General approach to glycosidase inhibitors. Enantioselective synthesis of deoxymannojirimycin and swainsonine

Article abstract of DOI:10.1021/jo048172s

(Chemical Equation Presented) Deoxymannojirimycin (2) and swainsonine (4) have been synthesized from each enantiomer of the same bicyclic carbarnate precursor 7. The key intermediate was prepared by a simple and efficient three-step synthesis involving RCM of the diene 8, which in turn is easily accessible in any configuration from enantiomerically enriched 2,3-epoxy-4-penten-1-ol 9.

Full text of DOI:10.1021/jo048172s

General Approach to Glycosidase
Inhibitors. Enantioselective Synthesis of
Deoxymannojirimycin and Swainsonine
Rube´n Mart´ın, Caterina Murruzzu,
Miquel A. Perica`s, and Antoni Riera*
Unitat de Recerca en S´ıntesi Asime`trica (URSA-PCB),
Parc Cientı´fic de Barcelona, and Departament de Qu´ımica
Orga`nica, Universitat de Barcelona, c/Josep Samitier,
1-5, 08028-Barcelona, Spain
ariera@pcb.ub.es
Received October 15, 2004
FIGURE 1. Representative glycosidase inhibitors with a
polyhydroxylated piperidine or indolizidine structure.
some of them being therapeutically useful substances.3-6
In addition, bicyclic compounds such as swainsonine (4),
2-epi-swainsonine (5), and castanospermine (6) also exhib-
it glycosidase inhibitory properties having been tested
for the treatment of cancer, HIV, or immunological
disorders.7
The absolute configuration of the stereogenic centers
is obviously crucial for their biological activity, and
therefore many stereoselective syntheses of these natural
alkaloids have been described to date.8-10 Moreover, in
efforts to increase the biological activity or/and selectivity
of these compounds, a great number of derivatives,
stereoisomers, and analogues of this family have also
been synthesized.11 Although most of the aforementioned
Deoxymannojirimycin (2) and swainsonine (4) have been syn-
thesized from each enantiomer of the same bicyclic carbam-
ate precursor 7. The key intermediate was prepared by a
simple and efficient three-step synthesis involving RCM of
the diene 8, which in turn is easily accessible in any config-
uration from enantiomerically enriched 2,3-epoxy-4-penten-
1-ol 9.
(3) Therapeutic utility. (a) Johnston, P. S.; Lebovitz, H. E.; Coniff, R.
F.; Simonson, D. C.; Raskin, P.; Munera, C. L. J. Clin. Endocrinol.
Metab. 1998, 83, 1515-1522. (b) Gruters, R. A.; Neefjes, J. J.; Tersmette,
M.; Goede, R. E. Y.; Tulp, A.; Huisman, H. G.; Miedema, F.; Ploegh,
H. L. Nature 1987, 330, 74-76. (c) Tsuruoka, T.; Fukuyasu, H.; Ishii,
M.; Usui, T.; Shibahara, S.; Inouye, S. J. Antibiot. 1996, 49, 155-161.
(4) Hughes, A. B.; Rudge, A. J. Nat. Prod. Rep. 1994, 11, 135-162.
(5) (a)Fellows, L. E.; Bell, E. A.; Lynn, D. G.; Pilkiewicz, F.; Miura,
I.; Nakanishi, K. J. Chem. Soc., Chem. Commun. 1979, 977-978. (b)
Fuhrmann, U.; Bause, E.; Legler, G.; Ploegh, H. Nature 1984, 307,
755-758.
Glycobiology is an emerging research field at the front-
ier of chemistry, enzymology, and biology. Many impor-
tant biological processes in which glycosidases play a cru-
cial role are being uncovered, hence opening the possibil-
ity of finding new therapeutic targets for the treatment
of diseases such as diabetes, AIDS, or cancer.1 Most gly-
cosidase inhibitors share two common structural fea-
tures: (i) a basic nitrogen that, at physiological pH, mimics
the positive charge formed during the hydrolysis of the
glycosidic bond and (ii) an array of hydroxyl groups in a
conformationally restricted motif that selectively fit into
the enzyme site.2 Consequently, the structures of many
glycosidase inhibitors include polyhydroxylated piperi-
dine, pyrrolidine, or indolizidine rings. Representative
examples are deoxynojirimycin (1), deoxymannojirimycin
(2), or deoxygalactostatin (3). In these compounds, refer-
red to as 1-deoxy-azasugars, or iminosugars, the hemiace-
talic function in the monosacharides has been substituted
by an aminomethylenene group. This simple modification
usually improves a compound’s inhibition of glycosidases,
(6) Miyake, Y.; Ebata, M. J. Antibiot. 1987, 40, 122-123.
(7) (a) Oredipe, O. A.; Furbert-Harris, P. M.; Green, W. R.; White,
S. L.; Olden, K.; Laniyan, I.; Parish-Gause, D.; Vaughn, T.; Griffin,
W. M.; Sridhar, R. Pharm. Res. 2003, 47, 69-74. (b) Oredipe, O. A.;
Furbert-harris, P. M.; Laniyan, I.; Green, W. R.; Griffin, W. M.; Sridhar,
R. Cell. Mol. Biol. 2003, 49, 1089-1099. (c) Dennis, J. W. Cancer Res.
1986, 46, 5131-5136. (d) Humphries, M. J.; Matsumoto, K.; White, S.
L.; Molyneux, R. J.; Olden, K. Cancer Res. 1988, 48, 1410-1415.
(8) Selected synthesis of 1-deoxymannojirimycin: (a) McDonnell, C.;
Cronin, L.; O’Brien, J. L.; Murphy, P. V. J. Org. Chem. 2004, 69, 3565-
3568. (b) Singh, O. V.; Han, H. Tetrahedron Lett. 2003, 44, 2387-2391.
(c) Knight, J. G.; Tchabanenko, K. Tetrahedron 2003, 59, 281-286.
(d) Meyers, A. I.; Andres, C. J.; Resek, J. E.; Woodall, C. C.;
McLaughlin, M. A.; Lee, P. H.; Price, D. A. Tetrahedron 1999, 55,
8931-8952. (e) Shirai, M.; Okamoto, S.; Sato, F. Tetrahedron Lett.
1999, 40, 5331-5332. (f) Wu, X.-D.; Khim, S.-K.; Zhang, X.; Ceder-
strom, E. M.; Mariano, P. S. J. Org. Chem. 1998, 63, 841-859. (g) Xu,
Y.-M.; Zhou, W.-S. J. Chem. Soc., Perkin Trans. 1 1997, 741-746. (h)
Baxter, E. W.; Reitz, A. B. J. Org. Chem. 1994, 59, 3175-3185. (i) Park,
K. H.; Yoon, Y. J.; Lee, S. G. J. Chem. Soc., Perkin Trans. 1 1994,
2621-2623. (j) Fleet, G. W. J.; Ramsden, N. G.; Witty, D. R. Tetrahe-
dron 1989, 45, 327-336. (k) Fleet, G. W. J.; Ramsden, N. G.; Witty,
D. R. Tetrahedron 1989, 45, 319-326. (l) Pederson, R. L.; Kim, M.-J.;
Wong, C.-H. Tetrahedron Lett. 1988, 29, 4645-4648. (m) Bernotas, R.
C.; Ganem, B. Tetrahedron Lett. 1985, 26, 1123-1126.
(1) (a) Iminosugars as Glycosidase Inhibitors: Nojirimycin and
Beyond; Stu¨tz, A. E., Ed.; Wiley-VCH: Weinheim, Germany, 1999. (b)
Elbein, A. D. FASEB J. 1991, 5, 3055.
(2) Reviews of glycosidase inhibitors: (a) Look, G. C.; Fotsch, C. H.;
Wong, C. H. Acc. Chem. Res. 1993, 26, 182-190. (b) van den Broek, L.
A. G. M.; Vermaas, D. J.; Heskamp, B. M.; van Boeckel, C. A. A.; Tan,
M. C. A. A.; Bolscher, J. G. M.; Ploegh, H. L.; van Kemenade, F. J.; de
Goede, R. E. Y.; Miedema, F. Recl. Trav. Chim. Pays-Bas 1993, 112,
82-94. (c) Junge, B.; Matzke, M.; Stoltefuss, J. Handb. Exp. Pharma-
col. 1996, 119, 411-482. (d) Bols, M. Acc. Chem. Res. 1998, 31, 1-8.
(9) Selected synthesis of 1-deoxynojirimycin: (a) Comins, D. L.; Fulp,
A. B. Tetrahedron Lett. 2001, 42, 6839-6841. (b) Lindstrom, U. M.;
Somfai, P. Tetrahedron Lett. 1998, 39, 7173-7176. (b) Berger, A.;
Ebner, M.; Stuetz, A. E. Tetrahedron Lett. 1995, 36, 4989-4990. (c)
Ermert, P.; Vasella, A. Helv. Chim. Acta 1991, 74, 2043-2053. (d)
Anzeveno, P. B.; Creemer, L. J. Tetrahedron Lett. 1990, 31, 2085-
2088. (e) Iida, H.; Yamazaki, N.; Kibayashi, C. J. Org. Chem. 1987,
52, 3337-3342.
10.1021/jo048172s CCC: $30.25 © 2005 American Chemical Society
Published on Web 02/17/2005
J. Org. Chem. 2005, 70, 2325-2328
2325

stereochemistry of which would be conveniently set by a
nucleophilic C-2 ring-opening of an enantiomerically
enriched epoxide. Consequently, we planned to start from
the known unsaturated epoxyalcohol 9 and to use allyl
isocyanate as a nucleophilic reagent. We describe here
full details of the straightforward preparation of inter-
mediate 7 as well as its conversion to 1-deoxymannojiri-
mycin (2), swainsonine (4), and epi-swainsonine (5).
Results and Discussion
Our synthesis started with the preparation of the
known epoxy alcohol 9. This epoxide can be obtained in
high enantiomeric purity by Sharpless epoxidation19 of
the readily available (E)-2,4-pentadien-1-ol.20 Although
pure (2S,3S)-2,3-epoxy-4-penten-1-ol (>91% ee) can be
obtained by distillation,21 the crude product from the
epoxidation was used to prepare the carbamate. Alter-
natively, 9 can be obtained in more than 99% ee by Payne
rearrangement of (2R,3S)-1,2-epoxy-4-penten-3-ol, pre-
pared by Sharpless epoxidation of the commercially
available bis allylic alcohol.22
FIGURE 2. General retrosynthetic analysis.
synthetic approaches involve chemical transformations
of monosaccharides, efficient syntheses based in asym-
metric catalysis have also been described. However, a
stereodivergent approach that could lead to different
stereoisomers of monocyclic and bicyclic compounds from
a common precursor is still lacking.
The crude epoxide 9, prepared by either way, was
treated by allyl isocyanate/Et3N in ether at reflux to
provide allyl carbamate 10 in 94% yield (59% overall yield
from 2,4-pentadienol; Scheme 1). The subsequent in-
tramolecular ring opening of 10 required extensive
Some years ago, in a project devoted to the preparation
of natural alkaloids using ring-closing metathesis12 as a
key step,13 we envisaged14 that oxazolidinylpiperidine 7
could be a key intermediate for the synthesis of many
glycosidase inhibitors. The trans stereochemistry be-
tween the nitrogen and the secondary hydroxyl is present
in many compounds such as deoxynojirimycin, deoxym-
annojirimycin, or swainsonine. The synthetic potential
of carbamate 7 for the construction of glycosidase inhibi-
tors has also been recognized by other groups.15-18
Our synthetic approach was based on the preparation
of 7 by RCM of the unsaturated carbamate 8, the
SCHEME 1
(10) (a) Lindsay, K. B.; Pyne, S. G. Aust. J. Chem. 2004, 57, 669-
672. (b) Song, L.; Duesler, E. N.; Mariano, P. S. J. Org. Chem. 2004,
69, 7284-7293. (c) Buschmann, N.; Rueckert, A.; Blechert, S. J. Org.
Chem. 2002, 67, 4325-4329. (d) Lindsay, K. B.; Pyne, S. G. J. Org.
Chem. 2002, 67, 7774-7780. (e) Zhao, H.; Hans, S.; Cheng, X.; Mootoo,
D. R. J. Org. Chem. 2001, 66, 1761-1767. (f) Punniyamurthy, T.; Irie,
R.; Katsuki, T. Chirality 2000, 12, 464-468. (g) De Vicente, J.; Arrayas,
R. G.; Canada, J.; Carretero, J. C. Synlett 2000, 53-56. (h) Mukai, C.;
Sugimoto, Y.-I.; Miyazawa, K.; Yamaguchi, S.; Hanaoka, M. J. Org.
Chem. 1998, 63, 6281-6287. (i) Ferreira, F.; Greck, C.; Genet, J. P.
Bull. Soc. Chim. Fr. 1997, 134, 615-621. (j) Pearson, W. H.; Hembre,
E. J. J. Org. Chem. 1996, 61, 7217-7221. (k) Zhou, W.-S.; Xie, W.-G.;
Lu, Z.-H.; Pan, X.-F. J. Chem. Soc., Perkin Trans. 1 1995, 2599-2604.
(l) Kang, S. H.; Kim, G. T. Tetrahedron Lett. 1995, 36, 5049-5052.
(m) Oishi, T.; Iwakuma, T.; Hirama, M.; Ito, S. Synlett 1995, 404-
406. (n) Zhou, W.-S.; Xie, W.-G.; Lu, Z.-H.; Pan, X.-F. Tetrahedron Lett.
1995, 36, 1291-1294. (o) Naruse, M.; Aoyagi, S.; Kibayashi, C. J. Org.
Chem. 1994, 59, 1358-1364.
experimentation since the standard conditions (NaH/
THF) gave a 1:1 mixture of the desired oxazolidinone 8
and the trans-acylated isomer 8b.23 Other bases such as
ButOK gave only slightly better yields, whereas treat-
ment with Lewis acid catalysts such as LiClO4 or
Ti(iPrO)4 led to decomposition of the starting material.
Ultimately, the use of sodium bis(trimethylsilyl)amide in
ether provided the desired oxazolidinone 8 in 88% yield
with no sign of the isomer 8b in the crude reaction
mixture.24 As expected, ring-closing metathesis on the bis
olefinic compound 8 took place cleanly using 5% mol
Grubbs’s catalyst25 in dichloromethane at room temper-
ature and afforded the target oxazolidinylpiperidine 7 in
(11) (a) Michael, J. P. Nat. Prod. Rep. 2003, 20, 458. (b) El Memr,
A. Tetrahedron 2000, 56, 8579-8629. (c) Carmona, A. T.; Fuentes, J.;
Robina, I.; Rodriguez-Garc´ıa, E.; Demange, R.; Vogel, P.; Winters, A.
L. J. Org. Chem. 2003, 68, 3874-3883.
(12) Reviews on this subject: (a) Felpin, F.-X.; Lebreton, J. Eur. J.
Org. Chem. 2003, 3693-3712. (b) Maier, M. E. Angew. Chem., Int Ed.
2000, 39, 2073-2077. (c) Phillips, A. J.; Abell, A. D. Aldrichimica Acta
1999, 32, 75-89. (d) Pandit, U. K.; Overkleeft, H. S.; Borer, B. C.;
Bieraugel, H. Eur. J. Org. Chem. 1999, 959-968.
(19) (a) Gao, Y.; Hanson, R. M.; Klunder, J. M.; Ko, S. Y.; Masamune,
H.; Sharpless, K. B. J. Am. Chem. Soc. 1987, 109, 5765-5780. (b)
Katsuki, T.; Mart´ın, V. S. Org. React. 1996, 48, 1-299.
(20) Schneider, M. P.; Goldbach, M. J. Am. Chem. Soc. 1980, 102,
6114-6116.
(21) Wershofen, S.; Scharf, H.-D. Synthesis 1988, 854-858.
(22) (a) Ja¨ger, V.; Hummer, W.; Stahl, U.; Gracza, T. Synthesis 1991,
769-776. (b) Ja¨ger, V.; Stahl, U.; Gracza, T.; Hummer, W. Synthesis
1991, 776-782. (c) Ja¨ger, V.; Schroeter, D.; Koppenhoefer, B. Tetra-
hedron 1991, 47, 2195-2210.
(23) Roush, W. R.; Adam, M. A. J. Org. Chem. 1985, 50, 3752-3757.
This methodology, as Roush already pointed out, has the serious draw-
back that the base also catalyzes an acyl migration, yielding a mixture
of oxazolidinones. Other conditions (TMS-Cl, imidazole DMF) lead to
the formation of the carbonate between both alcohols (see ref 14).
(13) (a) Martin, R.; Alco´n, M.; Perica`s, M. A.; Riera, A. J. Org. Chem.
2002, 67, 6896-6901. (b) Ginesta, X.; Perica`s, M. A.; Riera, A.
Tetrahedron Lett. 2002, 43, 779-782.
(14) Preliminary communication: Mart´ın, R.; Moyano, A.; Perica`s,
M. A.; Riera, A. Org. Lett. 2000, 1, 93-95.
(15) Asano, K.; Hakogi, T.; Iwama, S.; Katsumura, S. Chem.
Commun. 1999, 7, 41-42.
(16) Ciufolini, M. A.; Hermann, C. Y. W.; Dong, Q.; Shimizu, T.;
Swaminathan, S.; Xi, N. Synlett 1998, 105-114.
(17) Shirai, M.; Okamoto, S.; Sato, F. Tetrahedron Lett. 1999, 40,
5331-5332.
(18) Al-Rawi, S.; Hinderlich, S.; Reutter. W.; Giannis, A. Angew.
Chem., Int. Ed. 2004, 43, 4366-4370.
2326 J. Org. Chem., Vol. 70, No. 6, 2005

98% yield. The metathesis reaction was also performed
in supercritical CO2 (200 bar and 40 °C) providing an 88%
yield with only 2 mol % standard Grubbs’s catalyst.26
Once the key intermediate (-)-7 was acquired, the
study of its transformation into 1-deoxy-azasugars was
initiated. Katsumura et al.15 already described the trans-
formation of 7 into 1-deoxymannojirimycin 2 and to 1-de-
oxygalactostatin 3 through a tert-butyldimethylsilyl ether
derivative. In our hands, however, the silyl protective
group was too sensitive to the basic hydrolysis conditions
necessary for the deprotection of the carbamate group.
It was thus decided to use benzyl or benzhydryl groups
to protect the secondary hydroxyl functionality. In both
cases, the corresponding ethers 11a and 11b were
obtained in good yield (Scheme 2). The dihydroxylation
of both compounds took place in high diastereoselectivity
anti to the alkoxy group, but as might be anticipated,
the benzhydryl derivative 12b was obtained in a signifi-
cantly better diastereomeric ratio (12:1). Protection of the
vicinal diol in 12b as a dimethyl acetal followed by
chromatography provided diastereomerically pure 13, a
completely protected form of 1-deoxymannojirimycin 2.
Although in acetal 13 all protecting groups can, in
principle, be hydrolyzed by acid treatment to give 2, all
attempts to perform the hydrolysis of all three protecting
groups using HCl failed. On the other hand, basic
hydrolysis led to clean and selective deprotection of the
carbamate, affording 14 in excellent yield. Hydrogenoly-
sis as well as acid hydrolysis of the benzhydryl group in
14 proved to be more difficult than expected. Whereas
Pd/C or Pd(OH)2 catalytic hydrogenation only afforded
starting material, hydrolysis using HCl/MeOH gave the
desired product contaminated by a methyl ether deriva-
tive. Ultimately, it was found that hydrolysis in HCl/THF
cleaved both groups cleanly, affording the hydrochloride
15 (2‚HCl) in almost quantitative yield and high purity.
viously described synthesis. One of the advantages of us-
ing Sharpless epoxidation as a source of chirality is that
both enantiomers are readily accessible. Therefore, (-)-9
was prepared by simply changing the configuration of the
tartrate in the epoxidation step, and (+)-7 was synthesiz-
ed from it following the same reaction sequence. We plan
to build up the five-membered ring by introducing an un-
saturated ester via olefination of aldehyde 19, which in
turn could be prepared via deprotection of the carbamate
and subsequent oxidation of the primary alcohol. The syn
dihydroxylation of the unsaturated ester would control
the stereochemistry of the diol fragment in the cyclic
derivatives. Syn-dihydroxylation of E-20 would lead to a
trans diol, whereas syn-dihydroxylation of Z-20 would
afford the stereochemistry present in swainsonine 4.
The synthetic plan employed required the choice of an
adequate protecting group for the secondary alcohol after
hydrogenation of the double bond in (+)-7. The benzyl
group was chosen for the case at hand since it is more
resistant to hydrolysis than benzhydryl and we hoped
that its bulkiness would not be necessary in the diastereo-
selectivity control of the side chain dihydroxylation. Thus,
hydrogenation of (+)-7 gave alcohol 16 which, after pro-
tection as a benzyl ether gave 17 uneventfully (Scheme 3).
Basic hydrolysis of the carbamate took place smoothly, and
the free amine was protected in situ with a Boc group,
yielding hydroxymethylpiperidine 18 in an overall yield
of 87% from 17. Oxidation of the primary alcohol required
some experimentation. While Swern oxidation provided
aldehyde 19 in only 68% yield, the use of Dess-Martin
periodinane (DMP)27 as an oxidant in basic media afford-
ed 19 in nearly quantitative yield.
SCHEME 3
SCHEME 2
Horner-Emmons olefination of 19 afforded the unsat-
urated ester E-20 in good yield and diastereoselectivity
(86%, 10:1), although these results were improved using
Masamune-Roush conditions28 (94%, 14:1). On the other
hand, the corresponding Z isomer was conveniently ob-
tained using bis(trifluoroethoxy)phosphonate under Still
conditions,29 which gave Z-20 as the major isomer (5:1).
Gratifyingly, dihydroxylation using catalytic OsO4 and
NMO as a stoichiometric oxidant30 of both Z and E
The stereogenic centers in the piperidine ring of swain-
sonine have the opposite configuration of those in manno-
jirimycin. Consequently, their synthesis requires starting
from the enantiomer of the intermediate used in the pre-
(26) (a) Fu¨rstner, A.; Ackermann, L.; Beck, K.; Hori, H.; Koch, D.;
Langemann, K.; Liebl, M.; Six, C.; Leitner, W. J. Am. Chem. Soc. 2001,
123, 9000-9006. (b) Fu¨rstner, A.; Koch, D.; Langemann, K.; Leitner,
W.; Six, C. Angew. Chem., Int. Ed. Engl. 1997, 36, 2466-2469.
(27) Dess, D. B.; Martin, J. C. J. Am. Chem. Soc. 1991, 113, 7277-
7287.
(24) Other reports using NaN(TMS)2 in the ring opening of epoxy-
carbamates: (a) Bernet, B.; Vasella, A. Tetrahedron Lett. 1983, 24,
5491-5494. (b) Li, Y.-L.; Wu, Y.-L. Tetrahedron Lett. 1995, 36, 3875-
3876.
(25) (a) Grubbs, R. H.; Chang, S. Tetrahedron 1998, 54, 4413-4450.
(b) Grubbs, R. H.; Miller, S. J.; Fu, G. C. Acc. Chem. Res. 1995, 28,
446-552. (c) Schuster, M.; Blechert, S. Angew. Chem., Int. Ed. Engl.
1997, 36, 2036-2056. (d) Armstrong, S. K. J. Chem. Soc., Perkin Trans.
1 1998, 371-388. (e) Ivin, K. J. J. Mol. Catal. A: Chem. 1998, 133,
1-16.
(28) Blanchette, M. A.; Choy, W.; Davis, J. T.; Essenfeld, A. P.;
Masamune, S.; Roush, W. R.; Sakai, T. Tetrahedron Lett. 1984, 21,
2183-2186.
(29) Still, W. C.; Gennari, C. Tetrahedron Lett. 1983, 24, 4405-4408.
(30) Van Rheenen, V.; Kelly, R. C.; Cha, D. Y. Tetrahedron. Lett.
1976, 23, 1973-1976.
J. Org. Chem, Vol. 70, No. 6, 2005 2327

SCHEME 4
SCHEME 5
Thus, starting from E-20, syn-dihydroxylation with
osmium tetroxide afforded diol 21 in good yield and total
stereoselectivity (Scheme 4). Careful acid hydrolysis of
the tert-butyl carbamate, followed by treatment with
DIPEA, cleanly afforded the cyclic amide 22, which was
reduced by borane-dimethyl sulfide complex to the
protected 2-epi-swainsonine 23 in good yield.
FIGURE 3. Diastereoselectivity of the dihydroxylation of
unsaturated ester E-20.
unsaturated esters 20 took place with excellent diaste-
reoselectivity, affording syn-diols with 1′S configuration.
According to geometry optimizations at the ab initio
6-31G* level,31 in the most stable conformations of E-20
the alkene and benzyloxy are trans-diaxial (ca. 2.5 kcal/
mol more stable than the diequatorial). However, within
diaxial conformers the energetic difference of the struc-
tures arising from rotation of the C2-C1′ bond was too
small to justify the observed diastereoselectivity. Thus,
we performed a calculation of both transition states (TS-A
and TS-B), leading to the osmate intermediates resulting
of the OsO4 attack to each of the two diastereotopic
double-bond faces (Figure 3). The geometry of these TSs
was optimized at the semiempirical PM3 level, and their
corresponding vibrational analysis showed only one
imaginary frequency. Gratifyingly, the lowest energy
transition state (TS-A, ca. 6 kcal/mol more stable) is the
one leading to the 1′S configuration, as experimentally
observed. In the most stable TS, the approach of OsO4
takes place to an anti-periplanar conformation between
C2-H and C1′-H bonds, whereas in the less stable
transition state (TS-B) osmium approaches to a synclinal
conformation. As it can be readily seen, the lower energy
TS (TS-A) is free of strong steric repulsions because the
carbamate can avoid the pathway of the OsO4. On the
other hand, the osmylation by the opposite face encoun-
ters the steric repulsion of the C3-H bond as well as the
benzyl group. Taking this into account, the sense of the
diastereoselectivity can be easily understood, the most
stable TS-A being the one in which the attack of osmium
is opposite to the benzyloxy group (Figure 3). The same
arguments can be used in the understanding of the
dihydroxylation of Z-20.
To synthesize swainsonine 4, a diastereomerically en-
riched mixture of 20 (Z/E 5/1) was dihydroxylated under
the same conditions described above (Scheme 5). Diol 24
was obtained in excellent diastereoselectivity and then
cyclized as described above for 21 to give 25. Both pro-
ducts 24 and 25 were obtained in good yield but accom-
panied by a small amount of products derived from the
E isomer that could not be removed by chromatography.
Protection with 2,2-dimethoxypropane afforded diastereo-
merically pure 26 in good yield since the trans stereo-
chemistry of the products derived from E-20 impeded its
protection. The cyclic amide 26 was then reduced to ob-
tain the protected swainsonine 27,10a,e,k,o which was spec-
troscopically identical to that described.10,e,k,o Following
their procedure, 27 was easily converted into swainsonine
4 by hydrogenation with PdCl2 and HCl hydrolysis.
In summary, we have developed an efficient procedure
for the preparation of polyhydroxy piperidine and in-
dolizine alkaloids. The key intermediate 7 was prepared
in both enantiomeric forms from unsaturated epoxide 9,
easily obtained by Sharpless epoxidation. The same
intermediate 7 was used as a precursor of monocyclic and
bicyclic important glycosidase inhibitors, thereby il-
lustrating its high synthetic utility.
Acknowledgment. Financial support from CIRIT
(2001SGR-00050) and from DGICIT (BQU2002-02459)
is gratefully acknowledged. R.M. and C.M. thank CIRIT
(Generalitat de Catalunya) for doctoral fellowships.
Supporting Information Available: General experimen-
tal information, preparation of compounds 7, 8, and 10-27
1
and H and 13C NMR spectra of compounds E-20, Z-20, and
21-27. This material is available free of charge via the
Internet at http://pubs.acs.org.
(31) Spartan 02’; Wavefuction, Inc.; Irvine, CA.
JO048172S
2328 J. Org. Chem., Vol. 70, No. 6, 2005