A two-directional approach to the anatoxin alkaloids: Second synthesis of homoanatoxin and efficient synthesis of anatoxin-a

Article abstract of DOI:10.1039/b804304c

Total syntheses of the potent neurotoxins, environmental hazards, and biochemical probes anatoxin-a and homoanatoxin have been achieved from a common intermediate using a combined two-directional synthesis-tandem reaction strategy which includes key advan

Full text of DOI:10.1039/b804304c

View Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
A two-directional approach to the anatoxin alkaloids: second synthesis of
homoanatoxin and efficient synthesis of anatoxin-awz
Stephen J. Roe and Robert A. Stockman*y
Received (in Cambridge, UK) 12th March 2008, Accepted 22nd April 2008
First published as an Advance Article on the web 23rd May 2008
DOI: 10.1039/b804304c
Total syntheses of the potent neurotoxins, environmental
hazards, and biochemical probes anatoxin-a and homoanatoxin
have been achieved from a common intermediate using a com-
bined two-directional synthesis–tandem reaction strategy which
includes key advances in oxidative desymmetrisation, tandem
Michael–intramolecular Mannich cyclisations and a new method
for deprotection of N-tosyl anatoxin-a.
N-BOC anatoxin-a). In this paper, we present a 10 step total
synthesis of anatoxin-a (27.1% overall yield) and the second
total synthesis of homoanatoxin (10 steps, 14.8% overall yield)
using a combined two-directional synthesis and tandem reac-
tion strategy which allows both natural products to be
accessed from a common intermediate. Our retrosynthetic
analysis is shown in Fig. 1.
We envisaged that a Mannich and a Horner–Wads-
worth–Emmons disconnective approach would allow the use
of iminium synthon 3 as a precursor to both anatoxin-a and
homoanatoxin by choice of keto-phosphonate coupling part-
ner. Access to iminium 3 should be available from a symme-
trical dialdehyde derived from oxidative cleavage of dialkene
5, enabling the use of two-directional synthesis for the forma-
tion of our key common intermediate—iminium precursor 4.
The synthesis of key intermediate 4 is shown in Scheme 1.
Formation of the Grignard reagent of 3-butenyl bromide,
followed by reaction with 0.5 equivalents of ethyl formate,
resulted in the formation of symmetrical alcohol 6 in 76%
yield. Mitsunobu coupling of alcohol 6 with N-tosyl-tert-butyl
carbamate cleanly gave doubly protected amine 7, which was
then treated with TFA in order to remove the BOC protecting
group. We then exploited Schreiber’s desymmetrising cyclic
alkene ozonolysis methodology9 to access non-symmetrical
dialdehyde derivative 4: initial ring-closing metathesis was
carried out using 4.6 mol% of Grubb’s first generation cata-
lyst10 to convert dialkene 5 into cyclic alkene 8, followed by
the desymmetrising oxidative cleavage of 8 using ozone in
methanol followed by a reductive work-up, giving a quantita-
tive yield of aldehyde 4. This constitutes the first example of a
hemiaminal formation using Schreiber’s strategy, and the
proposed mechanism for this transformation is shown in
Anatoxin-a (1) was first identified after investigations into the
cause of incidents of fatal poisoning of wild and domestic
animals by cyanobacterial blooms in North America and
Europe,1 leading it to be initially called VFDF (very fast death
factor). Anatoxin-a was initially isolated from the blue-green
algae Anabaena flos aquae in the 1970’s,2 and has subsequently
been isolated from several other toxic strains of fresh water
cyanobacteria.3 Homoanatoxin (2) was first characterised as a
synthetic analogue of anatoxin-a by Gallagher et al. in 1992,4
and was subsequently isolated from strains of Oscillatoria
formosa5 and Raphidiopsis mediterranea.3f
Both anatoxin-a (1) and homoanatoxin (2) display potent
activity as agonists at nicotinic acetylcholine receptors
(nAChR) with a high degree of selectivity over muscarinic
acetylcholine receptors (mAChR), which has led to them being
used as tools to investigate muscle and neuronal nAChR.6 Due
to their modulatory role in the brain, nAChRs have attracted
attention as potential targets for therapeutic intervention in a
range of disease states including neurodegenerative diseases,
attention deficit disorder, pain, smoking cessation, schizophre-
nia, and epilepsy.7 Thus, the intriguing bicyclic structure and
enticing combination of potent and specific biological activity
have prompted a high degree of interest in the synthetic
community, resulting in numerous syntheses of anatoxin-a8
and surprisingly just one synthesis of homoanatoxin4 (from
School of Chemical Sciences and Pharmacy, University of East
Anglia, Norwich, UK NR7 4TJ
w This paper is dedicated to Professor Philip Magnus, FRS, on the
occasion of his 65th birthday.
z Electronic supplementary information (ESI) available: Full experi-
mental procedures and data. Crystallographic data. CCDC 681986
and 681987. See DOI: 10.1039/b804304c
y Present address: School of Chemistry, University of Nottingham,
Nottingham, UK NG7 2RD. E-mail: robert.stockman@nottingham.
ac.uk; Fax: þ44 (0) 1159513564; Tel: þ44 (0) 1159513252.
Fig. 1 Retrosynthetic analysis.
ꢀc
This journal is The Royal Society of Chemistry 2008
3432 | Chem. Commun., 2008, 3432–3434

View Online
Scheme 1 Synthesis of key intermediate 4.
Scheme 2. Thus the primary ozonide ring-opens, but due to the
tether, the formation of the secondary ozonide is slow, and the
carbonyl oxide is instead trapped as a hydroperoxy hemiacetal
by the methanol solvent. This allows hemiaminal formation
between the nitrogen on the chain and the aldehyde. Finally,
addition of dimethyl sulfide reduces the hydroperoxy hemi-
acetal and produces the desymmetrised aldehyde 4.
Scheme 3 Formal synthesis of anatoxin-a.
nature of the HCl–MeOH conditions, which have thus far
been the only reported protocol which has proven successful,13
we set about investigating alternative conditions for this key
transformation. It transpired that TMSI in dichloromethane
at ꢁ78 1C was able to promote a similar transformation
(presumably via iodo silyl enol ether 10b). The outcome of
this reaction was a good crude yield of N-tosyl anatoxin-a
accompanied by significant quantities of diastereomers of the
non-eliminated hydroiodide precursor. It was found that by
treating the crude mixture with DBU in toluene and stirring at
room temperature, any hydroiodide was transformed into
tosyl anatoxin-a (12), giving a 70% yield of 12 from 9 over
the two-step operation.
Somfai,11 and Trost14 have previously deprotected N-tosyl
anatoxin-a (12) using sodium amalgam in methanol at ꢁ40 1C
in the presence of a phosphate buffer, and thus at this stage
our synthesis of 12 in 8 steps and 33% overall yield from ethyl
formate represented a formal synthesis. However, we wished
to detail a scalable and potentially viable route for the
production of anatoxin-a on a multigram scale, and thus we
wished to deprotect our sample of 12. Unfortunately all of our
attempts to achieve the deprotection of our sample of (ꢂ)-12
using the procedures reported by both Somfai and Trost (on
enantiomerically enriched 12) failed, as we found our racemic
material to be insoluble under the conditions reported of
methanol at ꢁ40 1C (we have considered the possibility that
the racemic material has a higher lattice energy than the
enantioenriched material, and may be less soluble). Addition
of THF co-solvent led to complete solvation, but unfortu-
nately reduction of the enone was observed rather than
removal of the sulfonamide. We also tried a very wide range
of other reducing conditions for this transformation, but were
again unsuccessful, noting that if any reduction was seen, it
was of the enone functionality in preference to the sulfona-
mide. Unable to resolve this issue, we decided to protect the
Scheme 3 describes our formal synthesis of anatoxin-a from
key intermediate 4. Treatment of crude aldehyde 4 (which was
found to be unstable to chromatography and was thus used
unpurified) with diethylphosphono-acetone and sodium hy-
dride in THF gave enone 9 as a single alkene isomer but a
mixture of diastereomers across the pyrrolidine ring in 76%
yield over two steps from 8. Somfai and Ahman had already
shown that trans-E-9 was able to undergo a tandem iminium
ion formation–Baylis–Hillman type reaction via an intermedi-
ate of the type 10a to form a mixture of chlorides 11, which in
turn could be converted to N-tosyl anatoxin-a 12 by elimina-
tion of HCl with DBU in toluene at reflux in 67% overall
yield.11,12 In our hands, and using a mixture of diastereomers
of 9 as substrates, we were only able to attain moderate yields
of 12 at best using these conditions.13 Due to the capricious
Scheme 2 Mechanism of desymmetrising ozonolysis.
ꢀc
This journal is The Royal Society of Chemistry 2008
Chem. Commun., 2008, 3432–3434 | 3433

View Online
tactic of enone protection followed by reduction to give
homoanatoxin as the hydrochloride salt 2 in 10 steps and
14.8% overall yield from readily available starting materials.z
In conclusion, we have described a unified approach to the
synthesis of anatoxin-a (1) and homoanatoxin (2) from a
common precursor. Work is on-going in these laboratories
to extend this approach to the synthesis of a range of un-
natural anatoxin-a derivatives and for the multigram syntheses
of these important biological tools. Our reports on these
developments will be disclosed in due course.
Scheme 4 Deprotection of N-tosyl anatoxin-a.
Notes and references
1. W. W. Carmichael, D. F. Biggs and P. R. Gorham, Science, 1975,
187, 542.
2. J. P. Delvin, O. E. Edwards, P. R. Gorham, N. R. Hunter,
R. K. Pike and B. Stavric, Can. J. Chem., 1977, 55, 1367.
3. (a) M. Gugger, S. Lenoir, C. Berger, A. Ledreux, J. C. Druart,
J. F. Humbert, C. Guette and C. Bernard, Toxicon, 2005, 45, 919;
(b) C. Edwards, K. A. Beattie, C. M. Scrimgeour and G. A. Codd,
Toxicon, 1992, 30, 1165; (c) K. Sivonen, K. Himberg,
R. Luukkainen, S. I. Niemelae, G. K. Poon and G. A. Codd,
Toxic. Assess., 1989, 4, 339; (d) S. A. Wood, D. J. Stirling,
L. R. Briggs, J. Sprosen, P. Holland, J. G. Ruck and R. G. N.
Z. Wear, J. Mar. Freshwater Res., 2006, 40, 585; (e) E. Viaggiu,
S. Melchiorre, F. Volpi, A. Di Corcia, R. Mancini, L. Garibaldi,
G. Crichigno and M. Bruno, Environ. Toxicol., 2004, 19, 191;
(f) M. Namikoshi, T. Murakami, M. F. Watanabe, T. Oda,
J. Yamada, S. Tsujimura, H. Nagai and S. Oishi, Toxicon, 2003,
42, 533; (g) A. Ballot, L. Krienitz, K. Kotut, C. Wiegand and
S. Pflugmacher, Harmful Algae, 2005, 4, 139.
Scheme 5 Synthesis of homoanatoxin.
4. S. Wonnacott, K. L. Swanson, E. X. Albuquerque, N. J. S. Huby,
P. Thompson and T. Gallagher, Biochem. Pharmacol., 1992, 43,
419.
5. O. M. Skulberg, W. W. Carmichael, R. A. Andersen,
S. Matsunaga, R. E. Moore and R. Skulberg, Environ. Toxicol.
Chem., 1992, 11, 321.
6. For a recent review on the pharmacology of anatoxin-a and
analogues, see: S. Wonnacott and T. Gallagher, Mar. Drugs,
2006, 4, 228.
7. C. B. Cassels, I. Bermudez, F. Dajas, J. A. Abin-Carriquiry and
S. Wonnacott, Drug Discovery Today, 2005, 10, 1657.
8. For reviews on synthetic approaches to anatoxin-a, see:
H. L. Mansell, Tetrahedron, 1996, 52, 6025; V. Rodriguez, S. M.
e E. Pinto, C. M. P. Pereira and R. C. Braga, Quim. Nova, 2006,
29, 1365.
enone, such that more severe reducing conditions could be
utilised to remove the sulfonamide (Scheme 4).
Reaction of 12 with ethylene glycol under Dean–Stark
conditions gave the dioxolane 13 in 95% yield. In an analo-
gous reaction to the final step of Backvall and co-workers’s
ferruginine synthesis,15 sulfonamide 13 was treated with an
initial portion of magnesium powder (30 equivalents) in
methanol with sonication. After 40 minutes an additional 30
equivalents of magnesium powder were added and the sonica-
tion continued for a further hour before the addition of
aqueous HCl. It was found that the dioxolane protecting
group was removed and anatoxin-a was recovered as the
hydrochloride salt 1 in 82% yield over two steps from 12.
Overall, this constitutes a synthesis of anatoxin-a in 10 steps
and 27.1% overall yield from readily available starting
materials.
9. S. L. Schreiber, R. E. Claus and J. Regan, Tetrahedron Lett., 1982,
23, 3867.
10. T. M. Trnka and R. H. Grubbs, Acc. Chem. Res., 2001, 34, 18.
11. P. Somfai and J. Ahman, Tetrahedron Lett., 1992, 33, 3791. The
same cascade was also used by Speckamp et al. in their synthesis of
anatoxin-a, in that case with a carbomethoxy nitrogen protecting
group: K. H. Melching, H. Hiemstra, W. J. Klaver and
W. N. Speckamp, Tetrahedron Lett., 1986, 27, 4799.
With the synthesis of anatoxin-a (1) complete, our attention
turned to the adaption of this strategy for the synthesis of
homoanatoxin (2), which had previously been synthesised
by Gallagher in two steps and 19.2% yield from N-BOC
anatoxin-a.4 Scheme 5 shows our synthesis.
12. Royer and Husson also used this type of cyclisation promoted by
HCl in methanol in the synthesis of ferruginine: I. Gauthier,
J. Royer and H.-P. Husson, J. Org. Chem., 1997, 62, 6704.
13. These conditions had previously proved problematic to others too:
Tanner et al. reported a 51% yield for the HCl–MeOH cyclisation
and also observed racemisation of their enantioenriched system,
which used a carbomethoxy protecting group and an ethoxyhe-
miaminal iminium precursor: T. Hjelmgaard, I. Støtofte and
D. Tanner, J. Org. Chem., 2005, 70, 5688. In this paper, Tanner
describes several other sets of conditions that were investigated
and were not able to promote this cyclisation (H2SO4–MeOH,
TFA–CH2Cl2, TiCl4–CH2Cl2, DMAP–Sn(OTf)2–CH2Cl2).
14. B. M. Trost and J. D. Oslob, J. Am. Chem. Soc., 1999, 121, 3057.
15. S. Y. Jonsson, C. M. G. Lofstrom and J.-E. Backvall, J. Org.
Chem., 2000, 65, 8454.
Accordingly, alkene 8 was subjected to the desymmetrising
ozonolysis reaction to give aldehyde 4 which underwent olefi-
nation with 1-(diethylphosphono)butan-2-one to give enone
14. The TMSI-promoted tandem iminium formation–enol
ether formation–intramolecular Mannich reaction and subse-
quent elimination worked as on our previous substrate, giving
a 55% yield of 15 for this key step. Removal of the sulfona-
mide protecting group of 15 was accomplished using the same
ꢀc
This journal is The Royal Society of Chemistry 2008
3434 | Chem. Commun., 2008, 3432–3434