11-step enantioselective synthesis of (-)-lomaiviticin aglycon

Article abstract of DOI:10.1021/ja200034b

Lomaiviticins A and B are complex antitumor antibiotics that were isolated from a strain of Micromonospora. A confluence of several unusual structural features renders the lomaiviticins exceedingly challenging targets for chemical synthesis. We report an

Full text of DOI:10.1021/ja200034b

COMMUNICATION
pubs.acs.org/JACS
11-Step Enantioselective Synthesis of (-)-Lomaiviticin Aglycon
Seth B. Herzon,* Liang Lu,† Christina M. Woo,† and Shivajirao L. Gholap
Department of Chemistry, Yale University, New Haven, Connecticut 06520, United States
S Supporting Information
b
ABSTRACT: Lomaiviticins A and B are complex antitumor
antibiotics that were isolated from a strain of Micromonospora.
A confluence of several unusual structural features renders
the lomaiviticins exceedingly challenging targets for chemical
synthesis. We report an 11-step, enantioselective synthetic
route to lomaiviticin aglycon. Our route proceeds by late-stage,
stereoselective dimerization of two equivalent monomeric
intermediates, a transformation that may share parallels with
the natural products' biosyntheses. The route we describe is
scalable and convergent, and it lays the foundation for deter-
mination of the mode of action of these natural products.
Lomaiviticins A and B (1 and 2, Figure 1) are complex dimeric
bacterial metabolites that were obtained as minor constituents
from the fermentation broth of Micromonospora lomaivitiensis.1
The lomaiviticins, and the related monomeric isolates known as
the kinamycins (4-6),2 constitute a small family of natural products
which contain a diazo functional group. In 1-6, this function is
embedded in a highly oxidized tetracyclic carbon skeleton,
forming a diazotetrahydrobenzo[b]fluorene (diazofluorene, see
structure 7). Simple diazo-containing molecules, such as diazo-
methane, are prone to detonation.3 Thus, the presence of this
functional group within a natural product, itself formed under
conditions of microbial metabolism, is remarkable. Several addi-
tional structural features of the lomaiviticins—deoxyglycoside
residues, highly oxidized napthoquinones, and a sterically con-
gested carbon-carbon bond bridging the two moieties of each
metabolite—elevate their molecular complexity and render their
structures singular among all known natural products.
Both 1 and 2 exhibit powerful antimicrobial activities (MICs ≈
6-25 ng/spot, plate assay), and lomaiviticin A (1) is reported to
inhibit the growth of 25 cancer cell lines at low nanomolar to
picomolar concentrations. The activity profile of 1 differs from those
of other cancer chemotherapeutics, such as doxorubicin and mito-
mycin, suggesting that a distinct mechanism of action is operative.1
Model studies have provided evidence that the lomaiviticins form
reactive intermediates under reducing conditions.4 However, inves-
tigations of 1 and 2 directly have been impossible to initiate because
the lomaiviticins are obtained in minute quantities from their natural
source,1 and a total synthesis has not yet been realized.5,6
Herein, we report an 11-step, enantioselective synthesis of
lomaiviticin aglycon (3). This research addresses the decade-
long problem of synthesis of the dimeric diazofluorene structure
of the lomaiviticins. Many steps in our route may be executed
efficiently on gram scales, as shown. The brevity and convergence
of our synthesis renders it amenable to incorporation of affinity
probes for target identification experiments and to the synthesis
of analogues for structure-function studies.7
Figure 1. Structures of lomaiviticins A and B (1 and 2), lomaiviticin
aglycon (3), kinamycins A, C, and F (4-6), and diazotetrahydrobenzo-
[b]fluorene (diazofluorene, 7).
Our strategy to synthesize 3 was guided by the hypothesis that
nature prepares the lomaiviticins by direct oxidative dimerization
of monomeric diazofluorenes, as depicted in Scheme 1a. The
merit of this approach in a synthetic setting is that it prognos-
ticates the union of two identical monomers late-stage in the
synthesis, thereby vastly simplifying the route to 3. However,
several challenges to this bond construction are readily identified.
First, under strongly acidic or basic conditions, monomers such
as 8 are susceptible to a β-elimination-aromatization pathway
(Scheme 1b), and this reaction manifold severely limits the chem-
ical reagents that are compatible with such substrates. Second,
as a consequence of the steric hindrance about the incipient
bond, it was anticipated that the rate of dimerization may be too
slow to produce workable yields of coupled products (9). This
substitution pattern also made it difficult to predict the stereo-
chemical outcome of the coupling event. Finally, although the
oxidative couplings of ketones,8 their enolates,9 and their enoxy-
silanes10 have been used with great success in synthetic chemistry,
Received: January 3, 2011
Published: January 31, 2011
r
2011 American Chemical Society
7260
dx.doi.org/10.1021/ja200034b J. Am. Chem. Soc. 2011, 133, 7260–7263
|

Journal of the American Chemical Society
COMMUNICATION
Scheme 1. (a) Proposed Oxidative Dimerization To Form Lomaiviticins A and B (1 and 2), (b) Elimination-Aromatization
Pathway, and (c) Syntheses of the Diazofluorenes 18 and 19a
a Conditions: (a) Li, NH3, t-BuOH, THF, -78 f -33 f -78 °C, 98%. (b) (DHQD)2PHAL, K2OsO4, CH3SO2NH2, t-BuOCH3-t-BuOH-H2O,
-5 °C, 62%, 91% ee. (c) Pd(OAc)2, O2, DMSO, 24 °C, 92%. (d) MesCH(OCH3)2, PPTS, CH3CN, 24 f 50 °C, 85%. (e) TMSCH2MgCl, CuI,
HMPA, Et3N, TMSCl, THF, -30 f -60 f -78 °C; Pd(OAc)2, CH3CN, 24 °C, 82%. (f) TASF(Et), CH2Cl2, -78 °C, 81%. (g) Pd(OAc)2, polymer-
supported PPh3, Ag2CO3, toluene, 80 °C, 95%. (h) TfN3, DMAP, CH3CN, -20 °C, 51%.
the oxidative dimerization of diazofluorenes, such as 8, was
unknown at the outset of these studies.
to afford the γ-quinonylation products 16 in 81% yield (1 g scale).
The products 16 were then cyclized by heating in the presence of
palladium acetate and polymer-supported triphenylphosphine. Low-
ering the ratio of phosphine to palladium from 2.5:1 to 1.5:1
increased the yield of the cyclized products 17 to 95% (compared
to 66% yield in our kinamycin synthesis;6c 1.4 g scale). Finally, diazo
transfer to the cyclized products 17 (TfN3, DMAP) afforded the
diazofluorenes 18 and 19 in 51% combined yield, and these were
readily separated by flash column chromatography (550 mg scale).15
Transformation of 18 (or 19) to lomaiviticin aglycon (3)
proved to be a formidable challenge. An exhaustive evaluation of
conditions (>1500 experiments) using 18, 19, and various deriv-
atives of each was unsuccessful. Ultimately, however, a short
sequence was developed (Scheme 2a). First, the exo-mesityl
diazofluorene 18 was converted to its trimethylsilyl enol ether
derivative 20 by exposure to trimethylsilyl trifluoromethanesul-
fonate and triethylamine at 0 °C. The unpurified enoxysilane 20
was immediately re-dissolved in benzene and treated with a
solution of manganese tris(hexafluoroacetylacetonate) (21)16 in
benzene at 21 °C. Under these conditions, the oxidative coupling
products 22 and 23 were obtained in 38% combined yield on a
100 mg scale (26% isolated yield of 22, 12% isolated yield of 23,
30% of 18 recovered).17 A third dimeric product, tentatively
identified as the Cs-symmetric dimer, was detected by LC/MS ana-
lysis of the unpurified product mixture, but this compound was
not formed in quantities sufficient for detailed characterization.
The relative stereochemistry of the dimeric products 22
and 23 was deduced by conversion of 22 to lomaiviticin aglycon
Our pathway to lomaiviticin aglycon (3) began with synthesis
of the monomeric diazofluorene 18 by the sequence outlined in
Scheme 1c. Our route to 18 follows a pathway originally devel-
oped for the synthesis of kinamycin F (6),6c with several critical
modifications to increase flexibility and material throughput. The
synthesis began with silylation of the inexpensive commercial
reagent 3-ethylphenol (96%, >1 mol scale).11 Birch reduction of
the resulting silyl ether 10 followed by regio- and stereoselective
dihydroxylation12 afforded the enoxysilane 11 in 91% ee and 61%
yield over two steps (>60 g scale). In a departure from previous
studies, the diol 11 was oxidized directly to the R,β-unsaturated
ketone 12 by treatment with palladium acetate (10 mol %) under
an atmosphere of dioxygen (92%, 9.6 g scale).13 This modifica-
tion delayed installation of the protecting group, which acceler-
ated the development of this sequence. The R,β-unsaturated
ketone 12 was then protected as its mesitylaldehyde acetal (85%,
1:1 mixture of diastereomers, 4.4 g scale). The mixture of di-
astereomeric acetals 13 was homologated to the β-(trimethyl-
silylmethyl)-R,β-unsaturated ketones 14 by copper-mediated
1,4-addition of trimethylsilylmethylmagnesium chloride, trapping
with chlorotrimethylsilane, and reoxidation (82%, 1.5 g scale).
The β-(trimethylsilylmethyl)-R,β-unsaturated ketones 14 and
2,3-dibromo-5,8-bis(methoxymethyloxy)naphthoquinone (15, pre-
pared in one step from 2,3-dibromo-5,8-dihydroxynaphtho-
quinone)11,14 were efficiently coupled by treatment with tris-
(diethylamino)sulfonium trimethyldifluorosilicate [TASF(Et)],
7261
dx.doi.org/10.1021/ja200034b |J. Am. Chem. Soc. 2011, 133, 7260–7263

Journal of the American Chemical Society
COMMUNICATION
Scheme 2. (a) Dimerization of the exo-Mesityl Diazofluorene 18,a (b) Stereochemical Model for the Dimerization Event,
and (c) Cyclization of the Ketone Isomer (25) of Lomaiviticin Aglycon (3)
a Conditions: (a) TMSOTf, Et3N, CH2Cl2, 0 °C, then 21, PhH, 21 °C, 26% 22, 12% 23, 30% 18. (b) TFA, TBHP, CH2Cl2, -35 °C, 12 h, 39%. (c) TFA,
CH3OH, 50 °C, 41%.
(3, vide infra). The stereoselectivity in their formation (22:23 =
2:1) is believed to originate from a nonbonded interaction between
the ortho-methylarene substituents of 20 and the C-3 ethyl
substituent (Scheme 2b). This interaction is proposed to orient
the ethyl residue over the Si face of the enoxysilane, guiding the
approach of the enoxysilanes 20 from their respective Re faces.
Dimerization of the enoxysilane of 19, which contains the mesityl-
aldehyde acetal in the endo orientation, forms the (1S,10S)-dimer
exclusively (36% yield).11,18 This result is consistent with our
stereochemical model and the inherent preference for addition
exo to the cis-fused 5-6 system.
and HMBC experiments unequivocally established the presence
of 25 [a correlation was observed between H-1/H-10 (δ 3.34)
and C-2/C-20 (δ 190.8)]. The open-chain isomer (25) slowly
converted to lomaiviticin aglycon (3) on standing in chloroform
(ca. 50% conversion of 25 after 12 h at 24 °C) or, more rapidly, in
methanol (full conversion within 10 min). The C-2/C-20 carbon
resonance of 3 (δ 99.4, CDCl3), which was coupled to the
tertiary hydroxyl proton and to H-1/H-10 in the HMBC spectrum,
agrees well with that of lomaiviticin B (2, δ 96.5, CD3OD).1,11
These data support the assignment of the relative stereochem-
istry of 3 and, by inference, that of 22 and 23, as shown.
By comparison, intermediates in the (1S,10S)-series were ap-
preciably more robust. The (1S,10S)-dimer 23 was deprotected
by heating with trifluoroacetic acid in methanol at 50 °C, to
afford the diastereomeric lomaiviticin aglycon (24, 41%). The
increased stability of 24 may be attributed to the cis arrangement
of the R-hydrogen and β-oxygen substituents, which is expected
to decrease the rate of β-elimination. As expected, the (1S, 10S)-
diastereomer 24 showed no evidence of cyclization after standing
for 2 days at 24 °C in methanol-d4.
The deprotection of the (1R,10R)-dimer 22 was successfully
executed under carefully optimized conditions. Thus, slow addition
of trifluoroacetic acid to a mixture of 22 and excess anhydrous
tert-butylhydroperoxide19 in dichloromethane-decane solution
at -35 °C, followed by extractive workup and purification,
afforded lomaiviticin aglycon (3) in 39% yield (10 mg scale,
2.5 mg of 3 isolated; >15 mg of 3 has been prepared to date).
Careful examination of this reaction established that the open-
chain ketone isomer of lomaiviticin aglycon (25) is the product
that is first formed in the deprotection of 22 (Scheme 2c). The
ratio of the ketone isomer 25 to 3 was >8:1 throughout the
course of the reaction (LC/MS analysis, retention times of 25
and 3 = 1.55 and 1.26 min, respectively), but these mixtures
quantitatively converted to 3 upon purification by preparative
thin-layer chromatography. Mixtures of 25 and 3 could be
obtained by rapid isolation using flash-column chromatography,11
Several additional observations that arose during the course of
our studies are noteworthy. First, we found that enoxysilanes
derived from diazofluorenes bearing acyclic protecting groups on
the vicinal diol function were unstable above 0 °C. Attempts to
induce their direct coupling (at lower temperatures) invariably
resulted in elimination and aromatization, without formation of
detectable levels of dimeric products. Additionally, conventional
7262
dx.doi.org/10.1021/ja200034b |J. Am. Chem. Soc. 2011, 133, 7260–7263

Journal of the American Chemical Society
COMMUNICATION
oxidants, such as copper chloride9a or ceric ammonium nitrate
(CAN),10 led to rapid aromatization of the enoxysilane 20, even
at -35 °C. An extensive evaluation of alternative oxidants ulti-
mately led to the identification of manganese tris(hexafluoroacetyl-
acetonate) (21) as uniquely effective in this bond construction.
The synthesis of 21 proceeds in one step from commercial
reagents,16 and the volatility of the complex allows purification of
multigram quantities by sublimation (60-70 °C, 200 mTorr).
The application of 21 in the oxidative coupling of enoxysilanes
has not been described, to our knowledge, although Mayer has
conducted detailed studies of its reactivity toward activated C-H
bonds.16b The success of 21 in our studies may be attributed to
two factors. First, because the manganese atom in 21 is coordi-
natively saturated, it cannot behave as a Lewis acid toward the
β-oxygen of 20. Second, in contrast to copper chloride or CAN,
the manganese complex 21 is fully soluble (and stable) in benzene.
Use of this solvent in the oxidative coupling reaction may attenuate
competing elimination pathways.
1680. (c) Zhang, W.; Baranczak, A.; Sulikowski, G. A. Org. Lett. 2008, 10,
1939. (d) Gholap, S. L.; Woo, C. M.; Ravikumar, P. C.; Herzon, S. B. Org.
Lett. 2009, 11, 4322. (e) Nicolaou, K. C.; Nold, A. L.; Li, H. Angew.
Chem., Int. Ed. 2009, 48, 5860. (f) Morris, W. J.; Shair, M. D. Tetrahedron
Lett. 2010, 51, 4310. (g) Lee, H. G.; Ahn, J. Y.; Lee, A. S.; Shair, M. D.
Chem. -Eur. J. 2010, 16, 13058.
(6) For syntheses of various kinamycins, see: (a) Lei, X.; Porco, J. A.
J. Am. Chem. Soc. 2006, 128, 14790. (b) Nicolaou, K. C.; Li, H.; Nold,
A. L.; Pappo, D.; Lenzen, A. J. Am. Chem. Soc. 2007, 129, 10356.
(c) Woo, C. M.; Lu, L.; Gholap, S. L.; Smith, D. R.; Herzon, S. B. J. Am.
Chem. Soc. 2010, 132, 2540. For synthesis of (()-O-methyl-kinamycin
C, see: (d) Kumamoto, T.; Kitani, Y.; Tsuchiya, H.; Yamaguchi, K.; Seki,
H.; Ishikawa, T. Tetrahedron 2007, 63, 5189. For studies toward the
synthesis of kinamycin F (6), see: (e) Chen, N.; Carriere, M. B.; Laufer,
R. S.; Taylor, N. J.; Dmitrienko, G. I. Org. Lett. 2008, 10, 381.
(7) Wender, P. A.; Verma, V. A.; Paxton, T. J.; Pillow, T. H. Acc.
Chem. Res. 2008, 41, 40.
(8) Hawkins, E. G. E.; Large, R. J. Chem. Soc., Perkin Trans. 1974, 1,
280.
(9) For selected examples, see: (a) Ito, Y.; Konoike, T.; Harada, T.;
Saegusa, T. J. Am. Chem. Soc. 1977, 99, 1487. (b) Baran, P. S.;
DeMartino, M. P. Angew. Chem., Int. Ed. 2006, 45, 7083. (c) DeMartino,
M. P.; Chen, K.; Baran, P. S. J. Am. Chem. Soc. 2008, 130, 11546.
(10) For selected examples, see: (a) Baciocchi, E.; Casu, A.; Ruzziconi,
R. Tetrahedron Lett. 1989, 30, 3707. (b) Avetta, C. T.; Konkol, L. C.; Taylor,
C. N.; Dugan, K. C.; Stern, C. L.; Thomson, R. J. Org. Lett. 2008, 10, 5621.
(11) See Supporting Information.
’ ASSOCIATED CONTENT
S
Supporting Information. Experimental procedures and
b
detailed characterization data for all new compounds. This material is
available free of charge via the Internet at http://pubs.acs.org.
(12) Kolb, H. C.; VanNieuwenhze, M. S.; Sharpless, K. B. Chem. Rev.
1994, 94, 2483.
’ AUTHOR INFORMATION
(13) (a) Ito, Y.; Hirao, T.; Saegusa, T. J. Org. Chem. 1978, 43, 1011.
(b) Larock, R. C.; Hightower, T. R.; Kraus, G. A.; Hahn, P.; Zheng, D.
Tetrahedron Lett. 1995, 36, 2423.
Corresponding Author
seth.herzon@yale.edu
(14) Huot, R.; Brassard, P. Can. J. Chem. 1974, 52, 838.
(15) This reaction proceeds in 75% yield, as determined by 1H NMR
analysis of the unpurified reaction mixture against an internal standard.
However, under all purification conditions that we have examined, the
products 18 and 19 partially decompose. [Under optimized conditions,
elution of analytically pure samples of 18 (or 19) resulted in 70%
recovery.]
(16) (a) Evans, S.; Hamnett, A.; Orchard, A. F.; Lloyd, D. R. Faraday
Discuss. Chem. Soc. 1972, 54, 227. (b) Bryant, J. R.; Taves, J. E.; Mayer,
J. M. Inorg. Chem. 2002, 41, 2769.
Author Contributions
†These authors contributed equally to this work.
’ ACKNOWLEDGMENT
Financial support from Yale University, the National Science
Foundation (Graduate Research Fellowship to C.M.W.), Eli Lilly,
and the Searle Scholars Program is gratefully acknowledged.
We thank Dr. Haiyin He (Pfizer Pharmaceuticals) and Prof.
Andrew J. Phillips (Yale University) for helpful discussions. We
thank Patrick Lynch for design of the cover art.
(17) The ratio of 22 and 23 varied somewhat between experiments
but was consistently within the range 1.5-3:1. For example, in a separate
100 mg scale experiment, 22, 23, and 18 were obtained in 33%, 23%, and
15% yield, respectively.
’ REFERENCES
(18) Deprotection of the (1S,10S)-dimer derived from 19 (TFA,
CH3OH) provided 24 in 38% yield; see Supporting Information.
(19) Myers, A. G.; Fundy, M. A. M.; Lindstrom, P. A. Tetrahedron
Lett. 1988, 29, 5609.
(1) He, H.; Ding, W. D.; Bernan, V. S.; Richardson, A. D.; Ireland,
C. M.; Greenstein, M.; Ellestad, G. A.; Carter, G. T. J. Am. Chem. Soc.
2001, 123, 5362.
(2) (a) Ito, S.; Matsuya, T.; Omura, S.; Otani, M.; Nakagawa, A.
J. Antibiot. 1970, 23, 315. (b) Hata, T.; Omura, S.; Iwai, Y.; Nakagawa, A.;
Otani, M. J. Antibiot. 1971, 24, 353. (c) Omura, S.; Nakagawa, A.;
Yamada, H.; Hata, T.; Furusaki, A.; Watanabe, T. Chem. Pharm. Bull.
1971, 19, 2428. (d) Omura, S.; Nakagawa, A.; Yamada, H.; Hata, T.;
Furusaki, A. Chem. Pharm. Bull. 1973, 21, 931. (e) Seaton, P. J.; Gould,
S. J. J. Antibiot. 1989, 42, 189. For a review of kinamycin biosynthesis,
see: (f) Gould, S. J. Chem. Rev. 1997, 97, 2499.
(3) DeBoer, T. J.; Backer, H. J. Organic Syntheses; Wiley: New York,
1956; Vol. 36, p 16.
(4) (a) Laufer, R. S.; Dmitrienko, G. I. J. Am. Chem. Soc. 2002, 124,
1854. (b) Zeng, W.; Ballard, T. E.; Tkachenko, A. G.; Burns, V. A.;
Feldheim, D. L.; Melander, C. Bioorg. Med. Chem. Lett. 2006, 16, 5148.
(c) Feldman, K. S.; Eastman, K. J. J. Am. Chem. Soc. 2006, 128, 12562.
(d) Khdour, O.; Skibo, E. B. Org. Biomol. Chem. 2009, 7, 2140.
(5) For studies toward the syntheses of 1 and 2, see: (a) Nicolaou,
K. C.; Denton, R. M.; Lenzen, A.; Edmonds, D. J.; Li, A.; Milburn, R. R.;
Harrison, S. T. Angew. Chem., Int. Ed. 2006, 45, 2076. (b) Krygowski,
E. S.; Murphy-Benenato, K.; Shair, M. D. Angew. Chem., Int. Ed. 2008, 47,
7263
dx.doi.org/10.1021/ja200034b |J. Am. Chem. Soc. 2011, 133, 7260–7263