A concise asymmetric synthesis of the marine hepatotoxin 7-epicylindrospermopsin

Article abstract of DOI:10.1002/anie.200454208

Born from a simple amino acid ... the potent hepatotoxic cyanobacterial alkaloid 7-epicylindrospermopsin (see picture) has been synthesized through a concise asymmetric eighteen-step route. An intramolecular 1,3-dipolar cycloaddition and a nitroaldol reac

Full text of DOI:10.1002/anie.200454208

Communications
Palm Island, Australia.[1] It has since been isolated in Japan
from Umezakia natans and in Israel from Aphazinomenon
ovalisporum.[2] Following the discovery of the parent com-
pound, 7-epicylindrospermopsin (2) was isolated from A.
ovalisporum as a toxic minor metabolite.[3] The natural
congener 7-deoxycylindrospermopsin (3) was initially iso-
lated from C. raciborskii and has recently been co-isolated
with 1 in China from Raphidiopsis curvata.[4] Cylindrosper-
mopsin has been shown independently to be a potent
hepatotoxin (LD50 ~ 0.2 mgkgꢀ1 in mice), and 7-epicylindro-
spermopsin is equipotent with 1, whereas 3 is nontoxic.[1,4a,5] It
is thought that the toxicity of these substances results from a
general inhibition of protein synthesis, but it is unclear
whether this action occurs at the ribosomal or transcriptional
level by the interaction of oxidized metabolites with DNA.[6,7]
The threat posed to global public health by these molecules in
drinking water and the isolation of C. raciborskii in several
regions of the United States have prompted the National
Toxicology Program (NTP) of the National Institutes of
Health and the Unregulated Contaminant Monitoring Rule
(UCMR) of the Environmental Protection Agency to elect 1
for toxicological and environmental evaluation.[8] The paucity
of understanding of the biological functions of these agents
has prompted us to develop an efficient and flexible synthetic
approach that will be adapted to the production of uracil
analogues to be deployed for biomechanistic evaluation.
These highly polar compounds pose a significant synthetic
challenge as they contain almost as many heteroatoms as
carbon atoms, and their structures include a zwitterionic
guanidinium sulfate, a rare tetrasubstituted tricyclic guani-
dine, and a uracil moiety. Not surprisingly, these natural
products have attracted the attention of synthetic chemists for
over ten years.[9] Snider and co-workers completed the first
racemic total synthesis of cylindrospermopsin in 20 steps eight
years after its discovery.[10] However, their studies failed to
illuminate the misassigned stereogenic center at C7. The
configuration at that center was corrected by Weinreb and co-
workers in an elegant 30-step racemic, stereoselective syn-
thesis of 2, thus validating the illustrated structures.[11] Shortly
thereafter White and Hansen completed the first asymmetric
total synthesis of 2 in 28 steps and confirmed the absolute
stereochemistry as 7S, 8R, 10S, 12S, 13R, 14S.[12]
Natural Product Synthesis
A Concise Asymmetric Synthesis of the Marine
Hepatotoxin 7-Epicylindrospermopsin**
Ryan E. Looper and Robert M. Williams*
The cyanobacterial toxin cylindrospermopsin (1) was isolated
as the principal hepatotoxin from Cylindrospermopsis raci-
borskii in 1992 after suspicion of its involvement in an
outbreak of hepatoenteritis that hospitalized 150 people on
We envisaged that the three contiguous stereogenic
centers of the A ring of cylindrospermopsin could be con-
structed by a single intramolecular 1,3-dipolar cycloaddition
of the nitrone 4 to yield the tricyclic compound 5 (Scheme 1).
At the onset of our work, the absolute stereochemistry of
these natural products was not known. Retrosynthetic
analysis suggested that the nitrone 4 could be derived from
a simple amino acid, namely (R)- or (S)-crotylglycine, so that
1 could be constructed from a precursor with a single
stereogenic center, which would provide the inherent flexi-
bility to produce either enantiomer of the natural product.
The two remaining, remote stereogenic centers at C7 and C8
and their respective oxidation states define this family in
terms of both structure and function. Considering this, it was
anticipated that a nitroaldol (Henry reaction) strategy would
permit access to both C7 epimers (1 and 2), and through
reductive nitromethylation provide access to 3.
[*] R. E. Looper, Prof. Dr. R. M. Williams
Department of Chemistry
Colorado State University
Fort Collins, CO 80523 (USA)
Fax: (+1)970-491-3944
E-mail: rmw@chem.colostate.edu
[**] This work was supported by the National Institutes of Health (Grant
GM068011) and the National Science Foundation (Grant
CHE0202827). We are grateful to Array Biopharma for fellowship
support to R.E.L. Mass spectra were obtained on instruments
supported by the National Institutes of Health Shared Instrumen-
tation Grant (GM49631).
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
2930
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/anie.200454208
Angew. Chem. Int. Ed. 2004, 43, 2930 –2933

Angewandte
Chemie
The installation of N16 proved difficult as it
necessitated simultaneous oxidation-state
a
change at C15. It was eventually found that a
two-step protocol in which the lactone carbonyl
group was first reduced to the corresponding
lactol with DIBAL-H fulfilled this requirement
ꢀ
(Scheme 3). Tandem reductive amination/N O
bond cleavage gave a free diamine which was
more readily handled as the urea 13 after in situ
treatment with bis(p-nitrophenyl)carbonate.[18]
Selective oxidation of the primary alcohol with
TEMPO gave the sensitive aldehyde 6 in 75%
yield.[19] Most common oxidation protocols
showed selectivity for the secondary hydroxy
group or led to epimerization of the sensitive a-
ureidoaldehyde. Eventually it was found that the
addition of methanesulfonic acid (1 mol%) accel-
erated the reaction to give 6 without attendant
Scheme 1. Retrosynthesis of the cylindrospermopsins. PMB=p-methoxybenzyl.
1
To this end, the protected crotylglycine derivative 11 was
prepared from our oxazinone template 10 (Scheme 2).[13]
Thus, alkylation of 10 with (E)-crotyl iodide gave the
epimerization, as evidenced by H NMR NOE experiments.
The aldehyde 6 underwent nitromethylation upon treatment
with the lithium salt of nitromethane in THF. The resulting
Scheme 2. a) KHMDS, (E)-crotyl iodide, THF, ꢀ788C, 92%, >99% ee;
b) Li, NH3, THF, EtOH, 68–87%; c) AcCl, MeOH, 08C!RT; d) LiAlH4,
THF, 65% (2 steps); e) BrCH2CO2Ph, iPr2NEt, MeCN, 63–80%;
f) MCPBA, Na2HPO4, CH2Cl2, ꢀ788C, 84%; g) PhMe, 2008C (sealed
tube), 78%. KHMDS=potassium hexamethyldisilazide, MCPBA=3-
chloroperoxybenzoic acid.
Scheme 3. a) DIBAL-H, CH2Cl2, ꢀ788C, 87%; b) PMBNH2, H2/Pd/C, EtOAc, then
(p-NO2PhO)2CO, MeCN, 81%; c) TEMPO, PhI(OAc)2, MsOH (1 mol%), CDCl3,
75%; d) MeNO2, nBuLi, THF, room temperature, 84%; e) Ac2O, DMAP, CH2Cl2,
then NaBH4, EtOH, 67%; f) TFA (neat), reflux, 80%; g) Et3OBF4, Cs2CO3, CH2Cl2,
78%. DIBAL-H=diisobutylaluminum hydride, DMAP=4-dimethylaminopyridine,
MsOH=methanesulfonic acid, TEMPO=2,2,6,6-tetramethylpiperidine N-oxide,
TFA=trifluoroacetic acid.
corresponding crotylated lactone in 92% yield with
> 99% ee.[14] Reductive removal of the chiral auxiliary with
lithium in ammonia gave 11 in moderate to good yield. Acidic
removal of the Boc group with concomitant methyl ester
formation followed by reduction with lithium aluminum
hydride gave the optically pure alcohol 12. This substance was
then transformed into the free morpholinone in a one-pot
procedure by treatment with bromophenyl acetate.[15] Oxida-
tion of the secondary amine was most conveniently effected
by treatment with purified MCPBA in dichloromethane to
give the oxazinone N-oxide 4 in 84% yield.[16] Exposure of 4
to elevated temperatures gave the tricyclic isoxazolidine 5 in
78% yield as a 10:1 mixture with the minor product assumed
to arise from an endo approach of the alkene to the
nitrone.[9i,17] We also observed that the treatment of 4 with
scandium triflate at room temperature for up to three days
generated 5 in an improved 12:1 mixture. The relative
stereochemistry of 5 was confirmed by X-ray crystallogra-
phy.[9i]
nitroalcohol was peracetylated with acetic anhydride, and the
nitroalkene was then reduced to afford the nitroalkane 14 in
67% yield. The PMB group was removed cleanly in TFA at
reflux to give an 80% yield of the urea poised for alkylative
activation.[20] Attempts to methylate the oxygen atom of the
urea gave an unstable compound that decomposed. Fortu-
nately, O-ethylation with triethyloxonium tetrafluoroborate
gave the satisfactorily stable isourea 15 (78% yield), which
could be subjected to subsequent reactions.[21]
Initial attempts to effect the nitroaldol reaction yielded
essentially equimolar amounts of all four C7/C8 diastereom-
ers.[22,23] The treatment of 15 and 2,6-dimethoxypyrimidine-4-
carbaldehyde with 2 equivalents of tetra-n-butylammonium
fluoride for short reaction times, followed by reductive
guanidinylation of the nitro group, gave the best selectivities
(Scheme 4).[24,25] These conditions afforded a kinetic 1:0.8
mixture in favor of the diastereomer required for the syn-
Angew. Chem. Int. Ed. 2004, 43, 2930 –2933
www.angewandte.org
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2931

Communications
[2] a) K. Harada, I. Ohtani, K. Iwamoto, M. Suzuki, M. F. Wana-
tabe, M. Wanatabe, K. Terao, Toxicon 1994, 32, 73; b) R. Banker,
S. Carmeli, O. Hadas, B. Teltsch, R. Porat, A. Sukenik, J. Phycol.
1997, 33, 613.
[3] R. Banker, B. Teltsch, A. Sukenik, S. Carmeli, J. Nat. Prod. 2000,
63, 387.
[4] a) R. L. Norris, G. K. Eaglesham, G. Pierens, G. R. Shaw, M. J.
Smith, R. K. Chiswell, A. Seawright, M. R. Moore, Environ.
Toxicol. 1999, 14, 163; b) L. Renhui, W. W. Carmichael, S.
Brittain, G. K. Eaglesham, G. R. Shaw, Y. Liu, M. M. Watanabe,
J. Phycol. 2001, 37, 1121.
[5] M. T. Runnegar, C. Xie, B. B. Snider, G. A. Wallace, S. M.
Weinreb, J. Kuhlenkamp, Toxicol. Sci. 2002, 67, 81.
[6] a) M. T. Runnegar, S. Kong, Y. Z. Zhong, J. L. Ge, S. C. Lu,
Biochem. Biophys. Res. Commun. 1994, 201, 235; b) M. T.
Runnegar, S. M. Kong, Y. Z. Zhong, S. C. Lu, Biochem. Phar-
macol. 1995, 49, 219; c) S. M. Froscio, A. R. Humpage, P. C.
Burcham, I. R. Falconer, Environ. Toxicol. 2003, 18, 243; d) for a
review, see: D. J. Griffiths, M. L. Saker, Environ. Toxicol. 2003,
18, 78.
Scheme 4. a) 2,6-Dimethoxypyrimidine-4-carbaldehyde, TBAF(2.0 equiv),
ꢀ158C, THF; b) Pd(OH)2, H2, 5% AcOH/MeOH; c) conc. HCl, reflux, 32% for
16, 29% for 17 (3 steps); d) SO3·pyridine, DMF, MS (3 ꢀ), 59%. DMF=N,N-
dimethylformamide, TBAF=tetra-n-butylammonium fluoride.
[7] a) A. R. Humpage, M. Fenech, P. Thoma, I. R. Falconer, Mutat.
Res. 2000, 472, 155; b) V. Fessard, C. Bernard, Environ. Toxicol.
2003, 18, 353.
[8] For more information, see: http://www.epa.gov/safewater/stan-
dard/ucmr/, 2001 and http://ntp-server.niehs.nih.gov/htdocs/
Results_status/ResstatC/M000072.html, 2004.
[9] a) G. R. Heintzelman, M. Parvez, S. M. Weinreb, Synlett 1993,
551; b) B. B. Snider, T. C. Harvey, Tetrahedron Lett. 1995, 36,
4587 – 4590; c) G. R. Heintzelman, S. M. Weinreb, J. Org. Chem.
1996, 61, 4594; d) B. B. Snider, C. Xie, Tetrahedron Lett. 1998, 39,
7021; e) J. D. White, J. D. Hansen, Abstracts of Papers 219th
National Meeting of the American Chemical Society (San
Francisco, CA), American Chemical Society, Washington, DC,
2000, ORGN 812; f) I. J. McAlpine, R. W. Armstrong, Tetrahe-
dron Lett. 2000, 41, 1849; g) S. P. Keen, S. M. Weinreb, Tetrahe-
dron Lett. 2000, 41, 4307; h) J. F. Djung, D. J. Hart, E. R. R.
Young, J. Org.Chem. 2000, 65, 5668; i) R. E. Looper, R. M.
Williams Tetrahedron Lett. 2001, 42, 769.
thesis of 2. The other diastereomer in the mixture led to 17,
which is epimeric at C7 with a diastereomer described by
Snider and co-workers (relative configuration not
assigned).[10] It was found that the nitroaldol reaction must
be quenched with acetic acid and the products immediately
subjected to reduction conditions. Isolation of the products
after treatment with TBAF provided ~ 1:1:1:1 mixtures of the
nitroalcohols, thus indicating that the reaction is indeed
highly reversible. At this stage the diastereomeric dimethoxy-
pyrimidines were inseparable. Acidic hydrolysis of the
pyrimidines gave a separable mixture of 16 (32% yield from
15) and 17 (29%).[10,12] In our hands, the sulfonation of the
C12 hydroxy group proved capricious under the conditions
reported in the literature.[10–12] The use of sulfur trioxide/
pyridine complex in DMF with 3-Š molecular sieves gave 2
reproducibly in 59% yield,[26] along with the corresponding
bis(sulfate) (2:1 ratio by HPLC), as previously observed by
others.[11,12,27]
The asymmetric synthesis of 2 detailed herein represents
the shortest successful route toward this family of natural
products. 7-Epicylindrospermopsin has thus been obtained in
only eighteen steps with few protecting-group manipulations
from commercially available 10. Investigations directed at
controlling the diastereoselectivity of the nitroaldol process to
afford cylindrospermopsin (1) are ongoing and will be
reported in due course. The uracil moiety in 1 and 2 has
been deemed essential for their hepatotoxicity.[28] We believe
that the incorporation of the uracil synthon at a late stage in
the synthesis should render our strategy amenable to the
production of uracil analogues for evaluation of their bio-
logical and hepatotoxic activity.
[10] C. Xie, M. T. C. Runnegar, B. B. Snider, J. Am. Chem. Soc. 2000,
122, 5017.
[11] a) S. M. Weinreb, G. R. Heintzelman, W. K. Fang, S. P. Keen,
G. A. Wallace, J. Am. Chem. Soc. 2001, 123, 8851; b) G. R.
Heintzelman, W. K. Fang, S. P. Keen, G. A. Wallace, S. M.
Weinreb, J. Am. Chem. Soc. 2002, 124, 3939.
[12] J. D. White, J. D. Hansen, J. Am. Chem. Soc. 2002, 124, 4950.
[13] a) R. M. Williams, M.-N. Im, J. Am. Chem. Soc. 1991, 113, 9276;
b) R. M. Williams in Advances in Asymmetric Synthesis, Vol. 1
(Ed.: A. Hassner), JAI, New York, 1995, pp. 45 – 94; c) lactone
10 and its enantiomer are available from Aldrich Chemical Co.:
catalogue numbers C33,184-8 (10) and C33,181-3 (enantiomer of
10); d) for a similar preparation of (R)-allylglycine, see: R. M.
Williams, P. J. Sinclair, D. E. DeMong, Org. Synth. 2003, 80, 31.
[14] T. Kanai, S. Irifune, Y. Ishii, M. Ogawa, Synthesis 1989, 283.
[15] J. F. Dellaria, B. D. Santasiero, J. Org. Chem. 1989, 54, 3916.
[16] T. G. Traylor, A. R. Miksztal, J. Am. Chem. Soc. 1987, 109, 2770.
[17] a) P. N. Confalone, E. M. Huie in Organic Reactions, Vol. 36
(Ed.: A. S. Kende), Wiley, New York, 1998, pp. 3 – 173; b) W.
Oppolzer, R. L. Snowden, B. H. Bakker, M. Petrzilka, Tetrahe-
dron 1985, 41, 3497; c) for an application of chiral a-alkoxycar-
bonyl nitrones, see: O. Tamura, K. Gotanda, R. Terashima, M.
Kikuchi, T. Miyawaki, M. Sakamoto, Chem. Commun. 1996,
1861.
Received: March 5, 2004 [Z54208]
Published Online: May 6, 2004
Keywords: alkaloids · cyanobacteria · cycloaddition · guanidine ·
.
nitroaldol reaction
[18] J. Izdebski, D. Pawlak, Synthesis 1989, 6, 423.
[19] A. De Mico, R. Margarita, L. Parlanti, A. Vescovi, G. Piancatelli,
J. Org. Chem. 1997, 62, 6974.
[1] I. Ohtani, R. E. Moore, M. T. C. Runnegar, J. Am. Chem. Soc.
1992, 114, 7941.
[20] G. M. Brooke, S. Mohammed, M. C. Whiting, Chem. Commun.
1997, 1511.
2932
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2004, 43, 2930 –2933

Angewandte
Chemie
[21] H. Meerwein, Org. Synth. 1973, collect. vol. 5, 1080.
[22] For a recent review of the Henry reaction, see: F. A. Luzzio,
Tetrahedron 2001, 57, 915.
[23] Treatment with alkoxide bases in THF or DMSO mainly led to
ꢁ 1:1:1:1 mixtures of all C7/C8 diastereomers. The use of Lewis
acids, such as Zn, Mg, or Cu salts, retards the condensation of 15.
[24] E. L. Stogryn, J. Heterocycl. Chem. 1974, 11, 251.
[25] Contrary to previously reported results (S. Hanessian, P. V.
Devasthale, Tetrahedron Lett. 1996, 37, 987), the use of
2 equivalents of TBAF gave the highest selectivity. The use of
0.5 equivalents of TBAF led to a 10:9:7:5 mixture in favor of the
anti diastereomer required for the construction of 1, and the use
of 1 equivalent of TBAF produced a mixture in which 16 was
only slightly favored.
[26] Synthetic 2 had spectroscopic properties identical to those
reported: [a]D25 = ꢀ12.5 (c = 0.04, H2O); lit.:[3] [a]2D4 = ꢀ20.58(c =
0.04, H2O).
[27] N. Fujii, S. Futaki, S. Funakoshi, K. Akaji, H. Morimoto, R. Doi,
K. Inoue, M. Kogire, S. Sumi, M. Yun, T. Tobe, M. Aono, M.
Matsuda, H. Narusawa, M. Moriga, H. Yajima, Chem. Pharm.
Bull. 1988, 36, 3281.
[28] R. Banker, S. Carmeli, M. Werman, B. Teltsch, R. Porat, A.
Sukenik, J. Toxicol. Environ. Health Part A 2001, 62, 281.
Angew. Chem. Int. Ed. 2004, 43, 2930 –2933
www.angewandte.org
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2933