner, further conversion into anthraqui-
none 20 was undertaken (Scheme 3). In preparation for the
Scheme 3. Synthesis of anthraquinone 20. a) nBuLi, THF, DMPU,
ꢁ100!ꢁ908C, 30 min; PhSeBr, ꢁ908C, 10 min (38% for 17, 38% for
18); b) NaIO4, NaHCO3, THF, H2O, RT, 14 h (quant. from 17); H2O2,
NaHCO3, THF, H2O, RT, 14 h (85% from 18); c) CAN-SiO2, CH2Cl2,
H2O, RT, 10 min (88%); d) 5, benzene, CH2Cl2, 08C!RT, 3.5 h; SiO2,
benzene, CH2Cl2, RT, 2 h; K2CO3, MeOH, CH2Cl2, 08C, 20 min (94%).
DMPU=N,N’-dimethylpropylene urea, CAN=ceric ammonium nitrate.
ꢁ
introduction of the C2 C3 double bond, the lithiation at the
C2 position of isoxazole 16 was examined, and a suitable set of
conditions was found: nBuLi in a solvent mixture (THF,
DMPU, v/v = 5:1) at ꢁ908C for 30 min. Among methods
ꢁ
tested for introducing the C2 C3 double bond in 16, the
selenoxide-based protocol[12] proved effective; lithiation of
isoxazole 16 (see above) and subsequent treatment with
PhSeBr gave a 1:1 separable mixture of diastereomeric
selenides 17 and 18 in 76% combined yield. Upon oxidative
treatment, selenides 17 and 18 nicely converged into olefin 19
in high yield.[13] Oxidation of naphthalene 19 with CAN-
Scheme 2. Synthesis of octacycle 16. a) 11, tBuLi, THF, ꢁ788C,
10 min, then 10, ꢁ78!ꢁ408C, 20 min (quant., based on 10);
b) TfOH (25 mol%), CH2Cl2, RT, 13 h (96%); c) (ꢀ)-10-camphorsul-
fonic acid (CSA), toluene, reflux, 1 h (98%); d) DIBAL-H, CH2Cl2,
ꢁ1008C, 20 min; e) Ac2O, pyridine, DMAP, CH2Cl2, RT, 3 h (84%, 2
steps, d.r.=11:1); f) Me3Al, H2O, TMSOTf (see text), CH2Cl2,
ꢁ788C!RT, 12 h (92%). MOM=methoxymethyl, Tf=trifluorometh-
anesulfonyl, DIBAL-H=diisobutylaluminum hydride, DMAP=4-N,N-
dimethylaminopyridine, TMS=trimethylsilyl.
[14]
SiO2 afforded the corresponding naphthoquinone in 88%
yield, which was subjected to a Diels–Alder reaction with
siloxydiene 5[15] to give, after successive treatment with SiO2
and K2CO3, anthraquinone 20 in 94% yield.
Angew. Chem. Int. Ed. 2013, 52, 1 – 5
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3
These are not the final page numbers!

.
Angewandte
Communications
Having constructed the full carbon framework, the
remaining tasks toward the total synthesis were: 1) unmask-
ing of the isoxazole to give the 1,3-diketone, 2) deprotection,
and 3) introduction of the labile C3 hydroxy group. Extensive
model studies, however, suggested that only a narrow window
was open to the target owing to many limiting factors.
The order of these transformations was crucial, and the
third task proved extremely difficult; 1) no feasible conditions
were found for hydrating the olefin once the 1,3-diketo
functionality was unmasked, because the double bond was too
unreactive (Scheme 4a). 2) In the case of olefin 20 with the
isoxazole, even if the hydration was successful, the labile C3
hydroxy group would not survive the conditions for convert-
ing the isoxazole into the corresponding 1,3-diketone.
Scheme 5. Endgame. a) BBr3, CH2Cl2, ꢁ78!ꢁ108C, 3 h; b) Boc2O,
DMAP, CH2Cl2, ꢁ208C, 20 min (82%, 2 steps); c) N-bromosaccharin,
HCOOH, CH2Cl2, 08C, 20 min; d) Na2CO3, MeOH, RT, 1.5 h (83%, 2
steps); e) CF3CO2H, CH2Cl2, RT, 2 h; f) [Mo(CO)6], MeCN, H2O, 858C,
40 min (97%, 2 steps); g) benzyl bromide, Ag2O, DMF, RT, 1 h (93%);
h) tBuONO, TfOH, DMSO, RT, 40 min (61%); i) PhSeSePh, NaBH4,
AcOH, EtOH, THF, 08C, 20 min (82%); j) BBr3, CH2Cl2, ꢁ78!ꢁ558C,
1 h (76%). Boc=tert-butoxycarbonyl, DMAP=4-N,N-dimethylamino-
pyridine, Tf=trifluoromethanesulfonyl, DMSO=dimethyl sulfoxide.
Scheme 4. Problems and solution for installing the C3-hydroxy group.
The latter notion led us to focus on the C2–C3 epoxide as
a masked form of the C3 hydroxy group until the final
synthetic stage. However, the model epoxidation experiment
gave only the wrong stereoisomer, the b-epoxide
(Scheme 4b); this outcome gave insight into the molecular
ꢁ
geometry in the vicinity of the C2 C3 double bond. The b-
face is available for the approaching reagents (Scheme 4c).
We focused on the epoxide formation via a halohydrin
intermediate, expecting that the halonium ion formation from
the b-side followed by capture by an oxygen nucleophile
would give a halohydrin derivative that would be convertible
to the a-epoxide (Scheme 4d).
bromosaccharin[17] in formic acid at 08C, two products
formed. Careful structure assignment identified these prod-
ucts to be isomeric bromoformates 22 and 23 in a ratio of
1.8:1. Preferential formation of the anti-Markovnikov prod-
uct 22 was unexpected, but rationalized by the extreme steric
hindrance of the a-side at the C3 position (inside of the
concave pocket), thereby forcing the oxygen nucleophile to
attack the C2 position, albeit electronically unfavorable. This
unexpected result, however, caused no problem for our
purpose, as both regioisomers converged to the desired a-
epoxide 24 as a sole product in 83% yield (two steps) by
treatment with Na2CO3 (MeOH, RT).
Two additional issues arose: 1) The double bond next to
the isoxazole was also unreactive, and 2) the strong halogen-
ating reagents, thus needed, also halogenated other sites of
the aromatic rings. Therefore, Boc groups were employed for
the protection of two phenols to avoid halogenation on the A
and H rings (see 20!21, Scheme 5).
For enhancing the reactivity of the halogenating agent, we
focused on the acid activation,[16] and formic acid gave an
excellent result. Upon treatment of olefin 21 with N-
After the two Boc protecting groups in 24 were removed,
the isoxazole was cleaved with [Mo(CO)6][18] to give enami-
4
www.angewandte.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 1 – 5
These are not the final page numbers!

Angewandte
Chemie
[3] a) K. C. Nicolaou, Y. H. Lim, J. Becker, Angew. Chem. 2009, 121,
none 25. Our previous experience suggested that an enami-
none moiety in such a polycyclic construct would be resistant
to usual acid hydrolysis, for which a protocol through
diazotization was developed.[19,20] After the selective protec-
tion of phenol 25 with a benzyl group, the enaminone was
hydrolyzed by using tBuONO and TfOH to give the desired
diketone 26 in 61% yield. Epoxide 26 was subjected to the
reductive ring opening under Miyashita conditions[21] (PhSe-
SePh, NaBH4, EtOH) to give the corresponding tert-alcohol,
which was treated with BBr3 (CH2Cl2, ꢁ78!ꢁ558C). For-
tunately, in contrast to the pronounced lability of the tert-
alcohol, the Lewis acidic conditions lead to removal of the
benzyl protecting group at low temperature to afford the
long-sought target 1. The synthetic material was identical in
all respects (1H-, 13C NMR, IR, HRMS) to the natural sample.
Recrystallization from CDCl3 gave nice crystals (yellow
plates, m.p. 211–2138C), and X-ray analysis gave further
structural proof (Figure 2).[11]
3496 – 3500; Angew. Chem. Int. Ed. 2009, 48, 3444 – 3448;
b) K. C. Nicolaou, J. Becker, Y. H. Lim, A. Lemire, T. Neubauer,
A. Montero, J. Am. Chem. Soc. 2009, 131, 14812 – 14826; c) A. E.
Hayden, R. S. Paton, J. Becker, Y. H. Lim, K. C. Nicolaou, K. N.
Houk, J. Org. Chem. 2010, 75, 922 – 928; d) Highlight: D. C.
Rowley, Nat. Chem. 2009, 1, 110 – 112.
[4] a) J. W. Bode, Y. Hachisu, T. Matsuura, K. Suzuki, Org. Lett.
2003, 5, 391 – 394; b) J. W. Bode, Y. Hachisu, T. Matsuura, K.
Suzuki, Tetrahedron Lett. 2003, 44, 3555 – 3558; c) T. Matsuura,
J. W. Bode, Y. Hachisu, K. Suzuki, Synlett 2003, 1746 – 1748;
d) H. Takikawa, K. Hikita, K. Suzuki, Synlett 2007, 2252 – 2256;
e) Y. Hachisu, J. W. Bode, K. Suzuki, J. Am. Chem. Soc. 2003,
125, 8432 – 8433; f) H. Takikawa, Y. Hachisu, J. W. Bode, K.
Suzuki, Angew. Chem. 2006, 118, 3572 – 3574; Angew. Chem. Int.
Ed. 2006, 45, 3492 – 3494.
[5] a) K. Suzuki, H. Takikawa, Y. Hachisu, J. W. Bode, Angew.
Chem. 2007, 119, 3316 – 3318; Angew. Chem. Int. Ed. 2007, 46,
3252 – 3254; b) H. Takikawa, K. Hikita, K. Suzuki, Angew.
Chem. 2008, 120, 10035 – 10038; Angew. Chem. Int. Ed. 2008, 47,
9887 – 9890.
[6] a) G. Stork, S. Danishefsky, M. Ohashi, J. Am. Chem. Soc. 1967,
89, 5459 – 5460; b) G. Stagno DꢀAlcontres, Gazz. Chim. Ital.
1950, 80, 441 – 455; c) C. Kashima, Heterocycles 1979, 12, 1343 –
1368; d) P. G. Baraldi, A. Barco, S. Benetti, G. P. Pollini, D.
Simoni, Synthesis 1987, 857 – 869; e) A. I. Kotyatkina, V. N.
Zhabinsky, V. A. Khripach, Russ. Chem. Rev. 2001, 70, 641 – 653.
[7] For preparation of ketol 10 and naphthalene 11, see the
Supporting Information.
[8] K. Tomooka, K. Matsuzawa, K. Suzuki, G. Tsuchihashi, Tetra-
hedron Lett. 1987, 28, 6339 – 6342.
[9] M. Miyashita, M. Hoshino, A. Yoshikoshi, J. Org. Chem. 1991,
56, 6483 – 6485.
[10] The a-acetate also gave methylated product 16 in 96% yield
under the same reaction conditions.
[11] CCDC 923753 (10), 923754 (16), and 925606 (1) contain the
supplementary crystallographic data for this paper. These data
can be obtained free of charge from The Cambridge Crystallo-
graphic Data Centre via www.ccdc.cam.ac.uk/data_request/cif.
[12] a) H. J. Reich, I. L. Reich, J. M. Renga, J. Am. Chem. Soc. 1973,
95, 5813 – 5815; b) K. B. Sharpless, R. F. Lauer, A. Y. Teranishi,
J. Am. Chem. Soc. 1973, 95, 6137 – 6139.
[13] Upon treatment with NaIO4, a-selenide 17 underwent gradual
oxidation and elimination to give olefin 19 in quantitative yield.
b-Selenide 18 also gave olefin 19 in 85% yield by using H2O2 and
NaHCO3, which induced in situ epimerization of the selenoxide
owing to the enhanced acidity of C2 by the isoxazole and
selenoxide.
Figure 2. X-ray crystal structure of BE-43472B (1). Hydrogen atoms are
omitted for clarity. O red, C black. Thermal ellipsoids are shown at the
50% probability level.
In summary, total synthesis of (ꢀ)-BE-43472B (1) has
been accomplished in 27 steps and 3.0% yield from a com-
mercially available material. The synthesis features: 1) use of
isoxazole as a 1,3-diketone surrogate, 2) the isoxazole-
directed pinacol 1,2-shift for connecting two anthraquinone
precursors at the angular position, and 3) perfect control of
five contiguous stereogenic centers. The synthetic route
described herein is efficient enough to allow the synthesis of
the chiral, nonracemic ones as well as the various congeners
with potential biological activities. Further work along these
lines is in progress.
[14] M. H. Ali, M. Niedbalski, G. Bohnert, D. Bryant, Synth.
Commun. 2006, 36, 1751 – 1759.
[15] J. Savard, P. Brassard, Tetrahedron 1984, 40, 3455 – 3464.
[16] a) P. Konradsson, D. R. Mootoo, R. E. McDevitt, B. Fraser-Reid,
J. Chem. Soc. Chem. Commun. 1990, 270; b) F. L. Lambert,
W. D. Ellis, R. J. Parry, J. Org. Chem. 1965, 30, 304; c) G. A.
Olah, Q. Wang, G. Sandford, G. K. S. Prakash, J. Org. Chem.
1993, 58, 3194.
Received: February 24, 2013
Revised: April 13, 2013
Published online: && &&, &&&&
[17] D. Urankar, I. Rutar, B. Modec, D. Dolenc, Eur. J. Org. Chem.
2005, 2349 – 2353.
[18] M. Nitta, T. Kobayashi, J. Chem. Soc. Perkin Trans. 1 1985, 1401 –
1406.
[19] G. A. Olah, J. A. Olah, J. Org. Chem. 1965, 30, 2386 – 2387.
[20] A. Takada, Y. Hashimoto, H. Takikawa, K. Hikita, K. Suzuki,
Angew. Chem. 2011, 123, 2345 – 2349; Angew. Chem. Int. Ed.
2011, 50, 2297 – 2301.
[21] a) M. Miyashita, T. Suzuki, A. Yoshikoshi, Tetrahedron Lett.
1987, 28, 4293 – 4296; b) M. Miyashita, T. Suzuki, M. Hoshino, A.
Yoshikoshi, Tetrahedron 1997, 53, 12469 – 12486.
Keywords: antibiotics · bisanthraquinone · isoxazole ·
natural products · total synthesis
.
[1] H. Kushida, S. Nakajima, T. Koyama, H. Suzuki, K. Ojiri, H.
Suda, JP 08143569, 1996.
[2] a) A. M. Socha, D. Garcia, R. Sheffer, D. C. Rowley, J. Nat.
Prod. 2006, 69, 1070 – 1073; b) A. M. Socha, K. L. LaPlante,
D. C. Rowley, Bioorg. Med. Chem. 2006, 14, 8446 – 8454.
Angew. Chem. Int. Ed. 2013, 52, 1 – 5
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5
These are not the final page numbers!