New azadisaccharide analogs as potential antidiabetics

Article abstract of DOI:10.1016/S0040-4039(02)02016-6

The synthesis of various aminocyclitols displaying a N-substituted (or unsubstituted) polyhydroxycyclohexylamine structure is described. The strategy relies on two key steps: a tandem alkylation-cyclization of a C₂-symmetrical bis-epoxide deriv

Full text of DOI:10.1016/S0040-4039(02)02016-6

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 43 (2002) 8285–8288
New azadisaccharide analogs as potential antidiabetics
Christine Gravier-Pelletier, William Maton and Yves Le Merrer*
Universite´ Rene´ Descartes, Laboratoire de Chimie et Biochimie Pharmacologiques et Toxicologiques, UMR 8601 CNRS,
45, rue des Saints-Pe`res, 75270 Paris Cedex 06, France
Received 24 July 2002; accepted 16 September 2002
Abstract—The synthesis of various aminocyclitols displaying a N-substituted (or unsubstituted) polyhydroxycyclohexylamine
structure is described. The strategy relies on two key steps: a tandem alkylation–cyclization of a C2-symmetrical bis-epoxide
derived from
D-mannitol and a reductive amination with several amines of the resulting polyhydroxycyclohexanone. Reduction of
the latter also allows a rapid access to the corresponding carbasugars. © 2002 Published by Elsevier Science Ltd.
Due to the potential therapeutic applications of glycosi-
dases or glycosyltransferases inhibitors1 in diabetes,
cancer or viral infections, we have developed an inten-
sive program2 concerning the synthesis of carbohydrate
mimics. Our attention has first been focused on hetero-
cyclic compounds such as imino-,3 thio-,4 guanidino-
sugars5 or azadisaccharide analogs6 and we are now
interested in the design and the synthesis of aminocycli-
tols displaying various size for the ring. Thus, we
recently described access to C8-cyclitols via metathesis
of 1,9-diene7 and to C7-aminocyclitols.8 Considering the
powerful a-glucosidase inhibitory activity of both vogli-
bose and acarbose (Fig. 1) which are used in the
treatment of non-insulino-dependant diabetes mellitus,9
our goal is now to develop straightforward routes to
C6-aminocyclitols analogous to these compounds.
Due to their potential activity, intensive synthetic
efforts10 have been directed towards carbohydrate mim-
ics and their biological activity is believed to arise from
their ability to be protonated at physiological pH
allowing them to fit with the putative oxocarbenium
ion-like transition state. The charge can be localized
either at endocyclic oxygen atom as mimicked (Fig. 1)
by deoxynojirimycin (DNJ) or miglitol (another antidi-
abetic compound),11 or at anomeric carbon atom
as illustrated by isofagomine or finally at the exo-
cyclic oxygen atom as exemplified by voglibose or
acarbose.
The retrosynthesis of the target compounds is outlined
in Fig. 2 and relies on either the reduction or the
reductive amination of the enantiopure polyhydroxycy-
clohexanone resulting from
cyclization reaction of C2-symmetrical bis-epoxide
coming from -mannitol. According to the strategy we
a tandem alkylation–
D
recently described,12 the later reaction which involves
Brook rearrangement can be directed to the major
formation of either dithioketal of cyclohexanone or
cycloheptanone depending on the nature of the protec-
tive group for the central diol of the bis-epoxide. Thus,
benzyl ethers preferentially led to C7-rings while an
acetonide moiety mainly furnished the C6-ring. We
already described access to C7-aminocyclitols from C7-
rings8 and we now wish to report our results concerning
the synthesis of C6-carbasugars and C6-aminocyclitols
from C6-rings.
Figure 1.
Keywords: glycosidases; pseudo-azasaccharide; carbasugars; aminocy-
clitols; reductive amination.
* Corresponding author. Tel.: +33(0)142862176; fax: +33(0)
142868387; e-mail: yves.le-merrer@biomedicale.univ-paris5.fr
0040-4039/02/$ - see front matter © 2002 Published by Elsevier Science Ltd.
PII: S0040-4039(02)02016-6

8286
C. Gra6ier-Pelletier et al. / Tetrahedron Letters 43 (2002) 8285–8288
configuration at the newly created chiral carbon atom
1
was assigned by H NMR studies.16 The same reaction
done at −78°C for 3 hours led to a slight increase in
6a/6b to 96/4 in an identical yield. This result is in good
agreement with the described major formation of an
equatorial hydroxyl group promoted by NaBH4.17
Attempts to reverse the observed diastereoselectivity in
the presence of bulky hydride were then examined.
Surprisingly, DIBAL-H in THF at −78°C furnished the
same major compound as previously (6a/6b=78/22,
43% yield), however L-Selectride® and K-Selectride®
(−78°C, 30 min) in THF allowed the expected inversion
of the diastereoselectivity of the reaction and led to the
major formation of 6b (6a/6b up to 39/61, 50% yield).
The lower yield observed with DIBAL-H and L- or
K-Selectride® compared to that obtained with NaBH4
can be explained by the partial decomposition of inter-
mediate ketone 5. It has to be noted that the reduction
with K-Selectride® was also carried out on an isolated
pure cyclohexanone 5 sample and led to the same
isolated yield of 6a/6b (4/6, respectively, 50% yield).
Figure 2. Retrosynthetic analysis.
The preparation of polyhydroxycyclohexanes (Scheme
1) involved the regiospecific opening of the 1,2:5,6-
dianhydro-3,4-O-methylethylidene-L-iditol 1 by the
tert-butyllithium generated lithio derivative of 2-tert-
butyldimethylsilyl-1,3-dithiane in a 9:1 mixture of
THF:HMPA at −30°C, followed by in situ Brook rear-
rangement and carbocyclization which led to the
dithioketal of cyclohexanone 2 in 72% yield. It has to
be noted that the choice of
L-ido bis-epoxide was
directed by the absolute configuration of the asymmet-
ric carbon atoms of the resulting cyclohexane closed to
that of the compounds displaying interesting antidia-
betic properties. Protection of the primary alcohol func-
In order to examine an eventual induction by the free
alcohol function of the compound 2 during the reduc-
tion step, hydrolysis of its dithioketal moiety was car-
ried out in the same conditions as previously (NBS,
acetone–H2O, −50°C) but gave the corresponding
ketoalcohol 4 in only 10% yield revealing that these
conditions were not suitable with the free alcohol.
Alternatively, the alcohol 4 could be obtained by
deprotection of the silyl ether 5 with HF–pyridine in
40% non-optimized yield. However, the reduction of
this ketoalcohol 4 with NaBH4 in EtOH, at −78°C
afforded a 75/25 mixture of the corresponding diols
7a/7b revealing no modification of the diastereoselectiv-
ity of the reduction.
tion of
2
(tert-butyldimethylsilylchloride, DMF)
afforded the O-silylated derivative 3 in quantitative
yield. Then the dithioketal hydrolysis of 3 in the pres-
ence of N-bromosuccinimide in aqueous acetone13 at
−50°C, to avoid side reactions, gave the corresponding
cyclohexanone 514 which was submitted to the next
reduction without purification (see Table 1 for the
various conditions used for the reduction of the car-
bonyl moiety of 5).
Thus, in the presence of sodium borohydride in EtOH,
at 20°C, a 82/18 mixture of the alcohols 6a/6b15 was
obtained in 80% yield. These products were easily sepa-
rated by flash chromatography and the absolute
We next turned to the preparation of the targeted
aminocyclitols via the reductive amination of the O-
silylated keto-alcohol 5 (Scheme 2) in the presence of
Scheme 1.
Table 1. Reduction of ketone 4 or 5 with hydride reagents
Substrate
Reagents and conditions
Ratioa
Yield (%)
5
5
5
5
5
4
NaBH4, EtOH, 20°C, 20 h
NaBH4, EtOH, −78°C, 3 h
DIBAL-H, THF, −78°C, 1.5 h
L-Selectride®, THF, −78°C, 30 min
K-Selectride®, THF, −78°C, 30 min
NaBH4, EtOH, −78°C, 3 h
6a/6b=82/18
6a/6b=96/4
6a/6b=78/22
6a/6b=39/61
6a/6b=36/64
7a/7b=75/25
80
80
43
50
50
40
a The a/b ratio was determined by 1H NMR spectrum of the crude mixture prior to flash chromatographic separation.

C. Gra6ier-Pelletier et al. / Tetrahedron Letters 43 (2002) 8285–8288
8287
Scheme 2. Reductive amination and deprotections on the major compounds.
hydrogenolysis of the benzyl ethers in the presence of
palladium black in acetic acid to lead to the pseudo-
azadisaccharide 14a (70% yield). All compounds were
purified by ion exchange chromatography (Dowex
50X8-100).
various primary amines such as benzylamine, 1,3-di-
tert-butyldimethylsilyloxy-2-propylamine or methyl 6-
amino-6-deoxy-2,3,4-tri-O-benzyl-a-D
-glucopyranoside,18
respectively, chosen to afford unsubstituted C6-
aminocyclitol, an analog of voglibose and an analog of
disaccharide displaying partial acarbose structure.
Thus, the reductive amination was carried out under
Mattson et al. conditions19 and involved imine forma-
tion catalyzed by the mild and effective titanium(IV)
isopropoxide as Lewis acid followed by its in situ
In conclusion, we described powerful routes to new
carbasugars or aminocyclitols via respective stereoselec-
tive reduction or reductive amination of an enantiomer-
ically pure polyhydroxycyclohexanone easily obtained
by a tandem alkylation–cyclization of a bis-epoxide
reduction with sodium cyanoborohydride in EtOH.15
A
derived from
D-mannitol followed by careful dithioke-
mixture of the two possible aminocyclitols (8a/8b, 9a/9b
and 10a/10b, respectively) was obtained in moderate to
good yield (82, 58 and 61%, respectively). In all cases,
the ratio of the diastereomeric mixture (60/40 to 70/30)
was in favor of the amino group in equatorial posi-
tion,20 as precedently observed for the NaBH4 reduc-
tion of the cyclohexanone 5. Again, to examine the
possible induction of a free alcohol on the diastereose-
lectivity of the reductive amination, the keto-alcohol 4
was submitted to reductive amination with benzylamine
and afforded a 75/25 mixture of the corresponding
aminocyclitols 11a/11b in 52% yield, thus revealing no
modification in the diastereoselectivity of the reductive
amination.
tal hydrolysis. According to this strategy, the syn-
thetized aminocyclitols can be N-substituted by various
R-groups, such as R=CH(CH2OH)2 or methyl a-D-
glucoside as, respectively, voglibose or acarbose analogs
or unsubstituted to serve as a reference concerning the
eventual biological effect of N-substitution of such
derivatives. The biological activity of the synthetized
aminocyclitols towards various glycosidases and espe-
cially a-glucosidase will be evaluated soon and the
results of biological evaluation will be published later
on.
Acknowledgements
The deprotection of the aminocyclitols 9a or 9b was
easily achieved in a single step in the presence of a 9/1
mixture of trifluoroacetic acid/H2O to afford the
expected aminocyclitols 13a or 13b in good yield (86 or
80%, respectively). Hydrogenolysis of the N-benzyl
bond of the compound 8a in the presence of Pearlman
catalyst (H2, Pd(OH)2, 20°C) was followed by acidic
hydrolysis (TFA/H2O=9/1) to give the N-unsubsti-
tuted aminocyclitol 12a (54% yield). Deprotection of
the disaccharide analog 10a required first, acidic
hydrolysis (TFA/H2O/THF=1/1/3, 20°C) of both the
acetonide moiety and the silyl ethers followed by
Assistance of Dr. G. Bertho (LCBPT) in NMR studies
is gratefully acknowledged.
References
1. (a) For recent reviews on carbohydrates and glycobiol-
ogy, see: Hurtley, S.; Service, R.; Szuromi, P. Science
2001, 291, 2337 and references cited therein; (b) Mc
Carther, J. D.; Withers, S. G. Curr. Opin. Struc. Biol.
1994, 4, 885–892; (c) Ly, H. D.; Withers, S. G. Annu. Rev.

8288
C. Gra6ier-Pelletier et al. / Tetrahedron Letters 43 (2002) 8285–8288
10. (a) Carbohydrate Mimics, Concepts and Methods; Chap-
Biochem. 1999, 68, 487–522; (d) Elbein, A. D.; Molyneux,
R. J. Comprehensive Natural Products Chemistry; Pinto,
B. M., Ed.; Barton, D. H. R.; Nakanishi, K.; Meth-
Cohn, O., Ser. Eds.; Elsevier: UK, 1999; Vol. 3, Chapter
7.
leur, Y., Ed.; Wiley-VCH: New York, 1998; (b) Watson,
A. A.; Fleet, G. W. J.; Asano, N.; Molyneux, R. J.; Nash,
R. J. Phytochemistry 2001, 56, 265–295; (c) Asano, N.;
Nash, R. J.; Molyneux, R. J.; Fleet, G. W. J. Tetra-
hedron: Asymmetry 2000, 11, 1645–1680; (d) Iminosugars
as Glycosidases Inhibitors: Nojirimycin and Beyond; Stu¨tz,
A. E., Ed.; Wiley-VCH: New York, 1999.
2. Le Merrer, Y.; Gravier-Pelletier, C.; Depezay, J.-C.
Recent Res. Devel. Org. Chem. 2001, 5, 257–278 and
references cited therein.
11. Gaudillie`re, B. In Annu. Rep. Med. Chem.; Doherty, A.
M., Ed.; Academic Press: San Diego, 1999; Vol. 34, p.
325.
12. Le Merrer, Y.; Gravier-Pelletier, C.; Maton, W.; Numa,
M.; Depezay, J.-C. Synlett 1999, 1322–1324.
13. Corey, E. J.; Erickson, B. W. J. Org. Chem. 1971, 36,
3553–3560.
14. Due to the instability of the cyclohexanone 5 in basic
conditions, it is necessary to avoid basic conditions dur-
ing work-up.
15. Satisfactory analytical and/or spectroscopic data were
obtained on all new compounds.
16. For 6a: J1,2=9.1 Hz and for 6b: J1,2 <2 Hz.
17. Wigfield, D. C. Tetrahedron 1979, 35, 449–462.
18. This amine was prepared according to known routes. See:
Whistler, R. L.; Wolfrom, M. L. Methods Carbohydr.
Chem. 1963, 2, 244–245.
3. (a) Poitout, L.; Le Merrer, Y.; Depezay, J.-C. Tetra-
hedron Lett. 1994, 35, 3293–3296; (b) Le Merrer, Y.;
Poitout, L.; Depezay, J.-C.; Dosbaa, I.; Geoffroy, S.;
Foglietti, M.-J. Bioorg. Med. Chem. 1997, 5, 519–533.
4. Le Merrer, Y.; Fuzier, M.; Depezay, J.-C.; Dosbaa, I.;
Foglietti, M.-J. Tetrahedron 1997, 53, 16731–16746.
5. Le Merrer, Y.; Gauzy, L.; Gravier-Pelletier, C.; Depezay,
J.-C. Bioorg. Med. Chem. 2000, 8, 307–320.
6. Le Merrer, Y.; Sanie`re, M.; McCort, I.; Dupuy, C.;
Depezay, J.-C. Tetrahedron Lett. 2001, 42, 2661–2663.
7. Gravier-Pelletier, C.; Andriuzzi, O.; Le Merrer, Y. Tetra-
hedron Lett. 2002, 43, 245–248.
8. Gravier-Pelletier, C.; Maton, W.; Lecourt, T.; Le Merrer,
Y. Tetrahedron Lett. 2001, 42, 4475–4478.
9. (a) Balfour, J. A.; Mc Tavish, D. Drugs 1993, 46, 1025–
1054; (b) Shapiro, J. A.; Seaton, T. B. Arch. Inter. Med.
1994, 154, 2442–2448; (c) Cheng, X.-M. In Annu. Rep.
Med. Chem.; Bristol, J. A., Ed.; Academic Press: San
Diego, 1995; Vol. 30, p. 313.
19. Mattson, R. J.; Pham, K. M.; Leuck, D. J.; Cowen, K. A.
J. Org. Chem. 1990, 55, 2552–2554.
20. For 8a: J1,2=11.8 Hz and for 8b: J1,2 <2 Hz.