Synthesis of the putative structures of homopumiliotoxins 235C and 233F

Article abstract of DOI:10.1016/j.tetlet.2003.12.126

A novel approach to diene based quinolizidines, using an intramolecular Heck reaction in which the vinyl bromide double bond undergoes inversion of configuration, is reported. These quinolizidines have previously been proposed as tentative structures for homopumiliotoxin alkaloids 233F and 235C. The mass spectral data of the synthetic materials were different to those of the natural products confirming that the original structures need to be revised.

Full text of DOI:10.1016/j.tetlet.2003.12.126

Tetrahedron
Letters
Tetrahedron Letters 45 (2004) 1627–1630
Synthesis of the putative structures of homopumiliotoxins
235C and 233F
Paul Armstrong, Stephanie Feutren, Helena McAlonan,
Gavin OÕMahony and Paul J. Stevenson*
School of Chemistry, QueenÕs University, Belfast BT9 5AG, Northern Ireland, UK
Received 19 November 2003; revised 17 December 2003; accepted 23 December 2003
Abstract—A novel approach to diene based quinolizidines, using an intramolecular Heck reaction in which the vinyl bromide double
bond undergoes inversion of configuration, is reported. These quinolizidines have previously been proposed as tentative structures
for homopumiliotoxin alkaloids 233F and 235C. The mass spectral data of the synthetic materials were different to those of the
natural products confirming that the original structures need to be revised.
ꢀ 2004 Elsevier Ltd. All rights reserved.
Pumiliotoxin alkaloids, originally isolated from the
defensive skin secretions of the Dendrobates frogs, are
7
OH
8
6
5
indolizidines with a 6Z-alkylidine substituent.1;2 There
are a number of sub-classes, depending on the oxidation
level at C-7 and C-8 of the indolizidine core. Variation
in each sub-class is due to the nature of the alkylidene
unit. The vast majority of pumiliotoxins have a tertiary
alcohol at C-8, exemplified by pumiliotoxin 307A01 (Fig.
1). The allopumiliotoxins possess an additional hydroxyl
group at C-7, for example, allopumiliotoxin 339B3 (Fig.
1). Deoxypumiliotoxins, devoid of the hydroxyl group
at C-8, are rare and to date only one member of this
class deoxypumiliotoxin 251H (Fig. 1) has been iso-
lated.4 These alkaloids are essentially unavailable from
the natural source and this scarcity of material coupled
with intriguing biological activity make pumiliotoxins
ideal targets for synthesis and for developing new syn-
thetic methodologies.5 Of particular note was the
development of iminium ion chemistry,6–8 organonickel
chemistry,9;10 organopalladium chemistry,11;12 organo-
chromium chemistry13;14 and organotitanium chemis-
try15 for the synthesis of these alkaloids.
8a
H
OH
4
N
1
Pumiliotoxin 307A'
3
2
OH
OH
N
OH
H
OH
Allopumiliotoxin 339B'
OH
H
H
N
Deoxypumiliotoxin 251H
Figure 1. Representative examples of pumiliotoxin alkaloids together
with their numbering system.
pumiliotoxins, and have been the subject of some syn-
thetic studies.13;17 Recently a new class of diene based
homopumiliotoxin alkaloid has been isolated from
Madagascan frogs of the genus Mantella18 (Fig. 2). Due
to the scarcity of material, the structures of these com-
pounds were originally tentatively assigned solely by a
combination of GLC/MS and GLC/IR. The C-7 alkyl-
idene side chain was assumed to have a Z-geometry and
the absolute configuration at C-9a was assumed to be S
by analogy to C-8a of the pumiliotoxins. To date no
Homopumiliotoxins contain a quinolizidine core16 and
are much less widespread in distribution than the
Keywords: Homopumiliotoxin alkaloids; Heck reaction.
* Corresponding author. Tel.: +44-28-90274426; fax: +44-28-903821-
17/90274426; e-mail: p.stevenson@qub.ac.uk
CSS Ltd, Almac House, 20 Seagoe Industrial Estate, Craigavon
BT63 5PW, Northern Ireland.
0040-4039/$ - see front matter ꢀ 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2003.12.126

1628
P. Armstrong et al. / Tetrahedron Letters 45 (2004) 1627–1630
Recently, we have been investigating palladium-cataly-
8
7
9
H
sed Heck cyclisations of vinyl bromides as potential
routes to diene based indolizidines.19;20 We now report
that this chemistry can be extended to give diene based
quinolizidines and ultimately the structures, which have
tentatively been assigned to homopumiliotoxin alkaloids
233F and 235C (Scheme 1). Tertiary amines are prob-
lematic substrates for palladium-catalysed reactions so
amide precursor 11 was employed. The carbonyl group
was strategically placed in the first formed ring to pro-
vide optimum bond angles for the palladium-catalysed
cyclisation. The lactam moiety of the key cyclisation
substrate 11 was readily available from commercially
available (S)-2-aminoadipic acid. This amino acid was
converted to the diester hydrochloride salt 3 in 90%
yield by reaction with thionyl chloride in ethanol.21 The
corresponding free base proved to be much too water
soluble for isolation but could be generated in ethanol
by reaction with sodium ethoxide. Reductive amination
with benzaldehyde and sodium borohydride using a
standard procedure22 provided N-benzyl amino acid
diester 4. The overall yield for the N-benzylation of 3
9a
6
OH
N
5
1
4
2
1 (Homopumiliotoxin 235C)
3
H
O
N
N
2 (Homopumiliotoxin 233F)
OH
H
Homopumiliotoxin 221F
Figure 2. Tentative structures of diene based homopumiliotoxins iso-
lated from Mantella frogs.
synthetic work has been reported on these novel diene
based quinolizidine alkaloids.
CO2Et
CO2H
CO2Et
(i)
(ii)
CO2Et
CO2H
CO2Et
HO
HN
Bn
H2N
+H3N
Cl-
3
4
(iii)
EtO2C
BnN
H
H
H
H
(vi)
(iv)
(v)
HN
BnN
BnN
O
O
O
O
8
7
6
5
(viii), (ix)
O
Br
O
(vii)
Br
O
O
OH
Br
OH
10
9
(x) (xi)
O
H
H
8
Br
O
(xii)
11
O
O
H
H
N
N
O
12
13
N
O
N
11
O
O
(xiii)
14
(xiv)
H
H
OH
O
1
2
N
N
Scheme 1. Reagents and conditions: (i) SOCl2, EtOH, 90%; (ii) NaOEt, EtOH then PhCHO then NaBH4, 68%; (iii) EtOH, AcOH, reflux, 4 days,
86%; (iv) CH3MgI, THF, rt, 91%; (v) SOCl2, Et3N, THF, )78 ꢁC fi rt, 82%; (vi) Na, NH3(l), EtOH, 81%; (vii) [CpRu(MeCN)3]PF6, SnBr4, LiBr,
acetone, reflux, 43%; (viii) PBr3, ether, )20 ꢁC, 76%; (ix) ethylene glycol, p-TsOH, benzene, reflux, 95%; (x) KN(SiMe3)2, 8, THF, 0 ꢁC fi rt, 47%; (xi)
2 M aq HCl, THF, rt, 95%; (xii) Pd(OAc)2, PPh3, K2CO3, toluene, reflux, 48 h; (xiii) LiAlH4, AlCl3, THF, 62%; (xiv) Dess–Martin periodinane,
CH2Cl2, rt.

P. Armstrong et al. / Tetrahedron Letters 45 (2004) 1627–1630
1629
was 68%. Conversion of secondary amine 4 to d-lactam
5 proved to be much more difficult than the corre-
sponding reaction in the pyrrolidinone series, in which
cyclisation occurred on standing at room temperature.19
Literature precedent indicated that cyclisation to a
d-lactam was acid catalysed23 and experimentation
showed the cyclisation could be effected in 86% yield by
refluxing an ethanolic solution of amine 4, containing a
few drops of acetic acid, for four days. Reaction of ester
5 with a threefold excess of methylmagnesium iodide
gave the tertiary alcohol 6 in 91% yield. Dehydration of
tertiary alcohol 6 using a standard procedure6 gave 1,1-
disubstituted alkene 7 in 82% yield with none of the
regioisomeric dehydration product being observed.
Finally N-debenzylation using sodium in liquid ammo-
nia24 gave lactam 8 in 81% yield. The optical purity of
compound 8 was measured by making diastereomeric
urea derivatives with (R)- and (S)-a-methylbenzylisocy-
anate, and the ee was found to be 86%.25 The advantage
of these urea derivatives is that intramolecular hydrogen
bonding leads to sharp signals in the proton NMR
spectra aiding quantification.
exocyclic double bond of major compound 12 was
confirmed to be Z, as shown (Scheme 1), by NOE dif-
ference spectroscopy. Saturation of proton H-11 gave a
9.7% enhancement to proton H-8 leaving no doubt that
the vinyl bromide had indeed undergone stereochemical
inversion on the formal 6-endo-trig cyclisation. From the
previous model study on the indolizidinone series, for-
mation of double bond isomer 13 was disappointing and
unexpected. It is not clear if it is the more forcing con-
ditions that were responsible for the formation of 13 or
did a conformation difference between the indolizidine
and quinolizidine series determine the final double bond
position. On a second run, under apparently identical
conditions, an additional diene isomer 14 was isolated in
15% yield. Of particular note were the methyl doublet
and the additional triplet in the olefinic region of the
proton NMR spectrum confirming the structure of 14.
However, this compound proved to be very unstable
and on standing in deuteriochloroform for two days it
was quantitatively converted to the alternative enamide
isomer 13. This rapid isomerisation may well explain
why compound 14 was not observed in the first run.
Simultaneous reduction of the amide and the ketone
carbonyl groups of compound 12 gave quinolizidine 1 in
62% yield as a 1:1 mixture of secondary alcohol diaste-
reoisomers. Interestingly, the natural product was iso-
lated as a mixture of diastereoisomers at the secondary
alcohol centre. Finally Dess–Martin oxidation29 gave a
sample of diene 2.
Vinyl bromide 9, containing all the carbon atoms of the
alkylidene sub-unit, was rapidly assembled using the
three component ruthenium catalysed cross coupling
reaction recently developed by Trost.26;27 Addition of
propargyl alcohol to methyl vinyl ketone in the presence
of lithium bromide and 10 mol % [CpRu(MeCN)3]PF6 in
refluxing acetone gave a 10:1 mixture of Z:E-vinyl
bromide isomers from which the pure required Z-isomer
9 could be isolated in 43% yield. The moderate yield
may be attributed to the volatility of both starting
materials and product. Despite the moderate yield, this
reaction allowed rapid stereoselective assembly of the
desired vinyl bromide in 0.5 g batches. Conversion of
alcohol 9 to the corresponding allyl bromide followed by
protection of the ketone as a dioxolane gave 10 in 72%
yield for the two steps. N-Alkylation of lactam 8 with
allylic bromide 10 using a standard procedure28 but
substituting sodium hydride for potassium hexamethyl-
disilazane as base gave the N-alkylated product 11 in
47% yield. Despite extensive experimentation, the yield
for this alkylation could not be improved and we note
that the Z-vinyl bromides are much more problematic
substrates for N-alkylation than the corresponding
E-vinyl bromides. Model studies on the indolizidinone
series revealed that the bulky ketone protecting group
retarded the Heck cyclisation. Therefore the acetal
group was removed to afford ketone 11 in 95% yield.
Comparison of the mass spectral data of the synthetic
materials 1 and 2 with the data reported for the natural
products revealed that the molecular weights were
indeed the same at 235 and 233, respectively. However,
the fragmentation patterns were very different, suggest-
ing that the originally proposed structures for these
alkaloids are incorrect.30
Acknowledgements
We would like to thank DENI (H.M.A.), ESF (S.F.)
and DEL (G.OÕM.) and CSS Ltd for financial support.
References and notes
1. Daly, J.; Myers, C. Science 1967, 156, 970–973.
2. Daly, J.; Garraffo, H.; Spande, T. Amphibian alkaloids. In
The Alkaloids; Cordell, G., Ed.; Academic: New York,
1993; Vol. 43, p 185 (Chapter 3).
3. Tokuyama, T.; Daly, J. W.; Highet, R. J. Tetrahedron
1984, 40, 1183–1190.
4. Jain, P.; Garraffo, H. M.; Spande, T. F.; Yeh, H. J. C.;
Daly, J. W. J. Nat. Prod. 1995, 58, 100–104.
5. Franklin, A. S.; Overman, L. E. Chem. Rev. 1996, 96, 505–
522.
6. Overman, L. E.; Bell, K. L. J. Am. Chem. Soc. 1981, 103,
1851–1853.
We have previously demonstrated that Z-vinyl bromides
undergo cyclisation with inversion of alkene stereo-
chemistry to give indolizidines with the conjugated diene
as the sole product.19 However, the corresponding
chemistry with the quinolizidines proved much more
difficult and the conditions developed for the indoliz-
idines were not effective in the quinolizidine series. Much
more forcing conditions (toluene reflux compared to
acetonitrile reflux) and longer reaction times (48 h,
compared to 16 h) were required. On the first cyclisation
run, two compounds 12 and 13 were isolated in yields of
41% and 17%, respectively. The stereochemistry of the
7. Overman, L. E.; Bell, K. L.; Ito, F. J. Am. Chem. Soc.
1984, 106, 4192–4201.

1630
P. Armstrong et al. / Tetrahedron Letters 45 (2004) 1627–1630
8. Lin, N. H.; Overman, L. E.; Rabinovitz, M. H.; Robinson,
L. A.; Sharp, M. J.; Zablocki, J. J. Am. Chem. Soc. 1996,
118, 9062–9072.
9. Tang, X. Q.; Montgomery, J. J. Am. Chem. Soc. 1999,
121, 6098–6099.
20. McAlonan, H.; Montgomery, D.; Stevenson, P. J. Tetra-
hedron Lett. 1996, 37, 7151–7154.
21. Chauvel, E. N.; Llorenscortes, C.; Coric, P.; Wilk, S.;
Roques, B. P.; Fourniezaluski, M. C. J. Med. Chem. 1994,
37, 2950–2957.
10. Tang, X. Q.; Montgomery, J. J. Am. Chem. Soc. 2000,
122, 6950–6954.
22. Rosowsky, A.; Forsch, R. J. Med. Chem. 1982, 25, 1454–
1459.
11. Trost, B. M.; Scanlan, T. S. J. Am. Chem. Soc. 1989, 111,
4988–4990.
12. Kibayashi, C.; Aoyagi, S. J. Organomet. Chem. 2002, 653,
229–233.
13. Aoyagi, S.; Hasegawa, Y.; Hirashima, S.; Kibayashi, C.
Tetrahedron Lett. 1998, 39, 2149–2152.
14. Aoyagi, S.; Wang, T. C.; Kibayashi, C. J. Am. Chem. Soc.
1992, 114, 10653–10655.
15. Okamoto, S.; Iwakubo, M.; Kobayashi, K.; Sato, F.
J. Am. Chem. Soc. 1997, 119, 6984–6990.
16. Tokuyama, T.; Nishimori, N.; Shimada, A.; Edwards,
M. W.; Daly, J. W. Tetrahedron 1987, 43, 643–652.
17. Santos, L. S.; Pilli, R. A. Tetrahedron Lett. 2001, 42, 6999–
7001.
18. Garraffo, H. M.; Caceres, J.; Daly, J. W.; Spande, T. F.;
Andriamaharavo, N. R.; Andriantsiferana, M. J. Nat.
Prod. 1993, 56, 1016–1038.
23. Kubo, A.; Nakai, T.; Koizumi, Y.; Saito, N.; Mikami,
Y.; Yazawa, K.; Jun, U. Heterocycles 1992, 34, 1201–
1211.
24. Hodgson, S. T.; Hollinshead, D. M.; Ley, S. V. Tetra-
hedron 1985, 41, 5871–5878.
25. Pirkle, W. H.; Robertson, M. R.; Hyun, M. H. J. Org.
Chem. 1984, 49, 2433–2437.
26. Trost, B. M.; Pinkerton, A. B. Angew. Chem., Int. Ed.
2000, 39, 360–362.
27. Trost, B. M.; Pinkerton, A. B. J. Am. Chem. Soc. 2002,
124, 7376–7389.
28. Harkin, S. A.; Singh, O.; Thomas, E. J. J. Chem. Soc.,
Perkin Trans. 1 1984, 1489–1499.
29. Dess, D. B.; Martin, J. C. J. Org. Chem. 1983, 48, 4155–
4156.
30. There is a brief mention in a recent publication that the
structures of diene based homopumiliotoxins may need to
be revised see: Daly, J. W. J. Med. Chem. 2003, 46, 445–
452.
19. Feutren, S.; McAlonan, H.; Montgomery, D.; Stevenson,
P. J. J. Chem. Soc., Perkin Trans. 1 2000, 1129–1137.