Synthesis of fully substituted polyhydroxylated pyrrolizidines via Cope-House cyclization

Article abstract of DOI:10.1021/ol201749c

Total synthesis of the proposed structure of (-)-hyacinthacine C ₅ and its epimers at C6 and C7 is described. A key step of the synthesis was the construction of the bicyclic pyrrolizidine system by means of a nucleophilic addition of a dithian

Full text of DOI:10.1021/ol201749c

ORGANIC
LETTERS
2011
Vol. 13, No. 16
4414–4417
Synthesis of Fully Substituted
Polyhydroxylated Pyrrolizidines via
CopeꢀHouse Cyclization
Wei Zhang,† Kasumi Sato,‡ Atsushi Kato,‡ Yue-Mei Jia,† Xiang-Guo Hu,†
Francis X. Wilson,§ Renate van Well,§ Graeme Horne,§ George W. J. Fleet,
Robert J. Nash,^ and Chu-Yi Yu*,†
Beijing National Laboratory for Molecular Science (BNLMS), CAS Key Laboratory of
Molecular Recognition and Function, Institute of Chemistry, Chinese Academy of
Sciences, Beijing 100190, China, Department of Hospital Pharmacy, University of
Toyama, 2630 Sugitani, Toyama 930-0194, Japan, Summit PLC, 91, Milton Park,
Abingdon, Oxon OX14 4RY, U.K., Chemistry Research Laboratory, Department of
Chemistry, University of Oxford, Mansfield Road, Oxford, OX1 3TA, U.K., and
Phytoquest Limited, IBERS, Plas Gogerddan, Aberystwyth, Ceredigion SY23 3EB,
Wales, U.K.
yucy@iccas.ac.cn
Received June 29, 2011
ABSTRACT
Total synthesis of the proposed structure of (ꢀ)-hyacinthacine C5 and its epimers at C6 and C7 is described. A key step of the synthesis was the
construction of the bicyclic pyrrolizidine system by means of a nucleophilic addition of a dithiane to a cyclic nitrone followed by a CopeꢀHouse
cyclization.
Hyacinthacines, a series of polyhydroxylated pyrrolizi-
dines possessing a hydroxymethyl substituent at C3, were
isolated from the Hyacinthaceae family of plants
(Hyacinthoides nonscripta,1a Scilla campanulata,1a Muscari
armeniacum,1b Scilla sibirica,1c and Scilla socialis1d) from
1999 to 2007. Nineteen hyacinthacine alkaloids have been
† Chinese Academy of Sciences.
‡ University of Toyama.
§ Summit PLC.
(3) For carbohydrate as chiral pool, see: (a) Rambaud, L.; Compain,
P.; Martin, O. R. Tetrahedron: Asymmetry 2001, 12, 1807. (b) Izquierdo,
I.; Plaza, M. T.; Franco, F. Tetrahedron: Asymmetry 2003, 14, 3933. (c)
Cardona, F.; Faggi, E.; Liguori, F.; Cacciarini, M.; Goti, A. Tetrahedron
University of Oxford.
^ Phytoquest Limited.
(1) (a) Kato, A.; Adachi, I.; Miyauchi, M.; Ikeda, K.; Komae, T.;
Kizu, H.; Kameda, Y.; Watson, A. A.; Nash, R. J.; Wormald, M. R.;
Fleet, G. W. J.; Asano, N. Carbohydr. Res. 1999, 316, 95. (b) 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. (c) Yamashita, T.; Yasuda, K.; Kizu, H.; Kameda, Y.;
Watson, A. A.; Nash, R. J.; Fleet, G. W. J.; Asano, N. J. Nat. Prod. 2002,
65, 1875. (d) Kato, A.; Kato, N.; Adachi, I.; Hollinshead, J.; Fleet,
G. W. J.; Kuriyama, C.; Ikeda, K.; Asano, N.; Nash, R. J. J. Nat. Prod.
2007, 70, 993.
(2) (a) Asano, N. Curr. Top. Med. Chem. 2003, 3, 471. (b) Rempel,
B. P.; Withers, S. G. Glycobiology 2008, 18, 570. (c) Asano, N. In Modern
Alkaloids: Structure, Isolation, Synthesis and Biology; Fattorusso, E.,
Taglialatela-Scafati, O., Eds.; Wiley-VCH Verlag: Weinheim, 2008; pp
111ꢀ138.
ꢀ
Lett. 2003, 44, 2315. (d) Desvergnes, S.; Py, S.; Vallee, Y. J. Org. Chem.
2005, 70, 1459. (e) Chabaud, L.; Landais, Y.; Renaud, P. Org. Lett. 2005,
7, 2587. (f) Izquierdo, I.; Plaza, M. T.; Tamayo, J. A.; Yanez, V.; Lo, Re,
D.; Sanchez-Cantalejo, F. Tetrahedron 2008, 64, 4613. For amino acid
as a chiral pool, see:(g) Dewi-Wulfing, P.; Blechert, S. Eur. J. Org. Chem.
2006, 1852. (h) Sengoku, T.; Satoh, Y.; Takahashi, M.; Yoda, H.
Tetrahedron Lett. 2009, 50, 4937. (i) Ribes, C.; Falomir, E.; Carda,
M.; Marco, J. A. Tetrahedron 2009, 65, 6965. For diethyl tartrate as a
chiral pool, see: (j) Chandrasekhar, S.; Parida, B. B.; Rambabu, C. J.
Org. Chem. 2008, 73, 7826. (k) Liu, W. J.; Ye, J. L.; Huang, P. Q. Org.
Biomol. Chem. 2010, 8, 2085. For tropenone as a chiral pool, see:(l)
Affolter, O.; Baro, A.; Frey, W.; Laschat, S. Tetrahedron 2009, 65, 6626.
For 4-penten-2-ol as a chiral pool, see: (m) Au, C. W. G.; Nash, R. J.;
Pyne, S. G. Chem. Commun. 2010, 46, 713.
r
10.1021/ol201749c
Published on Web 07/28/2011
2011 American Chemical Society

isolated to date, and they were classified into three cate-
gories A1ꢀ7, B1ꢀ7, and C1ꢀ5 according to the oxygenation
pattern; their structures were initially assigned on the
basis of NMR analysis.1 They are inhibitors of R- and
β-glucosidases, β-galactosidases, amyloglucosidases, and
R-L-fucosidases1 and may have chemotherapeutic poten-
tial in the treatment of diabetes II, cancer, and viral
infections.2 Consequently, many efforts for devising gen-
eral strategies for accessing them and their congeners have
been prompted.
The majority of these synthetic methods start with a
chiral pool which have the identical stereocenters to the
alkaloid.3 Meanwhile, de novo approaches, which include
chemoenzymatic procedures utilizinganadolase,4 versatile
routes reliant on the enzymatic desymmetrization of
dihydropyrrole,5 and syntheses through the use of diaster-
eoselective dichloroketeneꢀchiral enol ether cycloaddi-
tion,6 making the hyacinthacines more available.
methods for this class of compound would allow evaluation
of their potential biological activities. Our retrosynthetic
strategy of hyacinthacine C5 is depicted in Scheme 1. Starting
from the cyclic nitrone 59 with three chiral centers, the stereo
center of C7a could be established via diastereoselective
addition of lithio-dithiane 6. The hydroxyl group at C7 could
be derived from dethioketalization and diastereoselective
reduction; the hydroxyl group at C6 could be initially
installed at the dithiane side chain. The chiral center at C5
was commonly introduced by intramolecular SN2 substitu-
tion3m,h or reductive amination3f,i,k or Bruylants alkyla-
tion,6b,9g but all these methods involved cumbersome steps.
Kaliappan’s method10 employing CopeꢀHouse cyclization11
gave excellent diastereoselectivity, and the stereochemical
pattern was identical with that of (ꢀ)-hyacinthacine C5.
Scheme 1. Retrosynthetic Analysis of (ꢀ)-Hyacinthacine C5
Ten hyacinthacines (A1ꢀA3, A5ꢀA7, B1ꢀB3, C2) have
3m
been synthesized and their absolute configurations as-
signed. Unambiguous syntheses of hyacinthacine B7
and C33h have shown that the initially proposed structures
were incorrect; the 1H and 13C NMR spectral data of the
synthetic and isolated compounds were not consistent.
There are no reports of the synthesis of hyacinthacines
7
C1, C4 and C5 with substituents at each carbon of the
pyrrolizidine nucleus so that their absolute configurations
have yet to be determined; only a few analogues have been
synthesized.8 This series of pyrrolizidines has four hydro-
xyl groups, one hydroxymethyl group and one methyl
group attached to each of the seven adjacent chiral centers
(Figure 1).
Since cyclic nitrone 5 could be easily prepared on a large
scale according to our improved approach,9c we investi-
gated the addition of 2-lithio-1, 3-dithiane derivatives to
nitrone 5. As a classical umpolung synthon, dithiane has
been widely applied in reversing the reactivity of carbonyl
groups.12 Although additions of 2-lithio-1,3-dithiane
(9) For the synthesis of nitrone 5, see: (a) Gurjar, M. K.; Borhade,
R. G.; Puranik, V. G.; Ramana, C. V. Tetrahedron Lett. 2006, 47, 6979.
(b) Yu, C. Y.; Huang, M. H. Org. Lett. 2006, 8, 3021. (c) Wang, W. B.;
Huang, M. H.; Li, Y. X.; Rui, P. X.; Hu, X. G.; Zhang, W.; Su, J. K.;
Zhang, Z. L.; Zhu, J. S.; Xu, W. H.; Xie, X. Q.; Jia, Y. M.; Yu, C. Y.
Synlett 2010, 488. For the application of cyclic nitrone, see:(d) Tsou,
E. L.; Chen, S. Y.; Yang, M. H.; Wang, S. C.; Cheng, T. R. R.; Cheng,
W. C. Bioorg. Med. Chem. 2008, 16, 10198. (e) Tsou, E. L.; Yeh, Y. T.;
Liang, P. H.; Cheng, W. C. Tetrahedron 2009, 65, 93. (f) Su, J. K.; Jia,
Y. M.; He, R. R.; Rui, P. X.; Han, N. Y.; He, X. H.; Xiang, J. F.; Chen,
X.; Zhu, J. H.; Yu, C. Y. Synlett 2010, 1609. (g) Hu, X. G.; Jia, Y. M.;
Xiang, J. F.; Yu, C. Y. Synlett 2010, 982. For recent reviews of cyclic
nitrone, see:(h) Revuelta, J.; Cicchi, S.; Goti, A.; Brandi, A. Synthesis
2007, 485. (i) Brandi, A.; Cardona, F.; Cicchi, S.; Cordero, F. M.; Goti,
A. Chem.;Eur. J. 2009, 15, 7808.
Figure 1. Examples of fully substituted pyrrolizidines.
These enticing structures pose a great challenge for their
total syntheses. The development of general synthetic
(10) Kaliappan, K. P.; Das, P. Synlett 2008, 841.
(11) For recent reviews, see: (a) Cooper, N. J.; Knight, D. W.
Tetrahedron 2004, 60, 243. (b) Bainbridge, N. P.; Currie, A. C.; Cooper,
N. J.; Muir, J. C.; Knight, D. W.; Walton, J. M. Tetrahedron Lett. 2007,
48, 7782. For application in iminosugar syntheses, see: (c) Palmer,
A. M.; Jager, V. Eur. J. Org. Chem. 2001, 2547. (d) Palmer, A. M.; Jager,
V. Eur. J. Org. Chem. 2001, 1293. (e) Palmer, A. M.; Jager, V. Synlett
2000, 1405. (f) Jager, V.; Bierer, L.; Dong, H. Q.; Palmer, A. M.; Shaw,
D.; Frey, W. J. Heterocycl. Chem. 2000, 37, 455.
(12) Smith, A. B.; Adams, C. M. Acc. Chem. Res. 2004, 37, 365.
(13) (a) Furneaux, R. H.; Schramm, V. L.; Tyler, P. C. Bioorg. Med.
Chem. 1999, 7, 2599. (b) Davis, F. A.; Ramachandar, T.; Liu, H. Org.
Lett. 2004, 6, 3393. (c) Xu, X.; Liu, J. Y.; Chen, D. J.; Timmons, C.; Li,
G. G. Eur. J. Org. Chem. 2005, 1805.
(4) Calveras, J.; Casas, J.; Parella, T.; Joglar, J.; Clapes, P. Adv.
Synth. Catal. 2007, 349, 1661.
(5) Donohoe, T. J.; Thomas, R. E.; Cheeseman, M. D.; Rigby, C. L.;
Bhalay, G.; Linney, I. D. Org. Lett. 2008, 10, 3615.
(6) (a) Reddy, P. V.; Veyron, A.; Koos, P.; Bayle, A.; Greene, A. E.;
Delair, P. Org. Biomol. Chem. 2008, 6, 1170. (b) Reddy, P. V.; Koos, P.;
Veyron, A.; Greene, A. E.; Delair, P. Synlett. 2009, 1141.
(7) Hyacinthacine C4 has been determined to be the same structure as
hyacinthacine C1 (see ref 1d).
ꢀ
(8) (a) Tamayo, J. A.; Franco, F.; Sanchez-Cantalejo, F. Tetrahedron
2010, 66, 7262. (b) Yu, C.; Gao, H. CN200610113357, 2006.
Org. Lett., Vol. 13, No. 16, 2011
4415

derivatives to imines are well documented,13 the addition
of 2-lithio-1,3-dithiane derivatives to nitrones has not been
reported previously.
Scheme 2. Construction of the Pyrrolizidine Skeleton
We first investigated the model reaction of dithiane 7a
and nitrone 5. The reaction was performed at ꢀ30 ꢀC
in THF with 1.5 equivalents of dithiane. The adduct 8a
was obtained in 56% yield with excellent diastereoselec-
tivity (>95:5) (Table 1, entry 1), together with unreacted
nitrone 5. Addition of a second equivalent of 7a gave 8a
in good yield (80%) (Table 1, entry 2) with nitrone com-
pletely consumed. Addition of TMEDA enhanced the
nucleophilicity of the dithiane anion and increased the
yield to 86% (Table 1, entry 3). Treatment of 7b and 7c
with nitrone 5 under the above conditions gave analogues
of 8 in good yields with excellent diastereoselectivity
(Table 1, entries 4 and 5). The relative configuration of
the newly formed stereocenters was determined by analysis
of the NOESY spectrum.
Table 1. Nucleophilic Addition of 2-Lithio-1,3-dithiane to Ni-
trone 5
Attempted hydrolysis of the dithioketal 12 by several
methods including the Stork protocol16 [e.g., bis(tri-
fluoroacetoxy)iodobenzene] failed to give the ketone 14.
17
NBS/AgNO3 gave a relatively good result but did not
allow the purification of the ketone 14. However, addition
of NaBH4 to the crude product from the NBS/AgNO3
treatment formed diol 15 as a single isomer in 31% yield,
the structure of which was firmly established by X-ray
crystallographic analysis (Scheme 3 and the Supporting
Information). Subsequent removal of the benzyl protect-
ing groups by hydrogenolysis afforded the final product
(ꢀ)-hyacinthacine C5 2.
entry dithiane
R
dithiane (equiv) additive yielda (%) drb
1
2
3
4
5
7a
7a
7a
7b
7c
H
H
H
1.5
3
56
80
86
95
63
>95:5
>95:5
>95:5
>95:5
>95:5
3
TMEDA
TMEDA
TMEDA
phenyl
butyl
3
3
1
a Isolated yield after column chromatography. b Determined by H
NMR analysis of a crude product.
Scheme 3. Synthesis of (ꢀ)-Hyacinthacine C5 (2)
Following the successful model reactions, we turned
our attention to the synthesis of hyacinthacine C5. The
additionofracemic dilithiospecies6, preparedfrom 1-(1,3-
dithian-2-yl)propan-1-ol,14 to nitrone 5 afforded hydro-
xylamine 9 which was subjected to CopeꢀHouse cycliza-
tion to give pyrrolizidine N-oxides 10 and 11 as the only
detectable products in 55% isolated yield in a ratio of 1:1.15
During the CopeꢀHouse cyclization, it was found that
some of the unsaturated hydroxylamine 9 remained. Re-
fluxing the reaction mixture in CHCl3 did not improve the
yield of products but resulted in decomposition of the
hydroxylamine. Reduction of the cyclized products by a
ZnꢀHOAc system gave the pyrrolizidines 12 and 13 in
excellent yields; X-ray crystallographic analysis confirmed
their relative configurations (Scheme 2 and the Supporting
Information).
Protection of the hydroxyl group in 12 with MOMCl
and unmasking of the carbonyl group with NBS/AgNO3
gave the ketone 17 in 61% yield. Reduction of the ketone
17withNaBH4 affordeda mixture oftwo diastereoisomers
in a ratio of 5:1. The selectivity resulted from the favored
formation of 18 by hydride attack from the less hindered side;
L-Selectride, a bulky, noncoordinating alkylborohydride,
(14) Annunziata, R.; Cinquini, M.; Cozzi, F.; Raimondi, L. Tetra-
hedron 1988, 44, 4645.
(15) It was necessary to leave the hydroxyl group of the dithiane
unprotected, since treatment of 2-(1-(benzyloxy)allyl)-1,3-dithiane with
n-BuLi caused elimination of the benzyloxy group.
(16) (a) Stork, G.; Zhao, K. Tetrahedron Lett. 1989, 30, 287. (b)
Fleming, F. F.; Funk, L.; Altundas, R.; Tu, Y. J. Org. Chem. 2001, 66,
6502.
(17) Corey, E. J.; Erickson, B. W. J. Org. Chem. 1971, 36, 3553.
4416
Org. Lett., Vol. 13, No. 16, 2011

cleanly gave 18 as a single stereoisomer. Finally, after MOM
deprotection, hydrogenolysis of diol 19 gave 7-epi-(ꢀ)-
hyacinthacine C5 3 in 91% yield (Scheme 4); no correlation
was observed between H-5 and H-7 in the NOESY NMR
analysis of 3 (see the Supporting Information). The minor
product from the reduction of the ketone 17 was subjected
to MOM deprotection, and the diol thus yielded had an
identical 1H and 13C NMR spectra to 15.
Scheme 5. Synthesis of 6-epi-(ꢀ)-Hyacinthacine C5 (4)
Scheme 4. Synthesis of 7-epi-(ꢀ)-Hyacinthacine C5 (3)
of R-glucosidase from rat intestinal maltase; IC50 = 64.2
μM of R-glucosidase from rice).
In summary, the total synthesis of fully substituted
pyrrolizidines with the structure originally proposed for
(ꢀ)-hyacinthacine C5 2, and of the C6 4 and C7 3 epimers,
was achieved; the key step was a CopeꢀHouse cyclization
of the adduct of a lithio-dithiane addition to a cyclic
nitrone. The inconsistency in spectral data between syn-
thetic compounds and the natural product proposed as 2
indicates further work is necessary to determine the actual
structure of the isolated natural product. This concise
synthetic strategy provides a general approach to this class
of compounds and allows evaluation of their structureꢀ
activity relationship.
6-epi-(-)-Hyacinthacine C5 4 was prepared in an analo-
gous route from 13 (Scheme 5). The results were similar
except for the reduction of ketone 21. In this reaction, both
NaBH4 and L-Selectride gave diastereoselective reduction
to 22 as the sole product, the structure of which was firmly
established by X-ray crystallographic analysis (see the
Supporting Information). Removal of the remaining
O-MOM and O-benzyl protecting groups yielded 6-epi-(ꢀ)-
Acknowledgment. This work is supported by Summit
PLC (UK), The National Basic Research Program of
China (No. 2009CB526511 and No. 2011CB808603),
The Ministry of Science and Technology, and the Ministry
of Health of the P.R. China (No. 2009ZX09501-006).
1
hyacinthacine C5 4 (Scheme 4). None of the H and 13C
NMR spectra of the three final products 2ꢀ4matched those
reported for the natural (þ)-hyacinthacine C5.1d Therefore,
revision of the stereochemical assignment of the natural
product may be required.
(ꢀ)-Hyacinthacine C5 2 and its C6, C7 epimers (4 and 3)
were assayed as potential glycosidase inhibitors of a range
of enzymes (see the Supporting Information). Compound 4
showed weak inhibition of R-glucosidases (IC50 = 58.5 μM
Supporting Information Available. Experimental pro-
1
cedures, characterization data, copies of H NMR and
13C NMR spectra, and crystallographic information files.
This material is available free of charge via the Internet at
http://pubs.acs.org.
Org. Lett., Vol. 13, No. 16, 2011
4417