Stereoselective synthesis of a tricyclic guanidinium model of cylindrospermopsin

Article abstract of DOI:10.1016/S0040-4039(99)02348-5

An 11-step enantioselective synthesis of the A-ring of cylindrospermopsin is described using an intramolecular conjugate addition as the key step to forming the piperidine ring. Further elaboration generates a tricyclic guanidine via a sequential double displacement strategy as a model for the cylindrospermopsin guanidinium core. (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0040-4039(99)02348-5

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 41 (2000) 1849–1853
Stereoselective synthesis of a tricyclic guanidinium model of
cylindrospermopsin
Indrawan J. McAlpine † and Robert W. Armstrong ∗,‡
Department of Chemistry and Biochemistry, University of California, Los Angeles, CA 90095, USA
Received 4 November 1999; revised 15 December 1999; accepted 16 December 1999
Abstract
An 11-step enantioselective synthesis of the A-ring of cylindrospermopsin is described using an intramolecular
conjugate addition as the key step to forming the piperidine ring. Further elaboration generates a tricyclic guanidine
via a sequential double displacement strategy as a model for the cylindrospermopsin guanidinium core. © 2000
Elsevier Science Ltd. All rights reserved.
Keywords: cylindrospermopsin; guanidines; piperidines.
In 1979, drinking water contaminated with cyanobacteria caused a hepatoenteritis outbreak in
Australia.1 The causative toxin, cylindrospermopsin (1), was first isolated by Moore2a in 1992 from
the cyanobacterium Cylindrosperopsis raciborskii and was later found in Umezakia natans2b and
Aphanizomenon ovalisporum (Scheme 1).2c Cylindrospermopsin is toxic to cultured rat hepatocytes and
is thought to inhibit reduced glutathione synthesis.3 Architecturally, cylindrospermopsin is a compact
zwitterionic molecule composed of a tricyclic guanidinium portion and uracil section. Currently, two
groups have published synthetic approaches to this complex natural product,4 including Snider and
Xie’s4a very concise synthesis of the hydroxymethyluracil fragment attached to a model of the AB-
bicyclic guandinium system. The dense functionality and multiple stereocenters make the synthesis of
cylindrospermopsin a very challenging prospect.
Our approach to cylindrospermopsin relies on forming the hydroxymethyl unit via a C-6 uracil anion5
addition into the appropriately protected aldehyde 2 (Scheme 2). Guanylation of piperidine 3 and double
displacement of the appropriate alcohols would yield the tricyclic guanidine core. The stereochemistry of
the acyclic amine 4 should give rise to a single diastereomer upon intramolecular conjugate addition. Here
∗
†
Corresponding author. E-mail: indra@slsiris.harvard.edu (I. J. McAlpine), roba@lilly.com (R. W. Armstrong)
Current address: Institute of Chemistry and Cell Biology, Harvard Medical School, 250 Longwood Ave., SGM604 Boston,
MA 02115. Tel: (617) 432-4957.
‡
Current address: Eli Lilly & Co., Lilly Corporate Center, DC 0530 Indianapolis, IN 46285. Tel: (317) 276-4678; fax: (317)
276-4523.
0040-4039/00/$ - see front matter © 2000 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(99)02348-5
tetl 16276

1850
Scheme 1.
we report the stereoselective synthesis of the protected tricyclic guanidine 27 as a model demonstrating
the feasibility of our synthetic design towards cylindrospermopsin.
Scheme 2.
Sodium borohydride reduction of the commercially available activated aspartic acid derivative 5
followed by TBS protection of the resulting primary alcohol gave 6 in 93% yield (Scheme 3). Modifying
conditions developed by Seebach6 for similar aspartic acid analogs, formation of the dianion of 6 at
−78°C using 2 equiv. of LDA followed by addition of iodomethane generated a 3:1 ratio of diastereomers
7 and 8 in 97% yield. At this point 7 was assigned the anti-configuration based on the precedent by
Seebach.6b DIBAL-H reduction of the t-butyl ester 7 created aldehyde 9 in 89% yield, and the primary
alcohol 10 in 11% yield which was easily oxidized to 9 using Dess–Martin periodinane (Scheme 3).7
Scheme 3.
Diastereoselective allylation of aldehyde 9 in an anti-Felkin–Anh fashion created the last of the three
contiguous stereocenters contained in the piperidine A-ring of cylindrospermopsin (Scheme 4). The
Brown allylboration reaction using (−)-isopinocamphorborane auxiliary8 gave the best diastereomeric
ratio (2:1) of 11 to 12 in an acceptable 70% yield. However, we did find that treating a solution of
aldehyde 9 and 2 equiv. of allyltributyltin in CH2Cl2 at −78°C with 1 equiv. of BF3OEt2 gave superior
yields of 84% but with diminished diastereoselectivity (1.5:1 of 11:12). Generation of the corresponding
cyclic carbamates 13 and 14 helped us to determine the absolute stereochemisty of 11 and 12. 2D NOESY
experiments indicated that the cyclic carbamates 13 and 14 had the stereochemistries and conformations

1851
as drawn.9 These data confirm the expected diastereoselectivity for the methylation of 6 and prove the
absolute stereochemistry of alcohols 11 and 12.
Scheme 4.
Protection of alcohol 11 with TBSOTf and 2,6-lutidine was quantitative. Ozonolysis of the terminal
alkene resulted in a mixture of cyclic hemiaminals 15 (Scheme 5). In order to avoid this undesired
cyclization the carbamate nitrogen of the bis-TBS diol was protected using TFAA and Et3N in 90%
yield. Subsequent ozonolysis generated aldehyde 16 in 91% yield. Reacting 16 with 2.5 equiv. of freshly
prepared phosphonate ylide generated from 17 in THF followed by methanolic sodium carbonate work-
up gave the α,β-unsaturated methyl ketone 18 in 64% yield along with hemiaminals 15 in 27% yield. As
predicted, treating 18 with a catalytic amount of pTSOH in refluxing benzene gave one diastereomer of
the corresponding Z-protected piperidine in 74% yield. Subsequent hydrogenation generated piperidine
19 in quantitative yield (Scheme 5).
Scheme 5.
2D NOESY experiments confirmed that the relative and absolute stereochemisty of 19 corres-
ponds to the stereochemistry of the A-ring of cylindrospermopsin.9 Treatment of 19 with bis-Z-
methylthiopseudourea 20,10 HgCl2 and Et3N in DMF guanylates the hindered piperidine nitrogen ge-
nerating 21 in 85% yield, along with a small amount of ring-opened product (Scheme 6). The first of the
two envisioned SN2 displacements requires the alcohol with appropriate stereochemistry resulting from
reduction of the methyl ketone. As a model study, treatment of 21 under non-asymmetric conditions using
sodium borohydride quantitatively reduced the methyl ketone into an inseparable mixture of alcohols 22
that exists in roughly a 5:1 ratio by NMR. Subjecting the mixture 22 to Mitsunobu conditions generated
the guanidine bicycles 23 and 24 in 75% combined yield (Scheme 6).11 2D NOESY experiments on the
minor bicycle 24 confirm the equatorial conformation of the methyl group in the newly formed B-ring.9
Due to the small amount of 24 formed, bicyclic guanidine 23 was chosen to test conditions for tricycle
formation. Sodium hydride in a THF:methanol (1:1) solution selectively deprotected one of the Z-groups
to generate 25 in 67% yield (Scheme 7).12 Treatment of the monoprotected guanidine with TBAF solution

1852
Scheme 6.
in THF removed both TBS groups to give 26 in 84% yield. Cyclization using Mitsunobu conditions
successfully generated the protected guanidine tricycle 27 as the only isolated product in 27% yield.
Scheme 7.
We have shown the utility of our synthetic plan towards cylindrospermopsin by synthesizing the
bicyclic guanidine 24, which possesses the relative and absolute stereochemistry of the natural product.
Construction of the natural product would require the appropriately protected α-hydroxyphosphonate
in place of 17 to generate aldehyde 2 following the same route. Finally, the synthesis of tricycle 27
demonstrates the feasability of a double displacement strategy to install the tricyclic guanidinium core of
cylindrospermopsin.
Acknowledgements
The authors thank Dr. J. Strouse and T. M. Stark for their help with the 2D NMR data. This work was
made possible by financial support from UCLA and NIH grant GM51095. I.J.M. was supported on a
NSF Predoctoral Fellowship.
References
1. (a) Hawkins, P. R.; Chandrasena, N. R.; Jones, G. J.; Humpage, A. R.; Falconer, I. R. Toxicon 1997, 35, 341. (b) Terao, K.;
Ohmori, S.; Igarashi, K.; Ohtani, I.; Watanabe, M. F.; Harada, K. I.; Ito, E.; Watanabe, M. Toxicon 1994, 32, 833. (c) Moore,
R. E.; Ohtani, I.; Moore, B. S.; de Koning, C. B.; Yoshida, W. Y.; Runnegar, M. T. C.; Carmichael, W. W. Gazz. Chim. Ital.
1993, 123, 329, and references cited therein.
2. (a) Ohtani, I.; Moore, R. E.; Runnegar, M. T. C. J. Am. Chem. Soc. 1992, 114, 7941. (b) Harada, K.-I.; Ohtani, I.; Iwamoto,
K.; Suzuki, M.; Watanabe, M. F.; Watanabe, M.; Terao, K. Toxicon 1994, 32, 73. (c) Banker, R.; Carmeli, S.; Hadas, O.;
Teltsch, B.; Porat, R.; Sukenik, A. J. Phycol. 1997, 33, 613.
3. (a) Runnegar, M. T.; Kong, S.-M.; Zhong, Y.-Z.; Lu, S. C. Biochem. Pharm. 1995, 49, 219. (b) Runnegar, M. T.; Kong,
S.-M.; Zhong, Y.-Z.; Ge, J.-L.; Lu, S. C. Biochem. Biophys. Res. Comm. 1994, 201, 235.
4. (a) Snider, B. B.; Xie, C. Tetrahedron Lett. 1998, 39, 7021. (b) Heintzelman, G. R.; Weinreb, S. M.; Parvez, M. J. Org.
Chem. 1996, 61, 4594. (c) Snider, B. B.; Harvey, T. C. Tetrahedron Lett. 1995, 36, 4587. (d) Heintzelman, G. R.; Weinreb,
S. M. Synlett 1993, 551.

1853
5. Du, X. Ph.D. Thesis, University of California-Los Angeles, 1997.
6. (a) Podlech, J.; Seebach, D. Liebigs Ann. 1995, 1217. (b) Estermann, H.; Seebach, D. Helv. Chim. Acta. 1988, 71, 1824. (c)
Seebach, D.; Wasmuth, D. Angew. Chem., Int. Ed. Engl. 1981, 20, 971.
7. Dess, D. B.; Martin, J. C. J. Org. Chem. 1983, 48, 4155.
8. Brown, H. C.; Jadhav, P. K. J. Am. Chem. Soc. 1983, 105, 2092.
9. 2D NOESY experiments revealed these relevant interactions:
10. (a) Su, W. Synth. Commun. 1996, 26, 407. (b) Kim, K. S.; Qian, L. Tetrahedron Lett. 1993, 34, 7677.
11. Dodd, D. S.; Kozikowski, A. P. Tetrahedron Lett. 1994, 35, 977.
12. We noticed similar selectivity using hydrazine on other bis-Z-protected cyclic guanidines.