Generalized Anomeric Effect in Action: Synthesis and Evaluation of Stable Reducing Indolizidine Glycomimetics as Glycosidase Inhibitors

Article abstract of DOI:10.1021/jo991242o

A series of aminoketalic castanospermine analogues incorporating a stereoelectronically anchored axial hydroxy group at the pseudoanomeric stereocenter (C-5) have been synthesized to satisfy the need for glucosidase inhibitors that are highly selective for α-glucosidases. The polyhydroxylated bicyclic system was built from readily available hexofuranose derivatives through a synthetic scheme that involved (i) the construction of a five-membered cyclic (thio)carbamate or (thio)urea moiety at the nonreducing end and (ii) the intramolecular nucleophilic addition of the heterocyclic thiocarbamic nitrogen atom to the masked aldehyde group of the monosaccharide. A biological screening of the resulting reducing 2-oxa- and 2-azaindolizidines against several glycosidase enzymes is reported.

Full text of DOI:10.1021/jo991242o

136
J . Org. Chem. 2000, 65, 136-143
Gen er a lized An om er ic Effect in Action : Syn th esis a n d Eva lu a tion
of Sta ble Red u cin g In d olizid in e Glycom im etics a s Glycosid a se
In h ibitor s
V´ıctor M. D´ıaz Pe´rez,† M. Isabel Garc´ıa Moreno,† Carmen Ortiz Mellet,† J ose´ Fuentes,†
J uan C. D´ıaz Arribas,‡ F. J avier Can˜ada,‡ and J ose´ M. Garc´ıa Ferna´ndez*,§
Departamento de Quı´mica Orga´nica, Facultad de Qu´ımica, Universidad de Sevilla, E-41071,
Sevilla, Spain, Instituto de Quı´mica Orga´nica General, CSIC, J uan de la Cierva 3, E-28006, Madrid,
Spain, and Instituto de Investigaciones Qu´ımicas, CSIC, Ame´rico Vespucio s/ n, Isla de la Cartuja,
E-41092, Sevilla, Spain
Received August 5, 1999
A series of aminoketalic castanospermine analogues incorporating a stereoelectronically anchored
axial hydroxy group at the pseudoanomeric stereocenter (C-5) have been synthesized to satisfy the
need for glucosidase inhibitors that are highly selective for R-glucosidases. The polyhydroxylated
bicyclic system was built from readily available hexofuranose derivatives through a synthetic scheme
that involved (i) the construction of a five-membered cyclic (thio)carbamate or (thio)urea moiety at
the nonreducing end and (ii) the intramolecular nucleophilic addition of the heterocyclic thiocarbamic
nitrogen atom to the masked aldehyde group of the monosaccharide. A biological screening of the
resulting reducing 2-oxa- and 2-azaindolizidines against several glycosidase enzymes is reported.
In tr od u ction
generically iminosugars (“azasugars”),5 are related to
nojirimycin (NJ , 1) or 1-deoxynojirimycin (DNJ , 2),6
nitrogenous-ringed stereochemical mimics of D-glucose,
or to the indolizidine alkaloid glucomimetic castanosper-
mine (CS, 3).7 Neither 1-3 nor most of the plethora of
synthetic analogues reported so far possess a defined
configuration at the pseudoanomeric center.8 It is there-
fore not surprising that 1-3 simultaneously inhibit
several R- and â-glucosidases, which may be particularly
problematic for clinical applications.4a
Natural and synthetic polyhydroxylated alkaloids with
glycosidase inhibitory properties have been receiving a
great deal of attention both as useful biological tools for
studies on glycoconjugate function, targeting, and turn-
over1 and as potential chemotherapeutic agents for the
treatment of viral infections,2 cancer,3 and metabolic
disorders such as diabetes.4 Most of the biologically
interesting members of this class of compounds, termed
* To whom correspondence should be addressed. Present address:
Instituto de Investigaciones Qu´ımicas, CSIC, Ame´rico Vespucio s/n,
Isla de la Cartuja, E-41092, Sevilla, Spain. Tel: +34 954489559. Fax:
+34 954460565. E-mail: jogarcia@cica.es.
† Universidad de Sevilla.
‡ Instituto de Qu´ımica Orga´nica General, CSIC, Madrid.
§ Instituto de Investigaciones Qu´ımicas, CSIC, Sevilla.
(1) For leading reviews, see: (a) Elbein, A. D.; Molyneux, R. J .
Alkaloid Glycosidase Inhibitors. In Comprehensive Natural Products
Chemistry; Barton, D., Nakanishi, K., Meth-Cohn, O., Eds.; Elsevier:
Oxford, 1999; Vol. 3, p 129. (b) Sears, P.; Wong, C.-H. Chem. Commun.
1998, 1161. (c) Dwek, R. A. Chem. Rev. 1996, 96, 683-720. (d) Kaushal,
G. P.; Elbein, A. D. Methods Enzymol. 1994, 230, 316. (e) Nishimura,
Y. In Studies in Natural Products Chemistry; Atta-ur-Rahman, Ed.;
One approach to improving the selectivity of imino-
sugar glycosidase ligands would be to exploit the stereo-
complementarity of the scissile aglyconic oxygen atom
with the key bilateral carboxylic groups, crucial for the
R- or â-glycosidic bond specificity, in the active site of the
enzyme.9 Yet, the development of aza-O-glycoside glyco-
mimetics is limited by the lability of the O/N acetal
Elsevier: Amsterdam, 1992; Vol 10,
p 495. (f) Legler, G. Adv.
Carbohydr. Chem. Biochem. 1990, 48, 319. (g) Sinnot, M. L. Chem.
Rev. 1990, 90, 1171. (h) Elbein, A. D. Annu. Rev. Biochem. 1987, 56,
497.
(2) (a) Carlson, G. B.; Butters, T. D.; Dwek, R. A.; Platt, F. M. J .
Biol. Chem. 1993, 268, 570. (b) Taylor, D. L.; Sunkara, P. S.; Liu, P.
S.; Kang, M. S.; Bowlin, T. L.; Tyms, A. S. AIDS 1991, 5, 693. (c)
Sunkara, P. S.; Taylor, D. L.; Kang, M. S.; Bowlin, T. L.; Liu, P. S.;
Tyms, A. S.; Sjoerdsma, A. Lancet 1989, 1206. (d) Karpas, A.; Fleet,
G. W. J .; Dwek, R. A.; Petursson, S.; Namgoong, S. K.; Ramsden, N.
G.; J acob, G. S.; Rademacher, T. W. Proc. Natl. Acad. Sci. U.S.A. 1988,
85, 9229.
(5) Although the term “azasugar” is widely used in the literature to
refer to glycomimetics in which the endocyclic oxygen atom has been
replaced by nitrogen, this is not strictly correct according to the
IUPAC-IUBM nomenclature recommendations for carbohydrates.
“Azahexose” would actually imply that
a carbon atom has been
exchanged into nitrogen. See: McNaught, A. D. Pure Appl. Chem. 1996,
68, 1919.
(3) For a review see: Gross, P. E.; Baker, M. A.; Carver, J . P.;
Dennis, J . W. Clin. Cancer Res. 1995, 1, 935.
(6) For reviews on DJ - and DNJ -based inhibitors, see: (a) Ganem,
B. B. Acc. Chem. Res. 1996, 29, 340. (b) Gijsen, H. J . M.; Qiao, L.;
Fitz, W.; Wong, C.-H. Chem. Rev. 1996, 96, 443. (c) Wong, C.-H.;
Halcomb, R. L.; Ichikawa, Y.; Kajimoto, T. Angew. Chem., Int. Ed. Engl.
1995, 34, 412. (c) Wong, C.-H.; Halcomb, R. L.; Ichikawa, Y.; Kajimoto,
T. Angew. Chem., Int. Ed. Engl. 1995, 34, 521.
(7) For reviews on CS-based inhibitors, see: (a) Michael, J . P. Nat.
Prod. Rep. 1997, 14, 21. (b) Cossy, J .; Vogel, P. In Studies in Natural
Products Chemistry; Atta-ur Rahman, Ed.; Elsevier: New York, 1993;
Vol. 12, p 275. (c) Burgess, K.; Henderson, I. Tetrahedron 1992, 48,
4045.
(4) (a) Witczak, Z. J . Carbohydrates as New and Old Targets for
Future Drug Design. In Carbohydrates in Drug Design; Witczak, Z.
J ., Ed.; Marcel Dekker Inc.: New York, 1997; p 1. (b) Balfour, J . A.;
McTavish, D. Drugs 1993, 46, 1025. (c) Robinson, K. M.; Begovic, M.
E.; Rhinehart, B. L.; Heineke, E. W.; Ducep, J .-B.; Kastner, P. R.;
Marshall, F. N.; Danzin, C. Diabetes 1991, 40, 825. (d) Anzeveno, P.
B.; Creemer, L. J .; Daniel, J . K.; King, C.-H.; Liu, P. S. J . Org. Chem.
1989, 54, 2539. (e) Platt, F. M.; Neises, G. R.; Reinkensmeier, G.;
Townsend, M. J .; Perry, V. H.; Proia, R. L.; Winchester, B.; Dwek, R.
A.; Butters, T. D. Science 1997, 276, 428.
10.1021/jo991242o CCC: $19.00 © 2000 American Chemical Society
Published on Web 12/16/1999

Generalized Anomeric Effect
J . Org. Chem., Vol. 65, No. 1, 2000 137
and II can be seen as conformationally restricted monosac-
charide analogues: 2-oxaindolizidines retain the hy-
droxylation profile of the parent hexose, whereas 2-aza-
indolizidines are, formerly, mimics of 6-amino-6-deoxy
sugars. Both share the reducing character with 1, but
they also have the aminoketalic hydroxy group, anchored
in the axial position, like the natural aglycons in R-gly-
copyranosides. The preparation of the key intermediates,
the scope and limitation of the methods, and the struc-
tural glycosidase inhibitory selectivities and potency
relationships are discussed.
F igu r e 1. Structures I, reducing 2-oxaindolizidine; II, reduc-
ing 2-azaindolizidine; III, aldohexofuranose.
function under hydrolytic conditions.10 Although the
anomeric hydroxyl group in 1 can be seen as an universal
surrogate for natural glycosidic links, reducing azasugars
suffer likewise from low stability, and moreover, rapid
anomerization occurs in aqueous solution. This paper
describes a novel approach that obviates these problems
based on the assumption that replacement of the imino
sp3 nitrogen by a (thio)carbamic-type nitrogen would
substantiate sp2 character11 and hence increases the
orbitalic contribution to the generalized anomeric effect12
in aminoketalic centers.13 The resulting N-(thio)carbonyl
iminosugars exhibit a much higher stability as well as
configurational integrity in water.14 This principle has
been translated into a practical synthesis of reducing
2-oxa- and 2-azaindolizidine glycomimetics (Figure 1,
structures I and II, respectively), a hitherto unknown
class of compounds, from cyclic (thio)carbamate and cyclic
(thio)urea carbohydrate precursors. Like 3, structures I
Resu lts
A retrosynthetic analysis revealed that the bicyclic
skeleton of 2-oxa- and 2-azaindolizidines can be con-
structed by the intramolecular nucleophilic addition of
the nitrogen atom of five-membered cyclic (thio)carbam-
ate or (thio)urea intermediates to a suitably located
carbonyl group, with simultaneous generation of the
aminoketal function (Figure 1). Our synthetic strategy
centers on the use of aldohexofuranose-derived azole
heterocycles with pseudo-C-nucleoside structure (III) as
precursors because the masked aldehyde group of the
monosaccharide can actually act as the electrophilic
target through the open-chain tautomeric form.
Syn th esis a n d Str u ctu r e of 5-Hyd r oxy-2-oxa in -
d olizid in es. The initial synthetic objective of this re-
search was the preparation of reducing 2-oxacastano-
spermine analogues, that is, oxazolidine-piperidine
bicyclic derivatives with a hydroxylation profile at the
six-membered ring identical to that of D-glucose. The
thiocarbonylation reaction of 5-amino-5-deoxy-1,2-O-iso-
propylidene-R-D-glucofuranose15 (4) with carbon disul-
fide-DCC at -10 °C afforded the required 6,5-(cyclic
thiocarbamate) 6 in a regioselective manner (Scheme 1).16
Because the epimeric amino sugar 5, with L-ido config-
uration, was readily accessible with minor modification
of the syntetic route leading to 4,15 it was of interest to
contemplate a parallel study on the reactivity and
structural properties for both sugar series to assess the
scope and limitations of the approach. Thus, compound
5 was subjected to thiocarbonylation with the same
reagent as that used with 4 to give 7 in 85% yield.
Carbonylation of 4 and 5 with bis(trichloromethyl)-
carbonate (triphosgene)17 and diisopropyl ethylamine
(DIPEA) led to the corresponding 2-oxazolidinone het-
erocycles 8 and 9, respectively, in virtually quantitative
yields (Scheme 1).18 The relative proportion of reactants
was found to have a strong influence on the outcome of
the latter reaction. Higher concentrations of triphosgene
resulted in further intra- and intermolecular carbonyla-
tion, thereby producing a subsequent loss of product
purity.
(8) The exceptions correspond to compounds having the exocyclic
glycosidic oxygen replaced by other atom or group of atoms. Noteworthy
examples are aza-C-glycosides and aza-S-glycosides. See: (a) Leewen-
burgh, M. A.; Picasso, S.; Overkleeft, H. S.; van der Marel, G. A.; Vogel,
P.; van Boom, J . H. Eur. J . Org. Chem. 1999, 1185, 5. (b) Zhu, Y.-H.;
Vogel, P. J . Org. Chem. 1999, 64, 666. (c) J ohns, B. A.; Pan, Y. T.;
Elbein, A. D.; J ohnson, C. R.; J . Am. Chem. Soc. 1997, 119, 4856. (d)
Fuchss, T.; Streicher, H.; Schmidt, R. R. Liebigs Ann./ Recl. 1997, 1315.
(e) Asano, N.; Nishida, M.; Kizu, H.; Matsui, K.; Watson, A. A.; Nash,
R. J . J . Nat. Prod. 1997, 60, 98. (f) Wong, C.-H.; Proveacher, L.; Porco,
J r., J . A.; J ung, S.-H.; Wang, Y.-F.; Chen, L.; Wang, R.; Steensma, D.
H. J . Org. Chem. 1995, 60, 1492. (g) Suzuki, K.; Hashimoto, H.
Tetrahedron Lett. 1994, 4119.
(9) For reviews on the reaction mechanism of enzymatic glycosidic
bond cleavage, see ref 5f,g and: (a) Heightman, T. D.; Vasella, A. T.
Angew. Chem., Int. Ed. Engl. 1999, 38, 750. (b) White, A.; Rose, D. R.
Curr. Opin. Struct. Biol. 1997, 7, 645. (c) Gorenstein, D. G. Chem. Rev.
1987, 87, 1047. (d) Kirby, A. J . Acc. Chem. Res. 1984, 17, 305.
(10) J ohnson, C. R.; Golebiowski, A.; Sundram, H.; Miller, M. W.;
Dwaihy, R. L. Tetrahedron Lett. 1994, 36, 653.
(11) For reviews on the comparative chemical and electronic proper-
ties of N-(thio)carbonyl compounds, see: (a) Molina, M. T.; Ya´n˜ez, M.;
Mo´, O.; Notario, R.; Abboud, J .-L. M. The Thiocarbonyl Group. In
Supplement A3, The Chemistry of Double-Bonded Functional Groups;
Patai, S., Ed.; J ohn Wiley & Sons: Chichester, 1997; Part 2, p 1355.
(b) Garc´ıa Ferna´ndez, J . M.; Ortiz Mellet, C. Sulfur Rep. 1996, 19, 61.
(c) Duus, F. Thiocarbonyl Compounds. In Comprehensive Organic
Chemistry; Barton, D., Ollis, W. D., Eds.; Pergamon Press: London,
1979; Vol. 3, p 373. (d) Walter, W.; Voss, J . The Chemistry of
Thioamides. In The Chemistry of Amides; Zabicky, J ., Ed.; J ohn Wiley
& Sons: London, 1970; p 383. (e) Stewart, W. E.; Siddall, T. H., III.
Chem. Rev. 1970, 70, 517.
(12) For reviews on the generalized anomeric effect, see: (a)
Tvaroska, I.; Bleha, T. Adv. Carbohydr. Chem. Biochem. 1989, 47, 45.
(b) J uaristi, E.; Cuevas, G. The Anomeric Effect; CRC Press: Boca
Raton, 1995. For recent papers on the controversial subject of the
relative contribution of orbitalic and electrostatic interactions to the
generalized anomeric effect in azaheterocycles, see: (c) Perrin, C. L.;
Armstrong, K. B.; Fabian, M. A. J . Am. Chem. Soc. 1994, 116, 715. (d)
Salzner, U. J . Org. Chem. 1995, 60, 986. (e) Ritter, J .; Gleiter, R.;
Irngartinger, H.; Oeser, T. J . Am. Chem. Soc. 1997, 119, 10599.
(13) (a) J ime´nez Blanco, J . L.; Ortiz Mellet, C.; Fuentes, J .; Garc´ıa
Ferna´ndez, J . M. Tetrahedron 1998, 54, 14123. (b) J ime´nez Blanco, J .
L.; Ortiz Mellet, C.; Fuentes, J .; Garc´ıa Ferna´ndez, J . M. Chem.
Commun. 1996, 2077.
(15) Dax, K.; Gaigg, B.; Grassberger, V.; Ko¨lblinger, B.; Stu¨tz, A.
E. J . Carbohydr. Chem. 1990, 9, 479.
(16) Previous work on the synthesis of 2-thioxo-1,3-O,N-heterocycles
from amino sugar templates had shown that â-amino alcohol segments
undergo regioselective cyclization to the corresponding oxazolidine-2-
thiones upon reaction with thiophosgene, provided that a cis relation-
ship of the reactive groups is allowed. See: (a) Garc´ıa Ferna´ndez, J .
M.; Ortiz Mellet, C.; Fuentes, J . J . Org. Chem. 1993, 58, 5192. (b)
Garc´ıa Ferna´ndez, J . M.; Ortiz Mellet, C.; J ime´nez Blanco, J . L.;
Fuentes, J . J . Org. Chem. 1994, 59, 5565. In the present case, this
procedure proved unsatisfactory, affording a rather modest 28% yield.
Simultaneous formation of oligomeric material was observed. Using
1,1-thiocarbonyldiimidazole as the thiocarbonylating reagent likewise
resulted in low yields.
(14) J ime´nez Blanco, J . L.; D´ıaz Pe´rez, V. M.; Ortiz Mellet, C.;
Fuentes, J .; Garc´ıa Ferna´ndez, J . M.; D´ıaz Arribas, J . C.; Can˜ada, F.
J . Chem. Commun. 1997, 1969.
(17) For a review on the synthetic applications of triphosgene, see:
Cotarca, L.; Delogu, P.; Nardelli, A.; Sunjic, V. Synthesis 1996, 553.

138 J . Org. Chem., Vol. 65, No. 1, 2000
D´ıaz Pe´rez et al.
Sch em e 1a
Sch em e 2a
a
Reagents: (i) 90% TFA-water; (ii) Amberlite IRA-68 (OH-)
(87-96%); (iii) 1:1 Ac2O/pyridine (85-97%).
the all-axial 5S diastereomer was detected exclusively
for the fully unprotected compounds (12 and 16) and the
tetra-O-acetates (13 and 17, respectively). The configu-
rational assignment was confirmed by the existence of a
long-range coupling constant between protons in W
arrangement and the lack of NOE contact between H-5
and H-8a.
a
Reagents: (i) CS2, DCC, CH2Cl2, 6-7 h; (ii) triphosgene (1.5
Syn th esis a n d Str u ctu r e of 5-Hyd r oxy-2-a za in -
d olizid in es. To implement this strategy of accessing
reducing 2-azacastanospermine analogues, the prepara-
tion of 5,6-diamino-5,6-dideoxyaldohexofuranose deriva-
tives was required. The reported methods for the syn-
thesis of these compounds are very scarce and suffer from
either low overall yields or formation of epimeric mix-
tures.20 Because the introduction of a first azido sub-
stituent at C-5 can be effected very efficiently, and with
total stereocontrol in the multigram scale in both D-gluco
and L-ido series (18 and 19, respectively), we decided to
introduce the second nitrogen substituent in a stepwise
manner: (i) first, we accomplished the tosylation of the
primary hydroxyl group and the in situ acetylation of the
remaining 3-OH (f20 and 21, respectively), and (ii) then
we effected the nucleophilic displacement of tosylate by
azide anion (f22 and 23) (Scheme 3). Deacetylation and
Staudinger reduction of the diazides 24 and 25 afforded
the corresponding diamines 26 and 27, in quantitative
yield21 (13C NMR monitoring), as hygroscopic solids which
were directly thiocarbonylated (f28 and 30) or carbo-
nylated (f29 and 31, respectively) following the protocols
previously optimized in the preparation of the cyclic
(thio)carbamate counterparts (Scheme 4).
equiv), DIPEA (10 equiv); CH2Cl2, 5 min.
Acid hydrolysis of the acetal protecting group in 6-9
with TFA-water led, initially, to R,â anomeric mixtures
of the corresponding furanoses, with the (thio)carbamate
group probably being protonated, as seen from 13C NMR
spectra of the crude reaction mixtures. After neutraliza-
tion, the equilibrium was spontaneously shifted toward
the target 2-oxaindolizidines (Scheme 2).
The castanospermine analogues 10 and 14 existed in
D2O solution as single diastereomers. The high field shift
of the C-1 resonance19 (77.4 and 77.0 ppm, respectively)
confirmed the aminoketalic bicyclic structure, whereas
3
the vicinal J H,H values around the piperidine ring
unambiguosly pointed to the 5R configuration for the new
stereocenter, with the pseudoanomeric hydroxy group in
axial position, fitting the anomeric effect. No traces,
either of the 5S epimer or of the furanose forms, were
detected even after conventional acetylation (f11 and
15, respectively). For the L-idose derived indolizidines,
(18) An alternative synthetic approach based on the aza-Wittig-type
reaction of 5-azido-5-deoxy-1,2-O-isopropylidene-R-D-glucofuranose, or
the L-ido epimer, with triphenylphosphine and CO2, proved unsuc-
cessful. When 1,1-carbonyldiimidazole was used as the carbonylating
reagent, the corresponding cyclic thiocarbamates were isolated in 50%
yield.
(19) For clarity of presentation, the authors choose not to use the
numbers resulting from the heterocyclic compounds (see the Experi-
mental Section) in the notation of atoms for NMR data. Instead, the
notation is kept consistent with the parent carbohydrate compounds.
See Figure 1 for atom notation equivalency.
(20) (a) Paulsen, H. Annalen 1963, 665, 166. (b) Paulsen, H.; Todt,
K. Chem. Ber. 1966, 99, 3450. (c) Rank, W.; Baer, H. H. Carbohydr.
Res. 1974, 35, 65.
(21) The acetylation/deacetylation step was found to be necessary
to avoid extensive formation of 3,6-anhydro derivatives, resulting from
intramolecular nucleophilic displacement of the tosylate group by
3-OH.

Generalized Anomeric Effect
Sch em e 3a
J . Org. Chem., Vol. 65, No. 1, 2000 139
Sch em e 4a
a
Reagents: (i) TsCl (1.1 equiv), pyridine, 0 (D-gluco) and -30
°C (L-ido), respectively, 1:1 Ac2O/pyridine (69-87%); (ii) NaN3,
DMF, 70 °C (86-98%); (iii) NaOMe, MeOH (97-98%), (iv) Ph3P,
dioxane/MeOH, aqueous NH4OH (quant).
The tendency of the cyclic (thio)ureas 28 and 29 toward
acid hydrolysis of the acetal protecting group was analo-
gous to that previously observed for the configurationally
related oxazolidines 6 and 8. Upon neutralization of the
reaction mixtures, the corresponding fused imidazoli-
dine-piperidine bicyclic glucomimetics 32 and 34 were
isolated in high yield (Scheme 4). Conventional acetyla-
tion of the thioxo derivative 32 afforded the penta-O,N-
acetyl compound 33, whereas the oxo analogue 34 was
transformed into the tetra-O-acetate 35 under identical
reaction conditions. Both the fully unprotected and the
acetylation products showed coupling constant values
around the six-membered ring, indicative of a conforma-
tion close to a chair with the pseudoanomeric hydroxy
or acetoxy group in axial position. In the case of the
imidazoline derivatives of L-idose 30 and 31, the outcome
of the reaction was radically different. Although the
thioxo compound 30 led to the expected all-axial polyhy-
droxy 2-azaindolizidine 36 with 5S configuration, albeit
in low yield, the analogous 3-oxo diazabicycle could not
be isolated and experienced further dehydration to give
a product for which structure 37 was proposed, on the
basis of microanalytical, MS, and NMR data (Scheme 4).
Eva lu a tion of th e In h ibitor y Selectivities a n d
P oten cies of Red u cin g In d olizid in es. First, 2-oxa-3-
thioxoindolizidines 10 and 12 were screened against
yeast R-glucosidase and sweet almond â-glucosidase.
Both enzymes have been widely used to check the R-
versus â-glucosidase selectivity for different inhibitors.
The conventional reducing iminosugar glucomimetic NJ
(1) inhibits â-glucosidase more potently (Ki ) 0.89 µM)
than R-glucosidase (Ki ) 6.3 µM).1f In stark contrast, the
inhibition constants for 10 (>100 000 and 40 µM, respec-
tively) indicated a reverse linkage specificity. Although
the inhibition potency for R-glucosidase is about 1 order
of magnitude lower as compared with 1, the selectivity
ratio is inverted, by a factor of above 104, for this
a
Reagents: (i) CS2, DCC (73-79%); (ii) triphosgene (1.5 equiv),
DIPEA (10 equiv) (61-70%); (iii) 90% TFA-water, Amberlite IRA-
68 (OH-); (iv) 1:1 AC2O/pyridine (85-98%).
Ta ble 1. Com p a r ison of In h ibitor y Activities (Ki, µM) for
2-Oxa (10 a n d 14) a n d 2-Aza ca sta n osp er m in es (32 a n d 34)
compound
enzyme
10
14
n.i.a
2.2
32
n.i.
4 200
n.i. 4 300
15 000 1 600
34
n.i.
5 700
â-glucosidase (almonds) 150 000
R-glucosidase (yeast)
R-glucosidase (rice)
amiloglucosidase
(Aspergillus niger)
trehalasa (pig kidney)
intestinal lactase
(newborn lamb)
40
1 600 1600
22 000 8 200
2 000
30 000
n.d.b
n.d.
n.d.
n.d.
n.d.
n.d.
a
b
No inbition detected. Not determined.
particular pair of enzymes. The R-linkage selectivity is
even much higher than that of DNJ (2), considered to be
a selective inhibitor of R-glucosidase (Ki ) 47 and 12.3
µM for â- and R-glucosidase, respectively),1f or CS (3; Ki
) 1.5 and >1500 µM).1f The L-idose derived diastereomer
12 did not inhibit any of these two glycosidases, discard-
ing the possibility of any significant contribution of
inspecific interactions with the cyclic thiocarbamate
portion to the Ki values.
We have further evaluated the ability of 10 to inhibit
a series of other glycosidases: R-glucosidase (rice), amilo-
glucosidase (Aspergillus niger), trehalase (pig kidney),
and intestinal â-glycosidase complex (intestinal lactase
purified from lamb). The Ki values obtained are collected
in Table 1. None of these enzymes were strongly inhib-
ited, having Ki values in the mM range. It may be

140 J . Org. Chem., Vol. 65, No. 1, 2000
D´ıaz Pe´rez et al.
concluded that 10 can discriminate not only between
sweet almond â-glucosidase and yeast R-glucosidase but
also between the latter and other R-glucosidases. This
effect was found to be more pronounced for the 3-oxo
analogue 14, which was 18-fold more potent as an
inhibitor of yeast R-glucosidase (Ki ) 2.2 µM) than 10
and, likewise, a very poor inhibitor of rice R-glucosidase
and amiloglucosidase. The replacement of the endocyclic
oxygen atom by nitrogen dramatically weakened the
inhibitory potency against yeast R-glucosidase and abol-
ished selectivity. Thus, 32 and 34 were found to be very
weak inhibitors of the above three R-glucosidases.
glycoside hydrolysis at the endocyclic oxygen atom re-
gion.24 In agreement with that assumption, the 2-ox-
acastanospermine glycomimetics 10 and 14 exhibited Ki
values against yeast R-glucosidase that are comparable
to those of the classical iminosugars 1 and 2; compound
14 was a stronger inhibitor by 3-fold and 6-fold, respec-
tively. In sharp contrast, they practically did not inhibit
â-glucosidases at all, which is in agreement with the
decisive role of anomeric stereochemistry in enzyme
specificity.
Indolizidine glycomimetics, such as castanospermine
(3), are known to be more potent and selective inhibitors
of some R-glucosidases than their related piperidine
analogues. This has been ascribed to the rigidity of the
bicyclic structure, which locks the homologous bond to
C-5-C-6 in hexopyranosides, thereby fixing the hydroxy
group equivalent to 6-OH in the correct position to fully
interact with the binding site of the enzyme.25 A com-
parative analysis of the topographical characteristics of
3, 10, and 14 at this region26 showed that the result of
replacing the 1-hydroxyindolizidine structure into a
2-oxaindolizidine skeleton is formally equivalent to a
rotation of ∼120° about the above bond. This brings the
oxygen atom equivalent to O-6 in hexoses from a gauche
arrangement into the anti position with respect to the
piperidine carbon atom C-8. This structural change is
probably responsible for the dramatic differences ob-
served in the glycosidase inhibitory properties: Although
castanospermine does not inhibit yeast R-glucosidase and
is a very strong inhibitor of rice R-glucosidase (Ki ) 0.015
µM)1f and amiloglucosidase (Ki ) 8 µM),27 the 2-ox-
acastanospermine analogues 10 and 14 have the reverse
selectivity (Table 1). These results strongly support the
notion that the prochiral O-6 in D-glucopyranosides binds
to R-glucosidases in a specific manner. Therefore, the
results suggest a way to discriminate the bound confor-
mation: those glucosidases that are inhibited by castano-
spermine require a disposition close to gauche-gauche
about C-5-C-6, whereas those that are inhibited by
2-oxacastanospermine derivatives bind in a gauche-
trans arrangement (Figure 2).28 Although differences in
the pKb of the inhibitors may also have an influence in
the R-glucosidase selectivity, the orientational properties
of the O-6 equivalent atom seem to be the major
structural requirement. Thus, replacement of this key
oxygen atom (e.g., with 1-deoxycastanospermine25 or
2-azacastanospermines 32 and 34), which does not
Discu ssion
We have designed and synthesized a new class of
glycosidase inhibitors, reducing polyhydroxyindolizidines
incorporating a (thio)carbamic segment in the bicyclic
skeleton. The synthetic scheme is straightforward and
employs readily available monosaccharide precursors.
The two configurations considered in this study, D-glucose
and L-idose, are limit cases regarding steric constraints.
In the first series, the nonanomeric substituents are in
the equatorial position in the final bicyclic glycomimetics,
whereas in the latter they are axially oriented. Neverthe-
less, the diastereomer having the aminoketalic hydroxy
group in axial arrangement was exclusively detected in
water solution in all cases. This result is in agreement
with the existence of a very strong and stabilizing
interaction between the π-type lone-pair orbital of the
nitrogen atom in the ground state of (thio)carbamic
functionalities and the σ* antibonding orbital of the
contiguous C-O bond, which is fully operative in polar
solvents. To the best of our knowledge, these are the first
examples of configurationally and conformationally stable
reducing glycomimetics.
Because the double-bond character of the N-C(Y)X
bond decreases in the sense thiocarbamate > carbamate
> thiourea > urea,11 a parallel abatement of the contri-
bution of the above orbitalic interaction to the generalized
anomeric effect on going from 2-oxa-3-thioxoindolizidines
to 2-aza-3-oxoindolizidines was expected. This probably
explains the lower stability of L-idose derived 2-azain-
dolizidines, for which steric repulsions balance or over-
come the anomeric effect stabilization. With the exception
of this particular case, all compounds here reported were
found to be stable for months as solids or as aqueous
solution in a refrigerator, which is notably different than
was found to be the case for NJ -type reducing iminosug-
ars.22
(24) Recent theoretical and experimental studies indicate that true
transition state analogues must be essentially neutral or zwitterionic
at physiological pH to keep specificity against the target enzyme. See:
(a) J eong, J .-H.; Murray, B. W.; Takayama, S.; Wong, C.-H. J . Am.
Chem. Soc. 1996, 118, 4227. (b) Legler, G.; Finken, M.-T. Carbohydr.
Res. 1996, 292, 103. (c) Ble´riot, Y.; Genre-Grandpierre, A.; Imberty,
A.; Tellier, C. J . Carbohydr. Chem. 1996, 15, 985. (d) Deng, H.; Chan,
A. W.-Y.; Bagdasianan, C. K.; Estupin˜a´n, B.; Ganem, B.; Callender,
R. H.; Scharam, V. L. Biochemistry 1996, 35, 6037. (e) Ernert, P.;
Vasella, A.; Weber, M.; Rupitz, K.; Withers, S. G. Carbohydr. Res. 1993,
250, 113.
(25) Winchester, B. G.; Cenci di Bello, I.; Richardson, A. C.; Nash,
R. J .; Fellows, L. E.; Ramsden, N. G.; Fleet, G. W. J . Biochem. J . 1990,
269, 227.
(26) Molecular mechanics calculations for 3, 10, and 14 were
performed using MM2* as integrated in MACROMODEL 6.0 and the
GB/SA continuum solvent model for water. The three-dimensional
geometries obtained were in agreement with the discussed NMR data.
(27) Molyneux, R.-J .; Pan, Y. T.; Tropea, J . E.; Benson, M.; Kaushal,
G. P.; Elbein, A. D. Biochemistry 1991, 30, 9981.
Iminosugars are thought to be good, but rather non-
specific, inhibitors of glycosidases because they mimic the
glycosyl cation, a proposed intermediate in the mecha-
nism of action for both R- and â-glycosidases, as a result
of the heterocyclic nitrogen atom being protonated at
physiological pH.1g In N-(thio)carbonyl compounds, the
basicity at the nitrogen center is drastically decreased
(by 14 pK units) as compared with the corresponding
amine,11,23 while keeping a positive charge density result-
ing from delocalization of the lone electron pair into the
(thio)carbonyl group. This scenario is probably closer to
that encountered in the transition state of enzymatic
(22) Paulsen, H.; Todt, K. Adv. Carbohydr. Chem. 1968, 23, 115.
(23) Humeres, E.; Debacher, N. A.; Sierra, M. M. de S.; Franco, J .
D.; Schutz, A. J . Org. Chem. 1998, 63, 1598.
(28) For the nomenclature used to denote the staggered rotamers
about the C-5-C-6 bond in hexopyranoses, see: Bock, K.; Duus, J . Ø.
J . Carbohydr. Chem. 1994, 13, 513.

Generalized Anomeric Effect
J . Org. Chem., Vol. 65, No. 1, 2000 141
between C-8 and O-2 that is equivalent to a gauche-
trans conformation about C-5-C-6 in glucopyranosides.
Exp er im en ta l Section
Gen er a l P r oced u r es. Optical rotations were measured at
room temperature in either 1-cm or 1-dm tubes. IR spectra
were recorded on an FT-IR instrument. UV spectra were
1
recorded with a Philips PU 8710 spectrophotometer. H and
13C NMR spectra were recorded at 500 (125.7) and 300 (75.5)
MHz. In the FABMS spectra, the primary beam consisted of
Xe atoms with a maximum energy of 8 keV. The samples were
dissolved in the matrices m-nitrobenzyl alcohol or thioglycerol
and the positive ions were separated and accelerated over a
potential of 7 keV. NaI was added as cationizing agent. TLC
was performed with E. Merck precoated TLC plates, silica gel
30F-245, with visualization by UV light and by charring with
10% sulfuric acid. Fully deprotected compounds were purified
by GPC. Acetylations were effected conventionally with 1:1
pyridine/Ac2O (10 mL per 1 g of sample). Microanalyses were
performed by the Instituto de Investigaciones Quı´micas (Sevi-
lla, Spain).
F igu r e 2. Major staggered rotamers about C-5sC-6 in the
conformational equilibrium of R-glucopyranosides.
Ma ter ia ls. 5-Azido-5-deoxy-1,2-O-isopropylidene-R-D-glu-
cofuranose (18) was prepared from commercial D-glucofura-
nurono 6,3-lactone in four steps, as reported.15 The L-ido
epimer 19 was obtained in three steps from the same lactone
(50% overall yield) by triflation at O-5, nucleophilic displace-
ment by an azide anion, and reduction of the lactone group
with lithium borohydride. Staudinger reduction29 of the 5-azido
derivatives with triphenylphosphine in dioxane-MeOH fol-
lowed by in situ hydrolysis of the resulting iminophosphoranes
(RsNdPPh3) with ammonium hydroxide afforded the corre-
sponding 5-amino-5-deoxy sugars with D-gluco (4) and L-ido
configuration (5). The 5,6-diamino-5,6-dideoxy-1,2-O-isopro-
pylidenehexofuranoses 26 and 27 were prepared from the
monoazides 18 and 19 by replacement of the primary OH-6
group by azide via tosylate and final Staudinger reduction
(Scheme 3; see the Supporting Information).
F igu r e 3. Probable mode of interaction between reducing
2-oxacastanospermine analogs and yeast R-glucosidase.
significantly affect the basicity at the nitrogen center,
abolishes strong inhibition in both series.
Further work is necessary to determine the exact mode
of interaction of reducing indolizidines with glycosidases.
Kinetic studies for 10 and 14 against yeast R-glucosidase
indicate fast and competitive inhibition. Most probably,
the bicyclic glucomimetics are bound in a ground-state-
like conformation, with the axial anomeric oxygen in the
necessary position to interact with the general acid at
the active site and the thiocarbamic nitrogen in position
to interact with the general base carboxylate (Figure 3).
Whether the enzyme is able to transform these potential
substrates into transition-state structures and, eventu-
ally, glycosyl-cation-like intermediates is still under
investigation. In any case, our work provides a new tool
for designing selective R-glycosidase inhibitors by exploit-
ing orbitalic interactions in bicyclic aminoketalic systems.
The solvents were commercial grade and were used as
supplied, with the following exceptions: DMF was distilled
from BaO. Methanol was distilled from methylmagnesium
iodide. Pyridine was distilled from KOH. Acetic anhydride was
distilled from freshly melted sodium acetate.
Gen er a l P r oced u r e for In h ibition Assa y. Inhibitory
potencies of the 2-oxa- and 2-azaindolizidines were determined
by spectrophotometrically measuring the residual hydrolytic
activities of the glycosidases against the corresponding p-
nitrophenyl R- or â-D-glucopyranoside, o-nitrophenyl R-D-
galactopyranoside (for lactase), or trehalose in the presence
of the corresponding 2-oxaindolizidine. The glycosidases used
were R-glucosidase (rice), R-glucosidase (yeast), amiloglucosi-
dase (Aspergillus niger), â-glucosidase (almond), intestinal
â-glycosidase complex (intestinal lactase, newborn lamb), and
trehalase (pig kidney). All enzymes were purchased from
Sigma, except the lactase, which was isolated and purified as
reported.30
Con clu sion s
Each assay was performed in a phosphate buffer at the
optimal pH for each enzyme. The reactions were initiated by
the addition of the enzyme to a solution of the substrate in
the presence or absence of various concentrations of inhibitor.
After the mixture was incubated for 10-30 min at 37 °C, the
reaction was quenched by the addition of 1 M Na2CO3 or a
solution of GLC-Trinder (Sigma, for trehalase). The absorbance
of the resulting mixture was determined at 400 nm (for
p-nitrophenol), 420 nm (for o-notrophenol), or 505 nm (for
D-glucose, in the case of trehalase). The Ki values were
determined with a Sigma Plot program (version 4.14, J andel
Scientific).
The concept, demonstrated herein, of stereoelectroni-
cally controlling the stereochemistry at aminoketalic
centers in aqueous solution, through the use of N-(thio)-
carbonyl functionalities, indicates a new design strategy
for the development of selective inhibitors of R-glycosi-
dases. By taking advantage of the capability of N-(thio)-
carbonyl functionalities to undergo nucleophilic attack
on the masked aldehydo group of reducing monosaccha-
ride derivatives, an efficient synthesis of reducing 2-oxa-
and 2-azaindolizidine glycomimetics has been achieved.
The observed R versus â linkage specificity is ascribable
to the presence of an axially anchored pseudoanomeric
hydroxy group. The selectivity for yeast R-glucosidase,
among other R-glucosidases, is probably due to the
ridigity of the bicyclic skeleton that fixes an anti position
(29) (a) Staudinger, H.; Meyer, J . Helv. Chim. Acta 1919, 2, 635.
(b) Gobolobov, Yu.; Zhmurova, I. N.; Kasukhin, L. F. Tetrahedron 1981,
37, 437.
(30) Rivera-Sagredo, A.; Can˜ada, F. J .; Nieto, O.; J ime´nez-Barbero,
J .; Mart´ın-Lomas, M. Eur. J . Biochem. 1992, 209, 415.

142 J . Org. Chem., Vol. 65, No. 1, 2000
D´ıaz Pe´rez et al.
5-Am in o-5-d eoxy-1,2-O-isop r op ylid en e-r-D-glu cofu r a -
n ose 5,6-(Cyclic th ioca r ba m a te) (6). To a stirred solution
of 5-amino-5-deoxy-1,2-O-isopropylidene-R-D-glucofuranose 4
(120 mg, 0.55 mmol) in CH2Cl2 (5 mL) at -10 °C were added
CS2 (0.3 mL, 4.5 mol) and DCC (114 mg, 0.55 mmol). The
reaction mixture was allowed to reach room temperature and
was stirred for 6-7 h. The solvent was removed under reduced
pressure, and the residue was purified by column chromatog-
raphy (2:1 EtOAc/petroleum ether) to give 6 (115 mg, 80%) as
an amorphous solid: Rf 0.38; [R]D ) -21.2 (c 1.0, acetone);
UV (MeOH) 242 nm (ꢀmM 16.3); IR (KBr) νmax 3337, 3262, 1530
(5R,6S,7R,8R,8a R)-5,6,7,8-Tetr a a cetoxy-2-oxa -3-th iox-
oin d olizid in e (11). Compound 11 was given as an amorphous
solid in 95% yield: Rf (EtOAc/petroleum ether) 0.43; [R]D
)
+33.5 (c 0.8, CH2Cl2); UV (CH2Cl2) 248 nm (ꢀmM 27.6); IR (KBr)
νmax 1753 cm-1; 1H NMR (500 MHz, CDCl3) Table 2 (Support-
ing Information) and δ 2.16, 2.08, 2.05, 2.02 (4 s, each 3 H);
13C NMR (125.5 MHz, CDCl3) δ 187.0, 169.8, 169.7, 169.2,
168.3, 74.8, 71.6, 71.3, 68.6, 68.3, 56.1, 20.5, 20.4, 20.2 (2C);
FABMS m/z 412 (100, [M
15H19NO9S: C, 46.27; H, 4.92; N, 3.60. Found: C, 46.22; H,
4.82; N, 3.55.
+
Na]+). Anal. Calcd for
C
cm-1
;
1H NMR (500 MHz, CD3OD) Table 2 (Supporting
(5S,6S,7R,8R,8a S)-5,6,7,8-Tetr a h yd r oxy-2-oxa -3-th iox-
oin d olizid in e (12). Compound 12 was given as a white foam
in 87% yield: Rf (45:5:3 EtOAc/EtOH/H2O) 0.59; [R]D ) +64.0
(c 0.8, H2O); UV (H2O) 245 nm (ꢀmM 16.6); IR (KBr) νmax 3584,
Information) and δ 1.44 (s, 3 H), 1.29 (s, 3 H); 13C NMR (75.5
MHz, CD3OD) δ 191.0, 113.0, 106.6, 86.7, 81.7, 75.4, 73.1, 56.9,
27.1, 26.4; FABMS m/z 284 (100, [M + Na]+). Anal. Calcd for
C
10H15NO5S: C, 45.96; H, 5.79; N, 5.36; S, 12.27. Found: C,
3324 cm-1 1H NMR (500 MHz, DMSO-d6 + D2O) Table 2
;
45.90; H, 5.62; N, 5.21; S, 12.25.
(Supporting Information); 13C NMR (125.5 MHz, DMSO-d6) δ
187.1, 80.8, 70.6, 68.8, 68.2 (2C), 53.4; FABMS m/z 244 (100,
[M + Na]+). Anal. Calcd for C7H11NO5S: C, 38.00; H, 5.01; N,
6.33; S, 14.49. Found: C, 37.68; H, 5.02; N, 6.20; S, 14.02.
(5S,6S,7R,8R,8a S)-5,6,7,8-Tet r a a cet oxy-2-oxa -3-t h iox-
oin d olizid in e (13). Compound 13 was given as an amorphous
solid in 97% yield: Rf (3:1 EtOAc/petroleum ether) 0.49; [R]D
) -35.8 (c 1.2, CH2Cl2); UV (CH2Cl2) 250 nm (ꢀmM 15.9); IR
5-Am in o-5-d e oxy-1,2-O-isop r op ylid e n e -â-L-id ofu r a -
n ose 5,6-(Cyclic th ioca r ba m a te) (7). Thiocarbonylation of
5 (120 mg, 0.55 mmol) following the above procedure afforded
7 (122 mg, 85%) as an amorphous solid: Rf 0.38; [R]D ) +8.8
(c 0.5, CH2Cl2); UV (MeOH) 242 nm (ꢀmM 21.4); IR (KBr) νmax
1
3360, 1651, 1516 cm-1; H NMR (300 MHz, CD3OD) Table 2
(Supporting Information) and δ 1.44 (s, 3 H), 1.29 (s, 3 H); 13
C
(KBr) νmax 1748 cm-1 1H NMR (500 MHz, CDCl3) Table 2
;
NMR (75.5 MHz, CD3OD) δ 191.4, 113.0, 106.6, 87.1, 83.2,
75.6, 73.2, 57.8, 27.2, 26.4; FABMS m/z 284 (100, [M + Na]+).
Anal. Calcd for C10H15NO5S: C, 45.96; H, 5.79; N, 5.36; S,
12.27. Found: C, 46.35; H, 5.84; N, 5.27; S, 12.49.
(Supporting Information) and δ 2.15, 2.14, 2.12, 2.10 (4 s, each
3 H); 13C NMR (75.5 MHz, CDCl3) δ 187.8,169.6, 168.6, 168.2,
167.7, 76.5, 67.9, 66.0, 65.7, 65.4, 52.4, 20.5 (2C), 20.2 (2C);
FABMS m/z 412 (100, [M
15H19NO9S: C, 46.27; H, 4.92; N, 3.60. Found: C, 46.25; H,
4.88; N, 3.55.
+
Na]+). Anal. Calcd for
5-Am in o-5-d eoxy-1,2-O-isop r op ylid en e-r-D-glu cofu r a -
n ose 5,6-(Cyclic ca r ba m a te) (8). To a stirred solution of 4
(120 mg, 0.55 mmol) in CH2Cl2 (8 mL) were added diisopropyl
ethylamine (DIPEA, 0.96 mL, 5.52 mmol) and triphosgene (82
mg, 0.29 mmol, 1.5 equiv). After 5 min, the solvent was
removed under reduced preassure, and the residue was
purified by column chromatography (1:1 toluene/acetone) to
C
(5R,6S,7R,8R,8a R)-5,6,7,8-Tetr a h yd r oxy-2-oxa -3-oxoin -
d olizid in e (14). Compound 14 was given as a white foam in
93% yield: Rf (2:1:1 BuOH/AcOH/H2O) 0.30; [R]D ) +43.6 (c
1.1, H2O); IR (KBr) νmax 3414, 3148, 1748 cm-1; 1H NMR (500
MHz, D2O) Table 2 (Supporting Information); 13C NMR (125.5
MHz, D2O) δ 160.1, 77.0, 76.0, 74.8, 73.5, 70.0, 55.6; FABMS
m/z 228 (50, [M + Na]+). Anal. Calcd for C7H11NO6: C, 40.98;
H, 5.40; N, 6.82. Found: C, 40.69; H, 5.48; N, 6.82.
give 8 (132 mg, 98%) as an amorphous solid: Rf 0.35; [R]D
)
-24.6 (c 1.0, MeOH); IR (KBr) νmax 3441, 3237, 1707 cm-1; 1H
NMR (300 MHz, CDCl3) Table 2 (Supporting Information) and
δ 6.74 (bs, 1 H), 1.65 (bs, 1 H), 1.51 (s, 3 H), 1.32 (s, 3 H); 13C
NMR (75.5 MHz, CDCl3) δ 158.2 , 110.3, 104.1, 84.4, 81.8, 71.9,
66.8, 49.5, 25.8, 25.1; FABMS m/z 268 (100, [M + Na]+). Anal.
Calcd for C10H15NO6: C, 48.98; H, 6.16; N, 5.71. Found: C,
48.96; H, 6.17; N, 5.73.
(5R,6S,7R,8R,8a R)-5,6,7,8-Tetr a a cetoxy-2-oxa -3-oxoin -
d olizid in e (15). Compound 15 was given as an amorphous
solid in 95% yield: Rf (2:1 EtOAc/petroleum ether) 0.42; [R]D
) +38.3 (c 0.6, CH2Cl2); IR (KBr) νmax 1778, 1751 cm-1 1H
;
NMR (500 MHz, CDCl3) Table 2 (Supporting Information) and
δ 2.12, 2.05, 2.03, 1.98 (4 s, each 3 H); 13C NMR (125.5 MHz,
CDCl3) δ 169.9, 169.7, 169.2, 168.5, 154.1, 72.2, 72.1, 68.9 (2C),
66.8, 52.5, 20.5, 20.4, 20.3, 20.2; FABMS m/z 396 (100, [M +
Na]+). Anal. Calcd for C15H19NO10: C, 48.26; H, 5.13; N, 3.75.
Found: C, 48.25; H, 5.09; N, 3.76.
5-Am in o-5-d e oxy-1,2-O-isop r op ylid e n e -â-L-id ofu r a -
n ose 5,6-(Cyclic ca r ba m a te) (9). Carbonylation of 5 (120 mg,
0.55 mmol) following the above procedure afforded 9 (123 mg,
91%) as an amorphous solid: Rf 0.35; [R]D ) +8.1 (c 0.3,
1
CH2Cl2); IR (KBr) νmax 3354, 1740 cm-1; H NMR (300 MHz,
CDCl3) Table 2 (Supporting Information) and δ 5.95 (bs, 1 H),
1.81 (bs, 1 H), 1.49, 1.32 (2 s, each 3 H); 13C NMR (75.5 MHz,
CDCl3) δ 160.1, 112.0, 104.9, 85.4, 80.6, 75.6, 67.4, 52.0, 26.7,
26.0; FABMS m/z 268 (100, [M + Na]+). Anal. Calcd for
(5S,6S,7R,8R,8a S)-5,6,7,8-Tetr a h yd r oxy-2-oxa -3-oxoin -
d olizid in e (16). Compound 16 was given as a white foam in
96% yield: Rf (2:1:1 BuOH/AcOH/H2O) 0.28; [R]D ) +4.0 (c
1.1, H2O); IR (KBr) νmax 3351, 1734 cm-1; 1H NMR (500 MHz,
D2O) Table 2 (Supporting Information); 13C NMR (125.5 MHz,
D2O) δ 159.0, 76.8, 69.7, 69.5, 67.8, 64.2, 49.2; FABMS m/z
228 (100, [M + Na]+). Anal. Calcd for C7H11NO6: C, 40.98; H,
5.40; N, 6.82. Found: C, 41.00; H, 5.35; N, 6.82.
C
10H15NO6: C, 48.98; H, 6.16; N, 5.71. Found: C, 48.97; H,
6.14; N, 5.71.
Gen er a l P r oced u r e for th e P r ep a r a tion of Red u cin g
2-Oxa in d olizid in es. A solution of the corresponding 5,6-cyclic
(thio)carbamate (6-9, 0.90 mmol) in 90% TFA-H2O (10 mL)
was stirred at room temperature for 2 h monitored by TLC
(45:5:3 EtOAc/EtOH/H2O). The solvent was eliminated under
reduced presure, and the residue was coevaporated with water
several times. The solvent was then neutralized with Amber-
lite IRA-68 (OH-) ion-exchange resin, purified by GPC (Sepha-
dex G-10, 1:1 MeOH/H2O), concentrated, and freeze-dried.
Peracetylated derivatives were obtained by conventional acety-
lation in virtually quantitative yields.
(5S,6S,7R,8R,8a S)-5,6,7,8-Tetr a a cetoxy-2-oxa -3-oxoin -
d olizid in e (17). Compound 17 was given as an amorphous
solid in 85% yield: Rf (2:1 EtOAc/petroleum ether) 0.51; [R]D
) -17.3 (c 0.9, CH2Cl2); IR (KBr) νmax 1767 cm-1 1H NMR
;
(500 MHz, CDCl3) Table 2 (Supporting Information) and δ 2.13,
2.12, 2.08, 2.07 (4 s, each 3 H); 13C NMR (125.5 MHz, CDCl3)
δ 168.6, 168.3 (2C), 168.0, 155.2, 74.4, 65.7 (2C), 65.4, 63.2,
48.4, 20.6, 20.5; FABMS m/z 396 (100, [M + Na]+). Anal. Calcd
for C15H19NO10: C, 48.26; H, 5.13; N, 3.75. Found: C, 48.30;
H, 5.19; N, 3.75.
(5R,6S,7R,8R,8a R)-5,6,7,8-Tetr a h yd r oxy-2-oxa -3-th iox-
oin d olizid in e (10). Compound 10 was given as a white foam
in 93% yield: Rf (45:5:3 AcOEt/EtOH/H2O) 0.61; [R]D ) +0.8
(c 0.8, H2O); UV (H2O) 245 nm (ꢀmM 17.4); IR (KBr) νmax 3372,
5,6-Dia m in o-5,6-d id eoxy-1,2-O-isop r op ylid en e-r-D-glu -
cofu r a n ose 5,6-(Cyclic th iou r ea ) (28). To a solution of
diamine 26 (290 mg, 1.33 mmol) in CH2Cl2 (15 mL) were added
CS2 (0.64 mL) and DCC (290 mg, 1.40 mmol). The reaction
mixture was stirred for 1 h and then concentrated. The
resulting residue was chromatographed on a silica gel column
(45:5:3 EtOAc/EtOH/H2O) to give 26 (273 mg, 79%) as an
amorphous solid: Rf 0.67; [R]D ) -41.8 (c 0.9, MeOH); UV
3291 cm-1 1H NMR (500 MHz, D2O) Table 2 (Supporting
;
Information); 13C NMR (125.5 MHz, D2O) δ 186.4, 77.4, 72.4,
71.6, 71.4, 69.8, 56.1; FABMS m/z 244 (100, [M + Na]+), 211
(40, [M + H]+). Anal. Calcd for C7H11NO5S: C, 38.00; H, 5.01;
N, 6.33; S, 14.49. Found: C, 38.11; H, 5.16; N, 6.31; S, 14.23.

Generalized Anomeric Effect
J . Org. Chem., Vol. 65, No. 1, 2000 143
(MeOH) 240 nm (ꢀmM 14.6); IR (KBr) νmax 3374, 3247, 1528
(H2O) 238 nm (ꢀmM 10.7); IR (KBr) νmax 3452, 3237, 1645 cm-1
1H NMR (500 MHz, D2O) Table 3 (Supporting Information);
13C NMR (125.5 MHz, D2O) δ 183.7, 87.2, 76.2, 75.7, 73.6, 59.6,
58.6; FABMS m/z 221 (70, [M + H]+). Anal. Calcd for
C7H12N2O4S: C, 38.17; H, 5.49; N, 12.72. Found: C, 37.93; H,
5.12; N, 12.47.
;
cm-1 1H NMR (300 MHz, CD3OD) Table 3 (Supporting
;
Information) and δ 1.45, 1.29 (2 s, each 3 H); 13C NMR (75.5
MHz, CD3OD) δ 184.6, 112.9, 106.4, 86.8, 82.7, 75.1, 56.6, 49.2,
27.0, 26.4; CIMS m/z 261 (50, [M + H]+). Anal. Calcd for
C
10H16N2O4S: C, 46.14; H, 6.20; N, 10.76. Found: C, 46.11;
H, 5.94; N, 10.62.
(5R,6S,7R,8R,8a R)-2-N-Acet yl-5,6,7,8-t et r a a cet oxy-2-
a za -3-th ioxoin d olizid in e (33). Compound 33 was given as
an amorphous solid in 85% yield: Rf (1:1 EtOAc/petroleum
ether) 0.44; [R]D +38.0 (c 0.8, H2O); UV (CH2Cl2) 237 nm (ꢀmM
5,6-Dia m in o-5,6-d id eoxy-1,2-O-isop r op ylid en e-r-D-glu -
cofu r a n ose 5,6-(Cyclic u r ea ) (29). A solution of triphosgene
(0.45 mmol) in CH2Cl2 (8 mL) was added dropwise to a stirred
solution of diamine 26 (295 mg, 1.35 mmol) and DIPEA (13.5
mmol) in CH2Cl2 (20 mL). The reaction mixture was stirred
at room temperature and monitored by TLC (45:5:3 EtOAc/
EtOH/H2O) until total consumption of the starting diamine
resulted (1 h). The mixture was then concentrated, and the
crude product was purified by column chromatography using
the same solvent to give 29 (229 mg, 70%) as an amorphous
solid: Rf 0.42; [R]D ) -31.7 (c 0.2, MeOH); IR (KBr) νmax 3441,
14.1); IR (KBr) νmax 1755, 1694 cm-1 1H NMR (500 MHz,
;
CDCl3) Table 3 (Supporting Information) and δ 2.85 (s, 3 H),
2.19, 2.12, 2.08, 2.06 (4 s, each 3 H); 13C NMR (75.5 MHz,
CDCl3) δ 178.6, 171.6, 169.9, 169.5, 169.3, 168.5, 74.2, 71.7,
69.3, 68.5, 52.8, 48.3, 26.6, 20.6, 20.4, 20.2 (2 C); FABMS m/z
453 (100, [M + Na]+). Anal. Calcd for C17H22N2O9S: C, 47.44;
H, 5.15; N, 6.51. Found: C, 47.33; H, 5.22; N, 6.48.
(5R,6S,7R,8R,8a R)-5,6,7,8-Tetr a h yd r oxy-2-a za -3-oxoin -
d olizid in e (34). Compound 34 was given as a white foam in
98% yield: [R]D ) +69.7 (c 0.9, H2O); IR (KBr) νmax 3364, 1690
cm-1; 1H NMR (300 MHz, D2O) (Table 3); 13C NMR (75.5 MHz,
D2O) δ 183.7, 79.9, 74.2, 73.2, 72.0, 53.8, 42.5. Anal. Calcd for
C7H12N2O5: C, 41.18; H, 5.92; N, 13.72. Found: C, 41.18; H,
5.66; N, 13.59.
3393, 1686 cm-1
;
1H NMR (300 MHz, CD3OD) Table 3
(Supporting Information) and δ 1.44, 1.29 (2 s, each 3 H); 13
C
NMR (125.5 MHz, CD3OD) δ 166.3, 112.8, 106.5, 86.9, 83.8,
75.2, 52.2, 45.1, 27.1, 26.4; FABMS m/z 267 (100, [M + Na]+).
Anal. Calcd for C10H16N2O5: C, 49.17; H, 6.60; N, 11.47.
Found: C, 49.14; H, 6.66; N, 11.44.
5,6-Dia m in o-5,6-d id eoxy-1,2-O-isop r op ylid en e-â-L-id -
ofu r a n ose 5,6-(Cyclic th iou r ea ) (30). Thiocarbonylation of
27 (290 mg, 1.33 mmol), following the procedure above decribed
for the preparation of 28 from 26, and purification by column
chromatography (45:5:3 EtOAc/EtOH/H2O) yielded 30 (252 mg,
73%) as an amophous solid: Rf 0.64; [R]D ) +10.3 (c 0.5,
MeOH); UV (MeOH) 241.6 nm (ꢀmM 7.1); IR (KBr) νmax 3125,
(5R,6S,7R,8R,8a R)-5,6,7,8-Tetr a a cetoxy-2-a za -3-oxoin -
d olizid in e (35). Compound 35 was given as an amorphous
solid in 75% yield: Rf (EtOAc) 0.51; [R]D ) +39.0 (c 0.5, CH2-
Cl2); IR (KBr) νmax 3441, 1753, 1647 cm-1; 1H NMR (300 MHz,
CDCl3) Table 3 (Supporting Information) and δ 2.12, 2.07, 2.04,
1.98 (4 s, each H); 13C NMR (125.5 MHz, CDCl3) δ 169.8 (2 C),
169.7 (2 C),157.4, 72.5, 72.0, 69.7, 69.5, 52.7, 42.4, 20.8, 20.5,
20.4, 20.2; FABMS m/z 395 (100, [M + Na]+). Anal. Calcd for
1
1516 cm-1; H NMR (300 MHz, CD3OD) Table 3 (Supporting
Information) and δ 1.43, 1.28 (2 s, each 3 H); 13C NMR (75.5
MHz, CD3OD) δ 184.3, 111.8, 106.5, 87.0, 83.6, 75.6, 57.9, 47.9,
27.1, 26.4; FABMS m/z 283 (100, [M + Na]+), 261 (25, [M +
H]+). Anal. Calcd for C10H16N2O4S: C, 46.14; H, 6.20; N, 10.76.
Found: C, 45.88; H, 6.22; N, 10.50.
C
15H20N2O9: C, 48.39; H, 5.41; N, 7.52. Found: C, 48.19; H,
5.11; N, 7.32.
(5S,6S,7R,8R,8a S)-5,6,7,8-Tetr a h yd r oxy-2-a za -3-th iox-
oin d olizid in e (36). Compound 36 was given as a white foam
1
in 32% yield: [R]D ) +59.7 (c 0.8, H2O); H NMR (300 MHz,
5,6-Dia m in o-5,6-d id eoxy-1,2-O-isop r op ylid en e-â-L-id -
ofu r a n ose 5,6-(Cyclic u r ea ) (31). Thiocarbonylation of
27(295 mg, 1.35 mmol), following the procedure above decribed
for the preparation of 29 from 26, and purification by column
chromatography (45:5:3 EtOAc/EtOH/H2O) yielded 31 (197 mg,
61%): Rf 0.37; [R]D ) +10.9 (c 0.7, MeOH); IR (KBr) νmax 3383,
D2O) Table 3 (Supporting Information); 13C NMR (75.5 MHz,
D2O) δ 180.8, 78.1, 70.0, 69.3, 68.8, 53.1, 42.3; FABMS m/z
221 (40, [M + H]+). Anal. Calcd for C7H12N2O4S: C, 38.17; H,
5.49; N, 12.72. Found: C, 37.99; H, 5.39; N, 12.57.
2-Aza -6-h yd r oxy-3-oxoin d olizin e (37). Compound 37 was
given as a syrup in 41% yield: 1H MNR (500 MHz, D2O) δ 8.71
(d, 1 H, J 5,7 ) 2.4 Hz, H-5), 8.32 (dd, 1 H, J 7,8 ) 8.9 Hz, H-7),
8.19 (d, 1 H, H-8); 13C NMR (75.5 MHz, D2O) δ 156.6, 151.3,
144.2, 137.2, 124.1, 121.4, 45.2; FABMS m/z 173 (40%, [M +
Na]+), 151 (100, [M + H]+); EIMS m/z 150 (100%, M+•). Anal.
Calcd for C7H6N2O2: C, 56.00; H, 4.03; N, 18.66. Found: C,
56.19; H, 3.89; N, 18.47.
1694 cm-1 1H NMR (300 MHz, CD3OD, 313 K) Table 3
;
(Supporting Information) and δ 1.44, 1.30 (2 s, each 3 H); 13
C
NMR (75.5 MHz, CD3OD) δ 166.4, 112.8, 106.5, 87.1, 84.4,
75.6, 53.6, 44.1, 27.1, 26.4; FABMS m/z 267 (100, [M + Na]+).
Anal. Calcd for C10H16N2O5: C, 49.17; H, 6.60; N, 11.47.
Found: C, 49.17; H, 6.55; N, 11.23.
Gen er a l P r oced u r e for th e P r ep a r a tion of Red u cin g
2-Aza in d olizid in es. Deacetalation of the cyclic (thio)ureas
28-30 (0.40 mmol) with 90% TFA-water, as above-described
for the preparation of reducing 2-oxaindolizidines, and puri-
fication by GPC (Sephadex G-10, 1:1 MeOH-H2O) afforded
the corresponding 2-azaindolizidines 32, 34, and 36, respec-
tively. When the cyclic urea derivative of L-idofuranose 31 was
subjected to identical reaction conditions, the 2-azaindolizine
derivative 37 was obtained. Conventional acetylation of 32 and
34 afforded the penta- and tetraacetates 33 and 35, respec-
tively.
Ack n ow led gm en t. We thank the Direccio´n General
de Investigacio´n Cient´ıfica y Te´cnica (grants PB 97/
0747, PB 96-0833, and PB 94-1440-C02-01) for financial
support.
Su p p or tin g In for m a tion Ava ila ble: Tables 2-4 contain-
1
ing H and 13C NMR data for all the new compounds and the
experimental details for the preparation of diamino sugars 26
and 27. This material is available free of charge via the
Internet at http://pubs.acs.org.
(5R,6S,7R,8R,8a R)-5,6,7,8-Tetr a h yd r oxy-2-a za -3-th iox-
oin d olizid in e (32). Compound 32 was given as a white foam
in 98% yield: Rf (MeOH) 0.48; [R]D ) -3.6 (c 1.1, H2O); UV
J O991242O