Total synthesis of hyacinthacine A1 and 3-epi-hyacinthacine A1

Article abstract of DOI:10.1021/ol050713s

(Chemical Equation Presented) Total synthesis of hyacinthacine A ₁ and its epimer at C3 is described. The synthesis includes a stereocontrolled carboazidation of a chiral allylsilane as a key step. C-Si bond oxidation and reduction of the azide

Full text of DOI:10.1021/ol050713s

ORGANIC
LETTERS
2005
Vol. 7, No. 13
2587-2590
Total Synthesis of Hyacinthacine A1 and
3-epi-Hyacinthacine A1
Laurent Chabaud,† Yannick Landais,*,† and Philippe Renaud*,‡
Laboratoire de Chimie Organique et Organome´tallique, UniVersite´ Bordeaux-I, 351,
Cours de la Libe´ration, F-33405 Talence Cedex, France, and Department of Chemistry
and Biochemistry, UniVersity of Berne, Freiestrasse 3, CH-3000 Berne 9, Switzerland
y.landais@lcoo.u-bordeaux.fr; philippe.renaud@ioc.unibe.ch
Received April 4, 2005
ABSTRACT
Total synthesis of hyacinthacine A1 and its epimer at C3 is described. The synthesis includes a stereocontrolled carboazidation of a chiral
allylsilane as a key step. C Si bond oxidation and reduction of the azide, with ring-closure, complete the total synthesis, which establishes
−
the absolute configuration of 3.
Polyhydroxylated pyrrolizidine alkaloids possessing a hy-
droxymethyl substituent at C3 are relatively rare in nature.
Several members of this class of alkaloids, including
australine 1 and alexine 2, have been isolated from legumi-
nosae (Castanospermum and Alexa) and exhibit promising
biological activity1 (e.g., 1 displays antiviral and anti-HIV
activity).2 More recently, Asano et al. reported the isolation,
from an extract of Muscari Armeniacum, of hyacinthacines
A1 (3), A2 (4), A3, and B3, new potent inhibitors of glycosi-
dase.3 The absolute configuration of hyacinthacine A1 remains
unknown. It was assigned structure 3 and was shown to be
an effective inhibitor of rat intestinal lactase (IC50 4.4 µM).
Scheme 1. Structures of Polyhydroxylated Pyrrolizidines
While several total syntheses of hyacinthacine A2 (4) have
been recently completed,4 the synthesis of 3 has not been
achieved to date.5 We report here the first total synthesis of
hyacinthacine A1 and one of its stereoisomers, thus establish-
ing the absolute configuration of the natural product. Our
† Universite´ Bordeaux-I.
‡ University of Berne.
(1) (a) Nash, R. J.; Fellows, L. E.; Dring, J. V.; Fleet, G. W. J.; Derome,
A. E.; Hamor, T. A.; Scofield, A. M.; Watkin, D. J. Tetrahedron Lett. 1988,
29, 2487-2490. (b) Molyneux, R. J.; Benson, M.; Wong, R. Y.; Tropea, J.
E.; Elbein, A. D. J. Nat. Prod. 1988, 51, 1198-1206. (c) Winchester, B.;
Al Dahler, S.; Carpenter, N. C.; Cenci Di Bello, I.; Choi, S. S.; Fairbanks,
A. J.; Fleet, G. W. J. Biochem. J. 1993, 290, 743-749.
(2) Taylor, D. L.; Nash, R.; Fellows, L. E.; Kang, M. S.; Tyms, A. S.
AntiViral Chem. Chemother. 1992, 3, 273-277.
(4) (a) Rambaud, L.; Compain, P.; Martin, O. R. Tetrahedron: Asym-
metry 2001, 12, 1807-1809. (b) Izquierdo, I.; Plaza, M. T.; Franco, F.
Tetrahedron: Asymmetry 2003, 14, 3933-3935. (c) Cardona, F.; Faggi,
E.; Liguori, F.; Cacciarini, M.; Goti, A. Tetrahedron Lett. 2003, 44, 2315-
2318. (d) Desvergnes, S.; Py, S.; Valle´e, Y. J. Org. Chem. 2005, 70, 1459-
1462.
(3) Asano, N.; Kuroi, H.; Ikeda, K.; Kizu, H.; Kameda, Y.; Kato, A.;
Adachi, I.; Watson, A. A.; Nash, R. J.; Fleet, G. W. J. Tetrahedron:
Asymmetry 2000, 11, 1-8.
(5) For an approach, see: Toyao, A.; Tamura, O.; Takagi, H.; Ishibashi,
H. Synlett 2003, 35-38.
10.1021/ol050713s CCC: $30.25
© 2005 American Chemical Society
Published on Web 05/21/2005

strategy is based on a free-radical carboazidation6 of a chiral
allylsilane as a key step,7 the silicon group being used both
to ensure a high level of 1,2-stereocontrol8 and as a latent
hydroxy group.9
As the absolute configuration of natural hyacinthacine A1
was unknown, we envisioned two different routes starting
from diastereomeric allylsilanes 6 and 7 that could lead to
both enantiomers of 3 (Scheme 2). Disconnection of 3 and
The scope of the carboazidation reaction was studied using
enantiomerically pure allylsilanes 10a-c, easily obtained
through Roush allylation of aldehyde 9.11 We first observed
that better yields were obtained when the alcohol function
â to silicon (C2) was protected. A rapid survey of the reaction
conditions and variation of the protective group was thus
carried out, the results of which are summarized in Table 1.
Table 1. Free-Radical Carboazidation of Allylsilanes 10a-c
Scheme 2. Retrosynthetic Analysis
olefin
(equiv)
xanthate
(equiv)
t
yield
(%)c
entry
(°C)
syn/anti
1
2
3
4
5
10a (2)
10a (2)
10a (1)
10b (2)
10c (2)
1
1
2
1
1
80
60
60
60
60
84:16a
85:15a,b
70:30a
80:20a
61:39a
74
84
87
30
52
1
a Ratio estimated from H NMR of the crude reaction mixture. b Ratio
estimated from 28Si NMR of the crude reaction mixture. c Isolated yield.
The acetate group afforded the best results, with the syn-
carboazidation product 11a obtained as the major isomer with
reasonable diastereocontrol (Table 1, entries 1 and 2).12
Interestingly, lowering the temperature led to no change in
the diastereocontrol but gave consistently better yields,
indicating that some of the carboazidation product may be
decomposed at 80 °C (Table 1, entry 2). The reaction is
generally best carried out using 2 equiv of allylsilane for 1
equiv of xanthate, the excess of allylsilane being recovered
by chromatography. Reversal of the ratio also led to good
yield, albeit with slightly lower diastereocontrol (Table 1,
entry 3).
The use of a MOM ether (i.e., 10c) gave modest diaste-
reocontrol and lower yield, probably because of hydrogen
abstraction on the MOM group (vide infra). The benzoate
10b, which was difficult to isolate in pure form, also afforded
a small amount of carboazidation product. Finally, the use
of 3-pyridylsulfonyl azide (PyrSO2N3) instead of PhSO2N3
made the purification of the products easier and provided
higher yields.13 The study was next extended to the car-
boazidation of the acetate-protected anti/syn allylsilane 12.
Surprisingly, reaction under the above optimized conditions
led to azide 14 as a single isomer (stereochemistry not
determined at the anomeric center) and no trace of the
ent-3 thus reveals that both rings should be formed late in
the synthesis, with the ring closure occurring from intermedi-
ates 4 and 5, having the four contiguous stereogenic centers
correctly set up. The configuration of the C3 center would
be imposed by the use of aldehyde 9, readily available from
D-mannitol.10 Intermediates 4 and 5 would be accessible
through carboazidation of allylsilanes 6 and 7, followed by
Tamao-Fleming oxidation of the PhMe2Si group with
retention of configuration. Reduction of the azido group of
4 or 5 with concomitant ring closure would then lead to 3
or ent-3.6b,c While ring closure from mesylate 5 is expected
to occur with inversion of configuration at C3, leading to
ent-3, the synthesis of 3 from 4 would require a double
inversion (retention) at C3. This study showed that a careful
choice of the alcohol protective groups and the inversion of
the C3 configuration prior to ring closure were critical for
completion of the total synthesis of 3.
(6) (a) Ollivier, C.; Renaud, P. J. Am. Chem. Soc. 2001, 123, 4717-
4727. (b) Ollivier, C.; Renaud, P.; Penchaud, P. Angew. Chem., Int. Ed.
2002, 41, 3460-3462. (c) Panchaud, P.; Ollivier, C.; Renaud, P.; Zigmantas,
S. J. Org. Chem. 2004, 69, 2755-2759.
(7) Chabaud, L.; Landais, Y.; Renaud, P. Org. Lett. 2002, 4, 4257-
4260.
(8) Chabaud, L.; James, P.; Landais, Y. Eur. J. Org. Chem. 2004, 3173-
(11) Roush, W. R.; Grover, P. T. Tetrahedron 1992, 48, 1981-1998.
(12) Relative configuration of our products were determined through a
stereospecific fluoride-mediated â-elimination of the â-silyl azide; see: (a)
Masterson, D. S.; Porter, N. D. Org. Lett. 2002, 4, 4253-4256. (b) Chabaud,
L.; Landais, Y. Tetrahedron Lett. 2003, 44, 6995-6998.
3199.
(9) (a) Fleming, I. Chemtracts: Org. Chem. 1996, 1-64. (b) Landais,
Y.; Jones, G. Tetrahedron 1996, 52, 7599-7662. (c) Tamao, K. In AdVances
in Silicon Chemistry; Jai Press Inc.: Greenwich, 1996; Vol. 3, pp 1-62.
(10) Schmid, C. R.; Bryant, J. D. Org. Synth. 1995, 72, 6-13.
(13) Panchaud, P.; Renaud, P. AdV. Synth. Catal. 2004, 34, 925-928.
2588
Org. Lett., Vol. 7, No. 13, 2005

Scheme 3. Carboazidation of Allylsilane 12; 1,5-Hydrogen
Scheme 4. Approach toward 3-epi-Hyacinthacine A1
Atom Transfer
â-azidosilane 13 (Scheme 3). Azide 14 probably results from
a 1,5-hydrogen atom transfer,14 occurring after addition of
the xanthate onto the allylsilane. The radical I, â to the silicon
group, is then suitably located to abstract a hydrogen on the
dioxolane ring to generate a more stable radical II, which
then reacts with PyrSO2N3 to produce 14. Increasing the
amount of sulfonyl azide up to 10 equiv finally led to a 7:3
13/14 ratio (28Si NMR), with â-azidosilane 13 formed in 57%
as an inseparable 8:2 mixture of syn/anti diastereomers.
Conformational effects are likely critical and may explain
the different behavior of allylsilane 12 as compared to its
isomer 10a (vide supra). As a comparison, treatment of the
triacetate analogue of 12 under similar carboazidation
conditions led to the expected â-azidosilane in 80% yield
and a 73:27 syn/anti ratio.
used to test the azide reduction step, which was expected to
provide the pyrrolizidine skeleton in a single step through
cyclization-lactamization. Reduction of the azide using
indium17 led efficiently to the monocylic amine 18, which
resisted further lactamization. As expected, the formation of
the pyrrolidine ring had occurred with inversion of config-
uration at C3. Attempts to carry out the second ring closure
unfortunately failed, whatever the conditions (vide infra).
Steric interactions between the ester function and the CH2-
OTBDMS group (endo) and/or the acetonide probably
prevent the attack of the amino group on the ester function
in a 5-exo-trig fashion.
On the basis of these results, the total synthesis was
pursued with the syn-azidosilane 11a, possessing the four
required contiguous stereogenic centers. Installation of the
OH group at C1 was carried out through the Tamao-Fleming
oxidation of the silicon group. While this oxidation appeared
challenging at first because of the presence of a â-acetox-
ysilane and a â-azidosilane, prone to elimination, the
oxidation finally occurred smoothly using a buffered medium
(AcONa) (Scheme 4).15 The alcohol was obtained in good
yield as a mixture of acetate 15a,b as a result of 1,2-acyl
migration. The acetates were saponified using a DOWEX
1-X-10 ion-exchange resin, leading to a diol that was directly
protected as an acetonide. Transesterification of the ethyl
ester into a methyl ester occurred concomitantly. The
selective deprotection of the bis-acetonide required much
effort, and the most satisfying conditions were those employ-
ing Zn(NO3)2 at 50 °C, which provided diol 16 in a corrected
75% yield (56% conversion).16 Protection of the primary
alcohol as a tert-butyldimethylsilyl (TBDMS) ether, followed
by mesylation, then provided the intermediate 17. This was
It was then decided to carry out cyclization using instead
a 5-exo-tet process. The ester function in 16 was thus reduced
into an alcohol and both hydroxyl groups converted into the
bis-mesylate 19 (Scheme 5). In contrast with the cyclization
Scheme 5. Synthesis of 3-epi-Hyacinthacine A1, 21
(14) For a review on hydrogen atom abstraction, see: Feray, L.;
Kuznetsov, N.; Renaud, P. In Radicals in Organic Synthesis; Renaud, P.,
Sibi, M. P., Eds; Wiley-VCH: Weinheim, 2001; Vol. 2, p 246.
(15) (a) Fleming, I.; Henning, R.; Parker, D. C.; Plaut, H. E.; Sanderson,
P. E. J. J. Chem. Soc., Perkin Trans. 1 1995, 317-337. (b) Panek, J. S.;
Zhang, J. J. Org. Chem. 1993, 58, 294-296. The use of Hg(OAc)2/AcOOH
conditions led to a complex mixture.
of 17, reduction of the azido group in 19 with indium or
PPh3-H2O18 led to poor results. Cyclization was finally
(17) Vidyasagar Reddy, G.; Venkat Rao, G.; Iyengar, D. S. Tetrahedron
Lett. 1999, 40, 3937-3938.
(16) Vijayasaradhi, S.; Singh, J.; Aidhen, I. J. Synlett 2000, 110-112.
Org. Lett., Vol. 7, No. 13, 2005
2589

successfully carried out using H2/Pd-C,19 leading to the
tricyclic system 20, from which acetonide and TBDMS
groups were then removed under acidic conditions. Purifica-
tion of the chlorhydrate through a basic DOWEX column
eventually provided 3-epi-hyacinthacine A1 (21), thus ob-
tained in 13 steps and 12% overall yield from allylsilane 8.
As mentioned above, the synthesis of natural hyacinthacine
A1 using this strategy entailed that a double inversion at C3
was performed on the carboazidation product 11a or one of
its derivatives (i.e., 16 or 17). This was eventually achieved
by converting the mesylate 17 into the corresponding epoxide
22 (Scheme 6).20 Treatment of 17 with a fluoride source led
indicates that the steric hindrance of the CH2OR group (exo
in 23) and the relative configuration at C3 influence the rate
of lactamization. It is also noticeable that no trace of 6-endo
product could be detected. Reduction of the lactam using
LiAlH4, followed by deprotection as above provided hya-
cinthacine A1 (3), the optical rotation of which matches that
of the natural product, demonstrating that the structure
attributed by Asano et al. to natural 3 was correct. Compound
3 was thus obtained in 13 steps and 8% overall yield from
allylsilane 8.
In summary, we report here on the total synthesis of
hyacinthacine A1 (3), a potent inhibitor of glycosidase, and
one of its epimers 21, thus establishing the absolute config-
uration of the natural product. In the course of this study,
we also faced for the first time a competitive [1,5]-hydrogen
transfer in carboazidation reactions of olefins. These results
underline the influence of conformational effects in free-
radical transformation and more particularly in this carboazi-
dation process. It may also be added that this approach is
not restricted to pyrrolizidines having an anti C1/C2 relative
configuration as the syn-isomers are also available, using an
alternative approach.22 Analogues of 3 may be thus at hand
by simply changing the relative configuration of the starting
allylsilane 6 or 7 (Scheme 2). Finally, this work further
demonstrates the utility of the silyl group, first in the control
of the stereochemistry in radical processes8,23 and as a masked
hydroxy group.
Scheme 6. Synthesis of Hyacinthacine A1, 3
Acknowledgment. This work was supported in France
by CNRS, MENRT (grant to L.C.) and the Institut Univer-
sitaire de France and in Switzerland by the State Secretariat
for Education and Research (SER) and by the Swiss National
Science Foundation (Grant 20-103627). We gratefully ac-
knowledge Dr. F.-X. Felpin and Dr. F. Robert for helpful
discussions and COST-D28 for financial support.
to the alcohol deprotection which was followed by the
cyclization in the presence of a base,21 providing 22 with
inversion of configuration at C3. Reduction of the azide with
H2-Pd/C then afforded the amine, which on treatment with
Et3N led to 23 by a subsequent cyclization-lactamization.
This result contrasts with the behavior of 17 (Scheme 4) and
Supporting Information Available: Representative ex-
perimental procedures including product characterization.
This material is available free of charge via the Internet at
http://pubs.acs.org.
(18) Vaultier, M.; Knouzi, N.; Carrie´, R. Tetrahedron Lett. 1983, 24,
763-764.
OL050713S
(19) Cossy, J.; Willis, C.; Bellosta, V.; Saint-Jalmes, L. Synthesis 2002,
951-957.
(22) (a) Yamamoto, Y.; Saito, Y.; Maruyama, K. J. Chem. Soc., Chem.
Commun. 1982, 1326-1328. (b) Yamamoto, Y.; Yatagai, H.; Ishihara, Y.;
Maeda, N.; Maruyama, K. Tetrahedron 1984, 40, 2239-2246.
(20) Molander, G. A.; Swallow, S. J. Org. Chem. 1994, 59, 7148-7151.
(21) Formation of the epoxide occurs spontaneously after deprotection
and is completed through addition of K2CO3.
(23) Landais, Y. C. R. Acad. Sci., Ser. IIc: Chem. 2005, 8, 823-832.
2590
Org. Lett., Vol. 7, No. 13, 2005