A new, stereocontrolled approach to iminosugar C-glycosides from L-sorbose

Article abstract of DOI:10.1021/ol0062493

The efficient synthesis of the iminoalditols derivatives 1 and 2 (nojirimycin α-C-glycosides) has been achieved in 10 steps from commercially available 2,3;4,6-di-O-isopropylidene-α-L-sorbofuranose in an overall yield of 23-27%.

Full text of DOI:10.1021/ol0062493

ORGANIC
LETTERS
2000
Vol. 2, No. 19
2971-2974
A New, Stereocontrolled Approach to
Iminosugar C-Glycosides from
L-Sorbose
Ge´raldine Masson, Philippe Compain, and Olivier R. Martin*
Institut de Chimie Organique et Analytique (I.C.O.A.), Faculte´ des Sciences,
BP 6759, 45067 Orle´ans Cedex 2, France
oliVier.martin@uniV-orleans.fr
Received June 26, 2000
ABSTRACT
The efficient synthesis of the iminoalditols derivatives 1 and 2 (nojirimycin r-C-glycosides) has been achieved in 10 steps from commercially
available 2,3;4,6-di-O-isopropylidene-r-L-sorbofuranose in an overall yield of 23−27%.
Owing to their remarkable biological activities, iminosugars
form undoubtedly the most important and attractive class of
carbohydrate mimics reported so far. As potent glycosidase1
and glycosyltransferase2 inhibitors, they promise a new
generation of carbohydrate-based therapeutics for the control
of various diseases including diabetes, cancer, and viral
infection.1 The main drawback associated with the use of
such “azasugars” is their lack of selectivity as R- or
â-glycosidase inhibitors, which may lead to detrimental side
effects in therapeutic applications. Therefore, despite a large
amount of synthetic effort in this area,3 there is still a need
for an efficient methodology to iminosugar C-glycosides of
predictable configuration from simple precursors to facilitate
the discovery of more selective inhibitors. To achieve this
challenging goal and as part of our continuing studies on
azaglycoside mimics,4 we have designed a flexible synthetic
strategy for the preparation of various types of piperidinoses
C-substituted at C-1.
In this paper, we wish to report our preliminary results
concerning the stereocontrolled synthesis of the R-1-C-
(1) (a) Stu¨tz, A. E. Iminosugars as Glycosidase Inhibitors: Nojirimycin
and Beyond; Wiley-VCH: Weinheim, 1999. (b) Winchester, B.; Fleet, G.
W. J. Glycobiology 1992, 2, 199-210.
(2) (a) Kim, Y. J.; Ichikawa, M.; Ichikawa, Y. J. Am. Chem. Soc. 1999,
121, 5829-5830. (b) Qiao, L.; Murray, B. W.; Shimazaki, M.; Schultz, J.;
Wong, C. H. J. Am. Chem. Soc. 1996, 118, 7653-7662. (c) Jefferies, I.;
Bowen, B. R. Bioorg. Med. Chem. Lett. 1997, 7, 1171-1174. (d) Platt, F.
M.; Reinkensmeier, G.; Dwek, R. A.; Butters, T. D. J. Biol. Chem. 1997,
272, 19365-19372. (e) Platt, F. M.; Neises, G. R.; Dwek, R. A.; Butters,
T. D. J. Biol. Chem. 1994, 269, 8362-8365.
(3) (a) Liotta, L. J.; Lee, J.; Ganem, B. Tetrahedron 1991, 47, 2433-
2447. (b) Johnson, C. R.; Golebiowski, A.; Sundram, H.; Miller, M. W.;
Dwaihy, R. L. Tetrahedron Lett. 1997, 653-654. (c) Johns, B. A.; Johnson,
C. R. Tetrahedron Lett. 1998, 39, 749-752 and references cited therein.
(d) Fuchss, T.; Streicher, H.; Schmidt, R. R. Liebigs Ann./Recl. 1997, 1315-
1321. (e) Fuchss, T.; Schmidt, R. R. Synthesis 2000, 259-264. (f) Zhu,
Y.-H.; Vogel, P. J. Org. Chem. 1999, 64, 666-669 and references cited
therein. (g) Cipolla, L.; Lay, L.; Nicotra, F.; Pangrazio, C.; Panza, L.
Tetrahedron 1995, 51, 4679-4690.
(4) Martin, O. R. In Carbohydrate Mimics: Concepts and Methods;
Chapleur, Y., Ed.; Wiley-VCH: New York, 1997.
(5) Hughes, A. B.; Rudge, A. J. Nat. Prod. Rep. 1994, 135-162.
(6) (a) Robinson, K. M.; Begovic, M. E.; Rhinehart, B. L.; Heinke, E.
W.; Ducep, J.-B.; Kastner, P. R.; Marshall, F. N.; Danzin, C. Diabetes 1991,
40, 825-830. (b) Dong, W.; Jespersen, T.; Bols, M.; Skrydstrup, T.; Sierks,
M. R. Biochemistry 1996, 35, 2788-2795. (c) Heightman, T. D.; Vasella,
A. T. Angew. Chem., Int. Ed. Engl. 1999, 38, 750-770 and references cited
therein.
(7) (a) Look, G. C.; Fotsch, C. H.; Wong, C.-H. Acc. Chem. Res. 1993,
26, 182-190. (b) Saavedra, O. M.; Martin, O. R. J. Org. Chem. 1996, 61,
6987-6993.
(8) Paulsen, H.; Sangster, I.; Heyns, K. Chem. Ber. 1967, 100, 802-
815.
(9) Whistler, R. L.; Doner, L. W.; Kosik, M. Methods Carbohydr. Chem.
1972, 6, 411-412.
10.1021/ol0062493 CCC: $19.00 © 2000 American Chemical Society
Published on Web 08/24/2000

substituted-1-deoxynojirimycin 1 and 2 and analogues. As
outlined in Scheme 1, our retrosynthetic analysis takes
Scheme 2a
Scheme 1
advantage of the chirality of L-sorbose, which provides the
three stereogenic centers (C-2, C-3, and C-4) of our target I
and acts as an internal chiral auxiliary. It is noteworthy that
L-sorbose has been used as the starting material for the first
synthesis, as well as for some of the shortest syntheses, of
1-deoxynojirimycin5 but not for the preparation of chain-
extended analogues. The diastereoselective chain extension
of imine IV is the key step of our synthetic strategy for the
following two reasons: first, the careful choice of reaction
conditions can give access selectively to the two possible
epimers at C-6 in III, i.e., the precursors of the R- or
â-epimers of imino-C-glycoside I. Second, at this stage,
structural diversity may be introduced at the “anomeric”
position by using the wide library of organometallic nucleo-
philes available. This point is particularly significant if one
wants to explore the affinity of the aglycon binding site
within a range of glycosidases to increase the selectivity of
potential inhibitors.6 The last steps of the synthetic plan
consisted in generating the final imino-sugar C-glycosidic
structure (I) by way of the intramolecular reductive amination
of the latent keto function of the chain-extended amino-
sorbofuranose derivative, a reaction that was expected to be
highly stereoselective and to give the desired epimer in the
D-series.7 Finally, an orthogonal protecting group strategy
was designed to facilitate the differentiation of the sugar
hydroxyl groups, notably at C-3 (structure I), thus opening
the way to oligosaccharide analogues or to other iminosugars
by controlled epimerisation of one of the OH group (see
Scheme 1).
The aldehyde 5 was obtained from 4 in 85% yield by
standard PCC oxidation.10 Finally, condensation of 5 with
1.1 equiv of benzylamine in dichloromethane in the presence
of molecular sieves (4 Å)11 afforded quantitatively the imine
6, as judged by proton NMR spectroscopy.
Having in hand the key intermediate 6, we first studied
the diastereoselectivity of the addition of vinylmagnesium
bromide to the CdN bond (Scheme 3).12 The resulting allylic
Scheme 3a
This approach required an efficient and easy access to
multigram quantities of imine 6. This key intermediate was
prepared from commercially available 2,3;4,6-di-O-isopro-
pylidene-R-L-sorbofuranose in seven steps and 69% overall
yield (Scheme 2). Quantitative benzylation of the primary
alcohol, followed by selective deprotection of the six-
membered isopropylidene acetal under aqueous acidic condi-
tions, in the presence of acetone, gave the diol 3 in 97%
yield.8 Benzylation of the secondary hydroxyl group was
performed by the following three-step procedure: selective
protection of the primary alcohol function of 3 as a trityl
ether, benzylation of the secondary OH, and cleavage of the
trityl group using a solution of HBr in glacial acetic acid.9
amine was of particular interest owing to the various
synthetic possibilities offered by the transformation of the
CdC double bond at different stages of the synthesis. A
solution of the N-benzylimine 6 in Et2O was added dropwise
to a cooled solution of 5 equiv of the Grignard reagent in
Et2O at 0 °C. The reaction mixture was slowly warmed to
room temperature and stirred for 24 h. Proton NMR analysis
(10) Nicotra, F.; Panza, L.; Russo, G.; Zucchelli, L. J. Org. Chem. 1992,
57, 2154-2158.
(11) Feldman, K. S.; Mingo, P. A.; Hawkins, P. C. D. Heterocycles 1999,
51, 1283-1294.
(12) Bloch, R. Chem. ReV. 1998, 98, 1407-1438 and references cited
therein.
2972
Org. Lett., Vol. 2, No. 19, 2000

of the crude product after workup revealed the presence of
a single diastereomer. The pure amine 7a was obtained in
74% yield after purification by flash chromatography. The
addition of allylmagnesium bromide (4 equiv) to 6 was also
found to be highly diastereoselective and proceeded to give
8a in a yield of 87%. The absolute configuration of the newly
created stereocenter was unambiguously established to be R
at the stage of the cyclic product 9.
Scheme 5a
The high stereoselectivity of the addition of the organo-
magnesium species to imine 6 can be best rationalized by
the chelated intermediate A12 involving as ligands the ring
oxygen atom of the sorbofuranose moiety and the nitrogen
atom of the N-benzylimine (Scheme 4). The propensity of
The relative configuration of the substituents in the
piperidine ring system was unambiguously established by
the 1H NMR spectra (COSY and NOESY) of 9 and 10 and
of their respective acetates 1 and 2,16 obtained in excellent
yield by acetylation using Ac2O in pyridine at 40 °C. As
shown by the data,16,17 both compounds have a pseudo-R-
D-gluco configuration and adopt predominantly a chair
conformation in which all substituents are in equatorial
position except the allyl/vinyl group.
Scheme 4
The intramolecular reductive amination of the alternate
6-benzylamino-6-deoxy-L-sorbose derivative 7b gave a labile
product, the configuration of which could not yet be
the ring oxygen in furanose derivatives to act as a coordinat-
ing Lewis base in additions of organometallic species to
5-aldehydo-pentofuranose derivatives is well documented.13
To modifiy the stereoselectivity of the reaction, we performed
the addition in the presence of a monodentate Lewis acid
[BF3‚Et2O (5 equiv)14 at -78 °C] with the goal of precom-
plexing the imine and suppressing chelation effects. Addition
of vinylmagnesium bromide (4 equiv) at -50 °C to 6 thus
preactivated and warming up the reaction mixture to -10
°C afforded, after purification, the epimeric amine 7b with
the opposite configuration at C-6, as expected according to
the open transition state model B12 depicted in Scheme 4.
The addition of allylmagnesium bromide to the imine 6 under
the same conditions gave, however, the expected product
with a very low degree of stereoselectivity.
(16) Data for 9: 1H NMR (500 MHz, CDCl3) δ (ppm) 3.02 (dt, 1H,
J5,6 ) 4.4, J5,4 ) 8.8 Hz, H-5), 3.39 (dd, 1H, J1,2 ) 4.9, J1,1′ ) 9.8 Hz,
H-1), 3.47 (t, 1H, J3,2 ) J3,4 ≈ 8.5 Hz, H-3), 3.55 (d, 1H, J ) 14.2 Hz,
NCHAHBPh), 3.75 (t, 1H, J4,5 ) J4,3 ≈ 8.5 Hz, H-4), 3.75 (dd, 1H, J2,3
)
9.2, J2,1 ≈ 5 Hz, H-2), 3.77 (dd, 1H, J6a,6b ≈ 10, J6a,5 ) 3.9 Hz, H-6a),
3.82 (dd, 1H, J6b,6a ) 10.3, J6b,5 ) 4.9 Hz, H-6b), 4.00 (d, 1H, J ) 14.2
Hz, NCHAHBPh), 4.46 (AB, 2H, J ) 11.7 Hz, OCH2Ph), 4.83 (AB, 2H, J
) 11.7 Hz, OCH2Ph), 5.19 (dd, 1H, Jgem ) 1.9, Jtrans ) 17.1 Hz, CHd
CHAHB), 5.45 (dd, 1H, Jgem ) 1.9, Jcis ) 10.5 Hz, CHdCHAHB), 5.99 (dt,
1H, J1,1′ ) 10, Jtrans ) 17.0, Jcis ) 10.5 Hz, CH1′dCH2), 7.05-7.50 (m,
15H, 3 C6H5); 13C NMR (62.9 MHz, CDCl3, δ CDCl3 ) 77.16) δ (ppm)
52.98 (NCH2Ph), 59.78, 63.40 (C-1, 5), 70.14 (C-6), 71.59, 72.67 (C-2, 4),
73.44 and 74.49 (2 OCH2C6H5), 83.53 (C-4), 122.39 (CHdCH2), 126.9-
128.6 (CArH), 131.19 (CHdCH2), 137.95, 138.85, 139.94 (3 CAr); MS m/z
460.0 [(M + H)+]. Data for 1: 1H NMR (500 MHz, CDCl3) δ (ppm) 1.86
(s, 3H, OAc), 1.90 (s, 3H, OAc), 3.16 (dt, 1H, J5,6 ) 3.9, J5,4 ) 9.8 Hz,
H-5), 3.51 (d, 2H, J6,5 ) 3.9 Hz, 2 H-6), 3.59 (dd, 1H, J1,2 ) 5.4, J1,1′
)
8.8 Hz, H-1), 3.68 (d, 1H, J ) 14.2 Hz) and 4.08 (d, 1H, J ) 14.2 Hz)
(AB, NCH2Ph), 3.76 (t, 1H, J3,4 ) J3,2 ) 9.5 Hz, H-3), 4.33 (d, 1H, J )
11.7 Hz) and 4.38 (d, 1H, J ) 11.7 Hz) (AB, OCH2Ph), 4.58 (d, 1H, J )
11.7 Hz) and 4.67 (d, 1H, J )11.7 Hz) (AB, OCH2Ph), 5.05 (dd, 1H, J2,3
) 9.8, J2,1 ) 5.4 Hz, H-2), 5.12 (br d, 1H, Jtrans ) 17.1 Hz, CHdCHAHB),
5.14 (t, 1H, J4,3 ) J4,5 ) 9.8 Hz, H-4), 5.36 (br d, 1H, Jcis ) 10.2 Hz,
CHdCHAHB), 5.92 (ddd, 1H, J1,1′ ) 8.8, Jtrans ) 17.1, Jcis ) 10.2 Hz,
CH1′dCH2), 7.15-7.35 (m, 15H, 3 C6H5); 13C NMR (62.9 MHz, CDCl3)
δ (ppm) 21.05, 21.07 (2 OAc), 52.66 (NCH2Ph), 57.91, 59.85 (C-1, 5),
69.04 (C-6), 72.07, 72.84 (C-2, 4), 73.19, 74.22 (2 OCH2Ph), 79.09 (C-3),
121.45 (CHdCH2), 126.0-128.67 (CArH), 130.66 (CHdCH2), 137.93,
138.54, 139.72 (3 CAr), 169.79, 170.07 (2 OAc); MS m/z 544.5 [(M +
H)+]; [R]20D +18.5 (c 1.2, CHCl3). Data for 2: 1H NMR (500 MHz, CDCl3)
δ (ppm) 1.89 (s, 3H, OAc), 1.95 (s, 3H, OAc), 2.24 (m, 1H) and 2.37 (m,
1H) (CH2CHdCH2), 3.19 (m, 2H, H-5 and H-1), 3.58 (m, 2H, H-6), 3.80
(t, 1H, J3,4 ) J3,2 ) 8.3 Hz, H-3), 3.96 (AB, 1H, J ) 14.2 Hz, NCH2Ph),
4.33 (near s, 2H, OCH2Ph), 4.62 (d, 1H, J ) 11.7 Hz) and 4.66 (d, 1H, J
) 11.7 Hz) (OCH2Ph), 4.96 (br d, 1H, Jcis ) 10.3, Jgem ) 1.5 Hz, CHd
CHAHB), 4.99 (dd, 1H, Jgem ) 1.5, Jtrans ) 17.1 Hz, CHdCHAHB), 5.15
(dd, 1H, J2,1 ) 5.4, J2,3 ) 8.8 Hz, H-2), 5.23 (t, 1H, J4,5 ≈ 8.5, J4,3 ≈ 8.5
Hz, H-4), 5.62 (m, 1H, J ) 6.8, 6.8, 10.3, 17.1 Hz, CH2′dCH2), 7.10-
7.45 (m, 15H, 3 C6H5); 13C NMR (62.9 MHz, CDCl3) δ (ppm) 21.17 (2
OAc), 30.71 (CH2CHdCH2), 52.64 (NCH2Ph), 55.85, 56.03 (C-1, 5), 69.59
(C-6), 71.44, 71.84 (C-2, 4), 73.18, 74.22 (2 OCH2Ph), 78.65 (C-3), 116.11
(CHdCH2), 126.00-128.5 (CArH), 136.10 (CHdCH2), 138.05, 138.44,
The intramolecular reductive amination of the amino-
sorbose hemiketal liberated upon acidic hydrolysis of the
isopropylidene group was then investigated.
The best experimental conditions consisted in using 90%
aqueous CF3COOH at room temperature for 24 h15 for the
hydrolysis step, followed by evaporation of the reaction
mixture and treatment of the crude intermediates with
NaBH3CN (4 equiv) in glacial acetic acid for the reduction
step. This sequence of reactions gave, after purification of
the final product by flash chromatography, the expected
diastereomerically pure piperidinols 9 and 10 in moderate
yield (48-49%) from the 6(R) amino-sorbose derivatives 7a
and 8a, respectively (Scheme 5).
(13) Inch, T. D. AdV. Carbohydr. Chem. Biochem. 1972, 27, 191-225
and references cited therein.
(14) van Delft, F. L.; de Kort, M.; van der Marel, G. A.; van Boom, J.
H. J. Org. Chem. 1996, 61, 1883-1885.
(15) Nortey, S. O.; Wu, W.-N.; Maryanoff, B. E. Carbohydr. Res. 1997,
304, 29-38.
140.12 (3 CAr), 169.97, 170.08 (2 OAc); MS: m/z 558.5 [(M + H)+]; [R]20
D
+26 (c 1, CHCl3).
(17) For example, in the case of the piperidinol 9, crucial NOE effects
were observed between H-1′ and H-5 and between H-1′ and H-3.
Org. Lett., Vol. 2, No. 19, 2000
2973

established. Further work on the epimers 7b and 8b, which
should give access to â-configured imino-C-glycosyl com-
pounds is currently in progress.
reaction sequence proceeds in an overall yield of 23-27%.
This strategy is currently being extended to other types of
organometallic nucleophiles, and the piperidinose moiety is
being modified with the goal of generating new types of
glycosyltransferase inhibitors.
In conclusion, the diastereoselective addition of organo-
metallic reagents onto the L-sorbose-derived imine 6, fol-
lowed by an internal reductive amination, provides a
stereocontrolled approach to nojirimycin R-C-glycosides and
related compounds; this is a very useful procedure since
R-substituted piperidinose-C-glycosides are generally of more
difficult access than the â-epimers. For example, double
internal reductive amination procedures yielded exclusively
1,5-cis-disubstituted piperidinose derivatives.7b The 10-step
Acknowledgment. The authors express their gratitude to
Jeroˆme Blu for assistance with synthetic work. Financial
support of this work by a grant from CNRS (Physique et
Chimie du Vivant) is also gratefully acknowledged.
OL0062493
2974
Org. Lett., Vol. 2, No. 19, 2000