Novel synthesis of carbamate-linked oligosaccharides by a modified curtius rearrangement

Article abstract of DOI:10.1002/ejoc.200600993

We describe a novel stereospecific synthesis of various carbamate-linked disaccharides using sugar carboxylic acids and sugar alcohols by a modified Curtius rearrangement. Furthermore, we applied this method to the synthesis of carbamate-linked oligosacch

Full text of DOI:10.1002/ejoc.200600993

SHORT COMMUNICATION
DOI: 10.1002/ejoc.200600993
Novel Synthesis of Carbamate-Linked Oligosaccharides by a Modified Curtius
Rearrangement
Daisuke Sawada,[a] Shinya Sasayama,[a] Hideyo Takahashi,[a] and Shiro Ikegami*[a]
Keywords: Carbamates / Carbohydrates / Dendrimers / Rearrangements / Stereospecific synthesis
We describe a novel stereospecific synthesis of various carba-
mate-linked disaccharides using sugar carboxylic acids and
bamate-linked oligosaccharides including a dendritic mole-
cule.
sugar alcohols by
Furthermore, we applied this method to the synthesis of car-
a
modified Curtius rearrangement.
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2007)
Introduction
a glycoside bond, and we have achieved the stereospecific
synthesis of carbamate-linked disaccharides and provide its
applications in oligosaccharide synthesis.[2,3]
Sugars equipped with several hydroxy groups stereo-
specifically can produce a large number of oligosaccharides
by variation of their bonding. It is well known that oligo-
The carbamate-linked disaccharide was first reported by
saccharides that interact with glycopeptides, glycolipids etc. Laupichler et al. in 1992,[2a] and recently, Ichikawa et al.
play an important role in a living body.[1] The glycoside and Prosperi et al. successively reported the stereospecific
bonds are, however, labile towards various chemical trans- synthesis of disaccharides; however, there is no example of
formations, and their stereoselective formation is often the synthesis of saccharides larger than disaccharides.[2c,2d]
troublesome. Inevitably, those problems restrict the synthe- Their synthetic methods utilize a glycosyl isocyanate as an
sis of oligosaccharides. If it were possible to substitute a intermediate, whose precursor is a glycosyl isocyanide.[4]
glycoside bond by a more stable linkage and to form this More recently, we reported the synthesis of carbamate- and
link stereoselectively, novel complex oligosaccharides would urea-linked glycoconjugates by a modified Curtius re-
be synthesized, which would have unique structures, new arrangement.[3] The advantages of our synthetic method are
functions, and interesting biological activity. Presently we as follows: the reaction shows high reactivity and proceeds
have focused on a carbamate linkage as a substitution for stereospecifically in a simple one-pot procedure by treating
Scheme 1. Stereospecific synthesis of the both anomeric isomers of carbamate-linked saccharides by using a glycosyl carboxylic acid and
a sugar alcohol.
a glycosyl carboxylic acid and a sugar alcohol with diphenyl
phosphoryl azide (DPPA).[5] Not only the elongation of the
[a] School of Pharmaceutical Sciences, Teikyo University
Sagamiko, Sagamihara, Kanagawa 199-0195, Japan
saccharide chains but also the synthesis of complex oligo-
saccharides are thought to be possible (Scheme 1).
Fax: +81-42-685-3729
E-mail: somc@pharm.teikyo-u.ac.jp
1064
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2007, 1064–1068

Carbamate-Linked Oligosaccharides
SHORT COMMUNICATION
Results and Discussion
and 2.[3] α-Carboxylic acid 1 was found to be less reactive
than β-carboxylic acid 2, and primary alcohol 7 had the
best reactivity among the acceptors 3 to 9. Although hemi-
acetal 3 was an anomeric mixture, the stereochemistry at the
1-position of 10 and 17 tended predominantly towards a β-
linkage.[9] The reason for this is that there is an equilibrium
between the anomeric isomers of 3, and the less-hindered
β-isomer reacts much faster than the more-hindered α-
alcohol. According to the overall results, the steric hin-
drance of the substrates reduces both the reactivity and the
yield. The rate-determining step could be the addition step
of a sugar alcohol to an isocyanate converted in situ from
a glycosyl carboxylic acid.
At first, we investigated the synthesis of a series of carba-
mate-linked disaccharides. As substrates, we utilized the ste-
reoisomers of carboxylic acids 1 and 2, which were derived
from the corresponding alcohols[6] by a simple oxidation
reaction,[7] as glycosyl donors and the hemiacetal 3 and
alcohols 4 to 9,[8] derived from -glucose, galactose, and
mannose, as glycosyl acceptors. A carboxylic acid and two
equivalents of an alcohol were heated at reflux with two
equivalents of DPPA and a base in benzene to afford the
desired carbamate-linked disaccharides in moderate to
good yields (Table 1). As reported previously K2CO3 was
the most effective base[3] (however, in entry 12 it can be seen
that the use of triethylamine also gave good results), and a
We then tried the elongation of a saccharide by using
catalytic amount of Ag2CO3 often improved the yield (en- carboxylic acid 24[3] and alcohol 25[10] (Scheme 2). Com-
tries 4, 5, 6, and 13). All the products were obtained with pounds 24 and 25 (the same number of equivalents) were
retention of the stereochemistry at the substrate glycosyl treated with DPPA and triethylamine for 4 h, and disaccha-
carboxylic acid. Thus, this procedure should stereospecifi- ride 26 was obtained in 84% yield. Carboxylic acid 27,
cally provide both anomeric isomers of disaccharide with 1 which was transformed from 26 by hydrolysis, was similarly
Table 1. Synthesis of carbamate-linked disaccharides.
[a] Ag2CO3 (0.1 equiv.) was added. [b] Triethylamine was used as a base.
Eur. J. Org. Chem. 2007, 1064–1068
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
1065

D. Sawada, S. Sasayama, H. Takahashi, S. Ikegami
SHORT COMMUNICATION
Scheme 2. Elongation of a saccharide chain both at the carboxylic acid end and alcohol end.
Scheme 3. Synthesis of dendritic molecules.
treated with 2 equiv. alcohol 25 to give trisaccharide 29 in to afford 29 in 88% yield (route b). The deprotection of
85% yield (route a). On the other hand, alcohol 28, con- compound 29 was successfully achieved by Pd·C, H2 in
verted from 26 with 2,3-dichloro-5,6-dicyano-1,4-benzoqui- MeOH, and CH2Cl2 to give 30 in 90% yield. As above, it
none (DDQ), was treated with 2 equiv. carboxylic acid 24 has been indicated that this synthetic method makes it pos-
1066
www.eurjoc.org
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2007, 1064–1068

Carbamate-Linked Oligosaccharides
SHORT COMMUNICATION
4.67 (d, J = 10.98 Hz, 1 H), 4.58–4.56 (m, 2 H), 4.55–4.28 (m, 6
H), 3.90 (t, J = 9.52 Hz, 1 H), 3.81 (m, 1 H), 3.70 (dd, J = 5.13,
9.27 Hz, 1 H), 3.63–3.53 (m, 7 H), 3.45 (dd, J = 5.86, 10.75 Hz, 1
H), 3.36 (s, 3 H) ppm. 13C NMR (100 MHz, CDCl3): δ = 154.65,
138.50, 138.20, 137.96, 137.89, 137.76, 137.74, 136.87, 128.38,
128.35, 128.34, 128.32, 128.28, 128.25, 128.19, 128.11, 128.07,
127.96, 127.91, 127.87, 127.83, 127.78, 127.73, 127.68, 127.64,
127.58, 127.48, 127.41, 127.36, 127.33, 98.07, 81.81, 79.19, 77.21,
77.04, 76.88, 75.52, 75.25, 74.94, 73.60, 73.53, 73.49, 73.42, 72.11,
70.71, 69.30, 68.89, 55.35 ppm. MS (FAB – NBA + NaI): m/z =
sible to elongate the saccharide chain both at the carboxylic
acid end and the alcohol end by simple transformations. In
addition, this method has a powerful application in carba-
mate-linked oligosaccharide synthesis.
On the basis of the above experiments, we set out to syn-
thesize dendritic compound 34 having a sugar as the core
structure, and started from the known monosaccharide
31,[11] which was soluble in benzene (Scheme 3). At first,
compound 31 was heated at reflux with 4 equiv. carboxylic
acid 2 and 8 equiv. DPPA and K2CO3, and after 17 h, grati- 1052 [M + Na]+. HRMS (FAB – NBA + NaI): calcd. for
fyingly, we obtained trisaccharide 32 in 84% yield in one
step with the carbamate-linked saccharide chains at the 2-
and 3-positions of the core glucose. The anomeric isomers
C63H67NNaO12 1052.4561; found 1052.4536.
Spectral Data for 29: [a]2D6 = –2.23° (c = 0.7, CHCl ). IR (neat): ν
˜
3
= 1738 (C=O), 3328 (NH) cm–1. 1H NMR (400 MHz, CDCl3): δ
of 32 were then separated, and the major α isomer was used = 7.21–7.32 (m, 45 H), 7.09–7.11 (m, 2 H), 6.80–6.82 (m, 2 H),
5.10–5.17 (m, 2 H), 4.68–4.92 (m, 17 H), 4.55 (d, J = 6.59 Hz, 1
H), 4.53 (d, J = 6.59 Hz, 1 H), 4.52–4.60 (m, 2 H), 4.47 (d, J =
10.26 Hz, 1 H), 4.32–4.38 (m, 4 H), 4.24–4.27 (dd, J = 3.72,
11.58 Hz, 2 H), 3.87 (d, J = 9.53 Hz, 1 H), 3.61–3.68 (m, 6 H), 3.74
(s, 3 H), 3.68 (s, 3 H), 3.57 (m, 1 H), 3.42–3.52 (m, 4 H), 3.28–3.34
(m, 2 H) ppm. 13C NMR (100 MHz, CDCl3): δ = 169.24, 159.19,
155.35, 155.21, 138.41, 138.24, 138.19, 137.76, 137.66, 137.51,
137.35, 129.92, 129.66, 128.57, 128.51, 128.46, 128.44, 128.42,
128.40, 128.39, 128.31, 128.20, 128.17, 128.13, 128.02, 127.97,
127.91, 127.88, 127.79, 127.75, 127.67, 127.62, 113.72, 86.14, 86.11,
in further transformations in order to attain a clear analysis
of the reactions. After the removal of the methoxybenzyl-
idene acetal group of 32 (α isomer), the given 4,6-diol was
similarly treated with 2, DPPA, and K2CO3 to give penta-
saccharide 33 in 89% yield. The allyl group of compound
33 was removed by PdCl2, and the resultant acetal should
then be the substrate for the introduction of the last carba-
mate-linked saccharide. Unfortunately, it was found that the
substrate acetal was not stable and decomposed under the
reaction conditions when K2CO3 was used. After several 85.83, 85.68, 81.79, 81.58, 80.33, 80.16, 79.90, 79.79, 78.13, 78.02,
77.82, 77.57, 77.50, 77.40, 76.20, 75.73, 75.62, 75.24, 75.16, 75.09,
75.06, 74.86, 74.80, 73.07, 67.69, 64.02, 63.72, 61.81, 60.37, 55.15,
52.44, 52.41 ppm. MS (FAB – NBA + NaI): m/z = 1586 [M +
Na]+. HRMS (FAB – NBA + NaI): calcd. for C93H98N2NaO20
1585.6613; found: 1585.6617.
attempts with triethylamine, Ag2CO3, and 6 equiv. 2, we fi-
nally succeeded in obtaining hexasaccharide 34 in 77%
yield.[12] Dendritic hexasaccharide 34 was afforded in only
5 steps in 34% overall yield by using α-32 obtained from
the simple monosaccharide 31.
Spectral Data for 30: [a]2D3 = –52.6° (c = 0.92, H O). IR (neat): ν =
˜
2
1717 (C=O), 3326 cm–1. 1H NMR (400 MHz, D2O): δ = 4.80–4.83
(m, 3 H), 4.45–4.50 (m, 2 H), 4.24–4.28 (m, 2 H), 4.02–4.05 (m, 1
H), 3.83–3.91 (m, 2 H), 3.67–3.74 (m, 4 H), 3.56 (s, 3 H), 3.35–3.58
(m, 7 H) ppm. 13C NMR (100 MHz, D2O): δ = 171.05, 170.45,
165.54, 152.08, 76.17, 74.07, 72.42, 71.88, 71.62, 71.41, 71.27,
70.85, 70.51, 62.23, 66.14, 65.92, 65.65, 65.55, 63.73, 63.52,
54.82 ppm. MS (FAB – NBA + NaI): m/z = 655 (M+ + Na)+.
HRMS (FAB – NBA + NaI): calcd. for C22H36N2NaO19 655.1811;
found: 655.1807.
Conclusions
We have accomplished the stereospecific synthesis of car-
bamate-linked oligosaccharides by using a modified Curtius
rearrangement. This synthetic method has made it possible
to synthesize unique oligosaccharides, which are chemically
stable, and would be applicable in the synthesis of func-
tional molecules. The synthesis of various complex mole-
cules is now in progress in our laboratory.
Spectral Data for 34: [a]2D4 = +4.19° (c = 0.42, CHCl3). IR (neat):
ν = 1746 (C=O), 3317 cm–1. 1H NMR (600 MHz, CDCl3): δ =
˜
7.41–7.04 (m, 100 H), 5.88 (m, 1 H), 5.71 (m, 1 H), 5.54 (m, 1 H),
5.39 (m, 1 H), 5.31–5.29 (m, 1 H), 5.19 (m, 1 H), 5.11–5.02 (m, 3
H), 4.90–4.39 (m, 47 H), 4.16 (m, 1 H), 3.93 (m, 1 H), 3.79 (m, 1
H), 3.47–3.68 (m, 23 H), 5.19 (m, 1 H), 3.33 (m, 1 H), 3.26–3.23
(m, 1 H), 3.10–3.19 (m, 3 H) ppm. 13C NMR (150 MHz, CDCl3):
δ = 155.04, 154.39, 153.28, 138.56, 138.44, 138.33, 138.22, 138.12,
138.08, 138.02, 137.95, 137.77, 137.43, 129.07, 128.97, 128.72,
128.62, 128.59, 128.49, 128.35, 128.29, 128.24, 128.15, 128.03,
127.85, 127.79, 127.71, 127.63, 127.53, 127.49, 92.90, 86.00, 85.90,
85.71, 82.15, 81.69, 80.56, 80.33, 77.89, 77.47, 76.41, 75.80, 75.67,
75.47, 74.96, 74.86, 74.81, 74.76, 74.48, 74.18, 73.46, 73.24, 68.51,
68.15 ppm. C181H187N5O36 (3008.48): calcd. C 72.26, H 6.27, N
2.33; found C 72.12, H 6.46, N 2.20.
Experimental Section
Typical procedure for the synthesis of a carbamate-linked saccha-
ride by the modified Curtius rearrangement (Table 1, entry 4): Di-
phenyl phosphoryl azide (29 uL, 0.132 mmol) was added to the
solution of 1 (38 mg, 0.066 mmol), 6 (63 mg, 0.132 mmol), K2CO3
(18 mg, 0.132 mmol), and Ag2CO3 (1.8 mg, 0.0066 mmol) in ben-
zene (7 mL), and the whole mixture was heated at reflux for 29 h.
Saturated aqueous NH4Cl solution (30 mL) was then added at
0 °C, and the mixture was extracted with ethyl acetate (3ϫ30 mL).
The combined organic layer was washed with brine (30 mL), dried
with Na2SO4, then filtered and evaporated to give the crude prod-
uct, which was purified by silica gel chromatography (hexane/
EtOAc, 4:1) to afford 13 (53 mg, 78%): [a]2D4 = +8.17° (c = 0.55,
Acknowledgments
CHCl ). IR (neat): ν = 1721 (C=O), 3285 (NH) cm–1. 1H NMR
˜
3
(400 MHz, CDCl3): δ = 7.27–7.01 (m, 35 H), 5.53 (d, J = 5.13 Hz,
1 H), 5.48 (m, 1 H); 4.96 (t, J = 9.52 Hz, 1 H), 4.85 (d, J =
10.98 Hz, 1 H), 4.79 (d, J = 11.48 Hz, 1 H), 4.77–4.68 (m, 4 H),
This work was partially supported by the Ministry of Education,
Culture, Sports, Science, and Technology of Japan, and the Uehara
Memorial Foundation (D. S.).
Eur. J. Org. Chem. 2007, 1064–1068
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
1067

D. Sawada, S. Sasayama, H. Takahashi, S. Ikegami
SHORT COMMUNICATION
[8]
a) N. Pravdic´, B. Danilov, Carbohyd. Res. 1974, 36, 167–180;
b) T. Lecourt, A. Herault, A. J. Pearce, M. Sollogoub, P. Sinaÿ,
Chem. Eur. J. 2004, 10, 2960–2971; c) J. R. Falck, D. K. Barma,
S. K. Venkataraman, R. Baati, C. Mioskowski, Tetrahedron
Lett. 2002, 43, 963–966; d) R. Eby, C. Schuerch, Carbohyd.
Res. 1982, 100, C41–C43; e) N. E. Franks, R. Montdomery,
Carbohyd. Res. 1968, 6, 286–298.
[1] A. R. Dwek, Chem. Rev. 1996, 96, 683–720.
[2] a) L. Laupichler, J. Thiem, Synlett 1992, 159; b) J. M.
García Fernández, C. O. Mellet, V. M. Díaz Pérez, J. Fuentes,
J. Kovács, I. Pintér, Tetrahedron Lett. 1997, 38, 4161–4164; c)
D. Prosperi, S. Ronchi, L. Panza, A. Rencurosi, G. Russo, Syn-
lett 2004, 9, 1529–1532; d) Y. Ichikawa, Y. Matsukawa, T. Ni-
shiyama, M. Isobe, Eur. J. Org. Chem. 2004, 586–591.
[3] D. Sawada, S. Sasayama, H. Takahashi, S. Ikegami, Tetrahe-
dron Lett. 2006, 47, 7219–7223.
[4] a) Y. Ichikawa, T. Nishiyama, M. Isobe, Synlett 2000, 9, 1253–
1256; b) Y. Ichikawa, T. Nishiyama, M. Isobe, J. Org. Chem.
2001, 66, 4200–4205; c) T. Nishiyama, Y. Ichikawa, M. Isobe,
Synlett 2003, 1, 47–50; d) D. Prosperi, S. Ronchi, L. Lay, A.
Rencurosi, G. Russo, Eur. J. Org. Chem. 2004, 395–405; e) Y.
Ichikawa, Y. Matsukawa, M. Isobe, J. Am. Chem. Soc. 2006,
128, 3934–3938.
[9]
By using 3, which is a mixture of the stereoisomers at the 1-
position in a 3:1 ratio (α/β), 10 was exclusively obtained with
a 1-β-linkage and 17 was obtained as a mixture of stereoiso-
mers at the 1-position in a 1:10 ratio (α/β).
Compound 25 was prepared from 24 in two steps – conversion
to the methyl ester with MeI and NaHCO3 (100% yield) and
then deprotection of the MPM group with DDQ (96% yield).
[10]
[11] T. Oka, K. Fujiwara, A. Murai, Tetrahedron 1998, 54, 21–44.
It is difficult to separate the anomeric isomers of compound
31.
[12] On the basis of the results in Table 1, the stereochemistry at
the 1-position of the core sugar in 34 should be a β-linkage,
and a trace amount of the stereoisomer could be isolated in
this reaction, whose structure was not determined.
Received: November 18, 2006
[5] T. Shioiri, K. Ninomiya, S. Yamada, J. Am. Chem. Soc. 1972,
94, 6203–6204.
[6] T. V. RajanBabu, G. S. Reddy, J. Org. Chem. 1986, 51, 5458–
5461.
[7] M. Zhao, J. Li, E. Mano, Z. Song, M. D. Tschaen, J. J. E. Gra-
bowski, J. P. Reider, J. Org. Chem. 1999, 64, 2564–2566.
Published Online: January 17, 2007
1068
www.eurjoc.org
© 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2007, 1064–1068