In a first approach, 5-azido-5-deoxy-1,2-O-isopropylidene-
R-D-glucofuranose 514 was considered as a suitable starting
material. Attempts to oxidize the primary hydroxyl into a
carboxylate group resulted, however, in concomitant 6,3-lactone
formation. To avoid this unwanted side reaction, the known 6-O-
tetrahydropyranyl derivative 614 was transformed into the
corresponding 3-O-acetyl derivative 7, which through a reaction
sequence involving acid hydrolysis of the tetrahydropyranyl
group, TEMPO oxidation of alcohol 8,15 and in situ esterification
of the resulting carboxylic acid with methanol and 2,6-
dichlorobenzoyl chloride (DCBzCl)16 afforded the R-azido
methyl ester 9. Catalytic hydrogenation of 9 to give the
corresponding R-amino ester intermediate proved to be trouble-
some due to O f N acetyl migration and 6,3-lactone formation,
as seen from a mass spectrum of the reaction mixture. To
minimize this problem the crude reduction mixture was reacted
with butyl isothiocyanate in pyridine in the presence of
triethylamine. The resulting thiourea derivative (10) underwent
spontaneous Edman’s-like reaction with formation of the
thiohydantoin heterocycle (f 11), followed by fast elimination
to the corresponding R,ꢀ-unsaturated carbonyl derivative (12),
cleavage of the acetonide group, and rearrangement of the
furanose pseudo-C-nucleoside intermediate into the target
indolizidine 13 (Scheme 1).
This tandem transformation is remarkable: although formation
of exocyclic double bonds in thiohydantoins bearing hydroxy-
alkyl substituents is well-documented,17 acetal cleavage under
basic conditions was unexpected. Most probably, the anomeric
position is activated intramolecularly by the thiohydantoin
moiety, thereby promoting hydrolysis of the isopropylidene
group in the presence of water, which is liberated during the
elimination reaction. The NH nitrogen atom then undergoes
intramolecular attack to the masked aldehyde group of the
reducing monosaccharide to zip up the bicyclic skeleton.
Deceivingly, the global yield was very low, mainly due to the
penalty associated with the formation of N-acyl byproduct during
reduction of 9.
FIGURE 1. Structure of (+)-castanospermine (1), 1-deoxynojirimycin
(2) and sp2-iminosugar-type castanospermine analogues (3 and 4).
General structure of the target thiohydantoin-castanospermine glyco-
mimetics (I) and the corresponding retrosynthetic analysis is included.
substituent, as in the sp2-iminosugars 3 and 4. Moreover, the
presence of a second endocyclic nitrogen offers the possibility
to generate molecular diversity at this point.
(Thio)hydantoins and their bi- and tricyclic derivatives
represent an important class of biologically active molecules
with broad medical7 (anticancer, anticonvulsant, antimuscarinic,
antiulcer, and antiarrhythmic), agrochemical8 (herbicidal and
fungicidal), and synthetic applications.9 Their preparation has
been broadly studied,10-12 with special attention to the synthesis
of C-glycosylated derivatives because of their resemblance to
natural nucleosides.13 According to the retrosynthetic scheme
depicted in Figure 1, the target thiohydantoin-piperidine fused
bicyclic core could be assembled by the intramolecular nucleo-
philic addition of the nitrogen atom of the preformed five-
membered heterocycle to the aldehyde group of a monosac-
charide precursor in the open-chain form (II). The key synthetic
intermediate would be the pseudo-C-nucleoside derivative III,
which at its turn can be constructed from the corresponding
furanose thioureidoester IV. A hydroxylation profile of stere-
ochemical complementarity with (+)-castanospermine in the
final bicyclic compounds implies the D-gluco configuration in
the furanose precursors.
Because configuration at the position equivalent to C-4 in
glucose (C-8 in the final compound) is lost during the synthesis,
starting from D-glucose or D-galactose precursors becomes
irrelevant. Previous work had shown that 5-amino-5-deoxy
galactofuranose derivatives bearing an acetyl group at C-3 do
not experience acetyl migration,6a the main handicap in the
glucose route. Accordingly, the D-galactofuranose-derived R-azi-
doester 16, obtained from the known 5-azido-3-O-acetyl-5-deoxy
derivative16 15 by oxidation and esterification as above, afforded
the thiohydantoin-castanospermine bicycle 13 in a more satis-
factory 70% yield after reduction and coupling with butyl
isothiocyanate (f 17 and 18). Conventional sodium methoxide
deacetylation of 13 afforded the fully unprotected derivative
14 (Scheme 1).
(7) (a) Spinks, A.; Waring, W. S. Prog. Med. Chem. 1963, 3, 313. (b) Struck,
R. F.; Kirk, M. C.; Rice, L. S.; Suling, W. J. J. Med. Chem. 1986, 29, 1319. (c)
Matsukura, M.; Daiku, Y.; Ueda, K.; Tanaka, S.; Igarashi, T.; Minami, N. Chem.
Pharm. Bull. 1992, 40, 1823. (d) Brouillete, W. J.; Brown, G. B. J. Med. Chem.
1994, 37, 3289. (e) Ahmed, I. K.; Philippe, B. Tetrahedron 1998, 54, 4859. (f)
Paquette, L. A.; Brand, S.; Behrens, C. J. Org. Chem. 1999, 64, 2010. (g) Somsa´k,
¨
L.; Kova´cs, L.; To´th, M.; Osz, E.; Szila´gyi, L.; Gyo¨rgydea´k, Z.; Dinya, Z.; Docsa,
T.; To´th, B.; Gergely, P. J. Med. Chem. 2001, 44, 2843.
(8) (a) Mizuno, T.; Kino, T.; Ito, T.; Miato, T. Synth. Commun. 2000, 30,
1675. (b) Mappes, C. J.; Pommer, E. H.; Rentzea, C.; Zeeh, B. U.S. Patent
4,198,423, 1980; Chem. Abstr. 93, 71784. (c) Mio, S.; Ichinose, R.; Goto, K.;
Sugai, S.; Sato, S. Tetrahedron 1991, 47, 2111. (d) Garribba, E.; Micera, G.;
Chruscinska, E. J. Chem. Res., Synop. 2001, 10, 408. (e) Nakajima, N.;
Matsumoto, M.; Kirihara, M.; Hashimoto, M.; Katoh, T.; Terashima, S.
Tetrahedron 1996, 52, 117. (f) Gerwick, B. C.; Crouse, G. D.; Murdoch, M. G.;
Green, S. B.; Heim, D. R. Pestic. Biochem. Physiol. 1996, 55, 210.
(9) (a) Ble´riot, Y.; Simone, M. I.; Wormald, M. R.; Dwek, R. A.; Watkin,
D. J.; Fleet, G. W. J. Tetrahedron: Asymmetry 2006, 17, 2276. (b) Ulgheri, F.;
Orru`, G.; Crisma, M.; Spanu, P. Tetrahedron Lett. 2004, 45, 1047. (c) Koo´s,
M.; Steiner, B.; Langer, V.; Gyepesova´, D.; Dur´ık, M. Carbohydr. Res. 2000,
328, 115. (d) Bodanszky, M. In Principles of Peptide Synthesis; Springer-Verlag:
Berlin, 1993; p 95.
Although the galactose route allowed proof of the synthetic
concept, the accessibility of the requested azidoalcohol 15 is
more costly as compared with the D-gluco precursor 6. To fully
exploit the potential of the general approach to introduce
molecular diversity in the new sp2-iminosugar structure, an
(10) De Oliveira, S. M.; Para´ıso da Silva, J. B.; Hernandes, M. Z.; Alves de
Lima, M. C.; Galdino, S. L.; da Rocha Pitta, I. Qu´ım. NoVa 2008, 31, 614.
(11) (a) Klason, P. Chem. Ztg. 1890, 14, 543. (b) Holan, G.; Samuel, E.
J. Chem. Soc. 1961, 4660.
(12) (a) Ware, E. Chem. ReV. 1950, 46, 403. (b) Edward, J. T. Chem. Org.
Sulfur Comp. 1966, 2, 287. (c) Li, J.-P.; Ma, C.-M.; Qu, G.-R. Synth. Commun.
2005, 35, 1203.
(14) Dax, K.; Gaigg, B.; Grassberger, V.; Ko¨lblinger, B.; Stu¨tz, A. E.
J. Carbohydr. Chem. 1990, 9, 479.
(15) Hadwiger, P.; Mayr, P.; Tauss, A.; Stu¨tz, A. E.; Nidetzky, B. Bioorg.
Med. Chem. Lett. 1999, 9, 1683.
(16) Brown, J. M.; Christodoulou, C.; Reese, C. B.; Sindona, G. J. Chem.
Soc, Perkin Trans. 1 1984, 1785.
(13) Khodair, A. I. A.; El Ashry, E. S. H.; Al-Masoudi, N. A. L. Monatsh.
Chem. 2004, 135, 1061.
(17) Winterfeld, G. A.; Khodair, A. I.; Schmidt, R. R. Eur. J. Org. Chem.
2003, 1009.
3596 J. Org. Chem. Vol. 74, No. 9, 2009