Chemoselective glycosidation strategy based on glycosyl donors and acceptors carrying phosphorus-containing leaving groups: A convergent synthesis of ganglioside GM3

Article abstract of DOI:10.1016/S0040-4039(00)01343-5

A convergent synthesis of ganglioside GM₃ has been achieved by capitalizing on three different phosphorus-containing leaving groups, in which the key features involve a chemo- and regioselective α-glycosidation of sialyl phosphite with partially benzylated galactosyl tetramethylphosphorodiamidate by the aid of trifluoromethanesulfonic acid and a traditionally uncommon coupling of an α-sialyl-(2→3)- galactosyl donor with a glucosylceramide building block which has the β-O-linked ceramide prebuilt into the glucose by a diphenyl phosphate method. (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0040-4039(00)01343-5

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 41 (2000) 7691–7695
Chemoselective glycosidation strategy based on glycosyl
donors and acceptors carrying phosphorus-containing
leaving groups: a convergent synthesis of ganglioside GM3
Hiroki Sakamoto, Seiichi Nakamura, Toshifumi Tsuda and Shunichi Hashimoto*
Graduate School of Pharmaceutical Sciences, Hokkaido University, Sapporo 060-0812, Japan
Received 13 July 2000; revised 3 August 2000; accepted 4 August 2000
Abstract
A convergent synthesis of ganglioside GM3 has been achieved by capitalizing on three different
phosphorus-containing leaving groups, in which the key features involve a chemo- and regioselective
a-glycosidation of sialyl phosphite with partially benzylated galactosyl tetramethylphosphorodiamidate by
the aid of trifluoromethanesulfonic acid and a traditionally uncommon coupling of an a-sialyl-(2“3)-
galactosyl donor with a glucosylceramide building block which has the b-O-linked ceramide prebuilt into
the glucose by a diphenyl phosphate method. © 2000 Elsevier Science Ltd. All rights reserved.
Keywords: glycosidation; glycolipids; phosphorous acids and derivatives; phosphoramidates.
With the advent of high-yielding and stereocontrolled glycosidation reactions, an enormous
amount of effort is currently being devoted to the device and implementation of innovative
strategies for the synthesis of various glycoconjugates as well as oligosaccharides.1,2 Our efforts
in this area have led to the development of a chemo- and stereoselective glycosidation method
based on glycosyl donors and acceptors carrying phosphorus-containing leaving groups, in
which a tetramethylphosphorodiamidate group plays a pivotal role as an anomeric protecting
group as well as a leaving group.3 The efficiency of this method was demonstrated by a
convergent and stereocontrolled synthesis of glycosphingolipid Gb3.4 To test the feasibility of
our glycosidation strategy for the synthesis of sialic acid-containing glycosphingolipids, we now
addressed a synthesis of ganglioside GM3 (1).5
Apart from its biological significance,6 GM3 (1) (a-NeuAc-(2“3)-b-
D
-Gal-(1“4)-b- -Glc-
D
ceramide) has presented itself as a target for evaluation of new glycosidation methodologies as
well as chemical and enzymatic sialylation reactions. Chemical7 and chemoenzymatic8 syntheses
of GM3 have already been achieved by five and four groups, respectively, the majority of which
commences with a-sialylation of a lactosyl acceptor followed by coupling with azidosphigosine
or ceramide derivatives at a late stage.
* Corresponding author. Fax: +81-11-706-4981; e-mail: hsmt@pharm.hokudai.ac.jp
0040-4039/00/$ - see front matter © 2000 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(00)01343-5

7692
Our convergent strategy to 1 relies on a-selective sialylation of a galactosyl acceptor by means
of the chemoselective glycosidation methodology followed by stereocontrolled coupling with a
glucosylceramide building block which have a b-O-linked ceramide prebuilt into the glucose. In
this context, the Schmidt and Wong groups have independently reported the effective use of sialyl
phosphites for high-yielding and a-selective sialylation,7d,9,10 in which the combinational use of
TMSOTf as a promoter and MeCN as a solvent is also highlighted. Coupled with our continuing
interest in the phosphite method,11 we explored TMSOTf-promoted coupling of sialyl diethyl
phosphite 2 with the partially benzoylated galactosyl tetramethylphosphorodiamidate 312,14
selected as an ideal ‘disarmed’ acceptor for the subsequent coupling with neighboring group
participation. However, the desired reaction did not work well (Table 1, entry 1). Instead,
decomposition of 2 occurred to give 2,3-dehydro compound 7 as a major product, probably due
to the poor nucleophilicity of 3.15
Table 1
Sialylation of galactosyl tetramethylphosphorodiamidates
In an attempt to enhance the reactivity of the acceptor alcohol, benzoyl protection on 3 was
switched to benzyl ether protection. As might be expected, however, the protective group
interchange presented another critical problem; the 2,6-di-O-benzyl-protected galactosyl phospho-
rodiamidate 416 as well as sialyl phosphite 2 could be activated by the use of TMSOTf in EtCN
at −78°C. Thus, we surveyed other promoters which could selectively activate the phosphite group
in 2 while leaving the phosphorodiamidate in 4 intact. Based on our findings that glycosidation
of the fully benzylated glycosyl diethyl phosphites could be effected with the aid of BF3·OEt2 at
−78°C whereas the corresponding tetramethylphosphorodiamidates were unreactive at below
−10°C,3 we first examined coupling of 2 with 4 under the influence of BF3·OEt2. Indeed,
BF3·OEt2-promoted coupling at −45°C displayed complete chemo- and regioselectivity to give
disaccharide 6 in 77% yield, but did not exhibit a reasonable level of stereoselectivity (entry 2),
although the use of EtCN as a solvent was expected to favor a-selective sialylation.18
Our attention was next turned to the use of TfOH, which had been once employed by Wong
and co-workers for glycosidation of fully benzylated glycosyl phosphites.19 Prior to the real
coupling, we found that sialylation of methyl 2,3,4-tri-O-benzyl-a-
D
-glucopyranoside with 2 in
EtCN could be promoted with the aid of TfOH at −78°C to give the corresponding sialosides

7693
in 84% yield with the a:b ratio of 86:14. To the best of our knowledge, this is the first example
that demonstrates the effectiveness of TfOH as a promoter for sialylation with sialyl phosphites.
In a separate experiment, we also confirmed that phosphorodiamidate 4 remained intact under
such conditions when kept at below −25°C. In the event, coupling of 2 with 4 in EtCN in the
presence of TfOH proceeded chemo- and regioselectively at −78°C to give disaccharide 6 in 86%
yield with the a:b ratio of 84:16 (entry 3),20 from which the desired a-anomer 6a could be easily
separated by column chromatography. It is worthy of note that the beneficial effect of EtCN on
a-selectivity was markedly pronounced with TfOH compared with BF3·OEt2 (entry 2 vs 3).21
With the chemoselective a-sialylation secured, the stage was now set for the installation of
glucosylceramide building block 134 (Scheme 1). To this end, the benzyl groups in 6a were
deblocked by catalytic hydrogenolysis. However, subsequent benzoylation was found to stop at
the monobenzoylation stage.22 It was thought that the sialyl moiety was too demanding as a
steric presence to allow benzoylation of the hydroxyl group at C2. Thus, we revised our scenario
so as to diminish the supposed steric hindrance by forming a rigid lactone system. In the event,
lactonization of 6a with DBU and debenzylation was followed by smooth benzoylation to
provide disaccharide donor 9 in 68% yield.23 Compound 13 was prepared by a slight modifica-
tion of the previously reported protocol,4 in which the chemical yield was improved by
employing diphenyl phosphate 10 as a glycosyl donor instead of the corresponding tetra-
methylphosphorodiamidate (94% vs 74%). Coupling of 9 and glucosylceramide 13 under
Scheme 1. Reagents and conditions: (a) DBU, CH2Cl2, 0°C, 1 h; (b) H2, Pd(OH)2/C, THF, 2 h; (c) Bz2O, pyridine,
DMAP; (d) 10:11:TMSOTf molar ratio=1.1:1.0:1.5, CH2Cl2, 0°C, 1 h; (e) H2NC(S)NH2, 2,6-lutidine, EtOH, 70°C,
2 h; (f) 9:13:TMSOTf molar ratio=1.1:1.0:3.0, CH2Cl2, 0°C, 1 h; (g) NaOMe, MeOH, then H2O; (h) Na, liq. NH3,
THF, −78 to −33°C, 15 min, then MeOH; (i) Ac2O, pyridine; (j) NaOMe, MeOH, then H2O

7694
promotion of TMSOTf in CH2Cl2 at 0°C proceeded uneventfully to give the protected GM3 14
in 80% yield and with complete stereocontrol.24 Deprotection of the acyl protecting groups and
lactone ring-opening were effected in one-pot by treatment of 14 with NaOMe in MeOH
followed by the action of water. Debenzylation of the product in the crude form with sodium
in liquid ammonia, followed by peracetylation (for the purpose of purification),25 per-deacetyla-
tion and saponification, furnished the target GM3 (1), [h]2D3 +1.92 (c 0.2, 1:1 CHCl3–MeOH)
1
[lit.,7b [h]D +1.8 (c 0.2, 1:1 CHCl3–MeOH)] in 67% yield from 14, which exhibited identical H
NMR data with those reported for this natural product.26
In summary, we have accomplished a convergent synthesis of GM3 based on the chemoselec-
tive glycosidation methodology, in which the anomeric reactivity of the glycosyl donor and
acceptor carrying different phosphorus-containing leaving groups can be regulated by a judi-
cious choice of the protecting groups as well as reaction conditions. We have also demonstrated
the effective use of TfOH as a promoter for sialylation with sialyl phosphites. Further extension
of the present method to the synthesis of more complex gangliosides is in progress.
Acknowledgements
This research was partially supported by Takeda Science Foundation and a Grant-in-Aid for
Scientific Research from the Ministry of Education, Science, Sports and Culture, Japan. H.S. is
grateful to the JSPS for a graduate fellowship.
References
1. For a review, see: Boons, G.-J. Tetrahedron 1996, 52, 1095.
2. For more recent works, see: (a) Zhu, T.; Boons, G.-J. Tetrahedron Lett. 1998, 39, 2187. (b) Douglas, N. L.; Ley,
S. V.; Lu¨cking, U.; Warriner, S. L. J. Chem. Soc., Perkin Trans. 1 1998, 51. (c) Yamada, H.; Kato, T.; Takahashi,
T. Tetrahedron Lett. 1999, 40, 4581. (d) Zhang, Z.; Ollmann, I. R.; Ye, X.-S.; Wischnat, R.; Baasov, T.; Wong,
C.-H. J. Am. Chem. Soc. 1999, 121, 734.
3. Hashimoto, S.; Sakamoto, H.; Honda, T.; Ikegami, S. Tetrahedron Lett. 1997, 38, 5181.
4. Hashimoto, S.; Sakamoto, H.; Honda, T.; Abe, H.; Nakamura, S.; Ikegami, S. Tetrahedron Lett. 1997, 38, 8969.
5. Yamakawa, T.; Suzuki, S. J. Biochem. (Tokyo) 1951, 38, 199.
6. (a) Bremer, E. G.; Schlessinger, J.; Hakomori, S. J. Biol. Chem. 1986, 261, 2434. (b) Hanai, N.; Nores, G.;
Torres-Me´ndez, C.-R.; Hakomori, S. Biochem. Biophys. Res. Commun. 1987, 147, 127.
7. (a) Sugimoto, M.; Ogawa, T. Glycoconjugate J. 1985, 2, 5. (b) Murase, T.; Ishida, H.; Kiso, M.; Hasegawa, A.
Carbohydr. Res. 1989, 188, 71. (c) Numata, M.; Sugimoto, M.; Ito, Y.; Ogawa, T. Carbohydr. Res. 1990, 203, 205.
(d) Martin, T. J.; Brescello, R.; Toepfer, A.; Schmidt, R. R. Glycoconjugate J. 1993, 10, 16. (e) Tomoo, T.;
Kondo, T.; Abe, H.; Tsukamoto, S.; Isobe, M.; Goto, T. Carbohydr. Res. 1996, 284, 207. (f) Tietze, L. F.;
Gretzke, D. Eur. J. Org. Chem. 1998, 1895.
8. (a) Ito, Y.; Paulson, J. C. J. Am. Chem. Soc. 1993, 115, 1603. (b) Liu, K. K.-C.; Danishefsky, S. J. J. Am. Chem.
Soc. 1993, 115, 4933. (c) Yamada, K.; Fujita, E.; Nishimura, S.-I. Carbohydr. Res. 1998, 305, 443. (d) Zehavi, U.;
Tuchinsky, A. Glycoconjugate J. 1998, 15, 657.
9. Martin, T. J.; Schmidt, R. R. Tetrahedron Lett. 1992, 33, 6123.
10. Kondo, H.; Ichikawa, Y.; Wong, C.-H. J. Am. Chem. Soc. 1992, 114, 8748.
11. Tanaka, H.; Sakamoto, H.; Sano, A.; Nakamura, S.; Nakajima, M.; Hashimoto, S. Chem. Commun. 1999, 1259
and references cited therein.
12. Compound 3 was prepared from allyl 2,6-di-O-benzoyl-a-
PhCH(OMe)2, CSA, CHCl3, 5 h, 65%; (2) Pd(PPh3)4, DABCO, EtOH–H2O (9:1), reflux, 6 h; (3) I2, pyridine,
D
-galactopyranoside13 by the following sequence: (1)

7695
THF–H2O (4:1), 30 min, 80% (two steps); (4) n-BuLi, THF, −78°C, then ClPO(NMe2)2, HMPA, −30°C, 2 h,
73%; (5) H2, Pd(OH)2/C, EtOH, 24 h, 85%.
13. Ziegler, T.; Herold, G. Liebigs Ann. Chem. 1994, 859.
14. All new compounds exhibited satisfactory spectral (500 MHz 1H NMR, 125 MHz 13C NMR, 109 MHz 31P NMR
1
and IR) and high-resolution mass spectral characteristics. Selected spectroscopic data: 6a: H NMR (CDCl3) l
2.15 (dd, J=13.2, 13.6 Hz, H-3%), 2.54 (dd, J=4.5, 13.2 Hz, H-3%), 3.81 (s, CO2CH3), 3.99 (dd, J=5.0, 12.5 Hz,
H-9%), 4.28 (d, J=12.5 Hz, H-9%), 5.92 (dd, J=3.5, 8.0 Hz, H-1); 13C NMR (CDCl3) l 93.7 (d, JCꢀP=5.0 Hz, C-1),
1
98.7 (C-2%). 6b: H NMR (CDCl3) l 2.02 (m, H-3%), 2.60 (dd, J=4.6, 13.2 Hz, H-3%), 3.59 (s, CO2CH3), 4.00–4.06
(m, H-9%), 4.75 (dd, J=1.9, 12.5 Hz, H-9%), 5.99 (dd, J=3.3, 8.2 Hz, H-1); 13C NMR (CDCl3) l 92.8 (C-1), 99.2
(C-2%). 14: 1H NMR (CDCl3) l 4.71 (d, J=8.1 Hz, H-1%); 13C NMR (CDCl3) l 99.6 (C-1%). 1: 1H NMR
(DMSO-d6–D2O=49:1) l 4.13 (d, J=7.7 Hz, H-1), 4.17 (d, J=7.8 Hz, H-1%); 13C NMR (DMSO-d6–D2O=49:1)
l 99.5, 103.5, 104.1; FAB-HRMS calcd for C59H108N2NaO21 (M+Na+) 1203.7342, found 1203.7270.
15. Birberg, W.; Lo¨nn, H. Tetrahedron Lett. 1991, 32, 7457.
16. Compound 4 was prepared from allyl 3,4-di-O-benzoyl-2,6-di-O-benzyl-a-D
-galactopyranoside17 by the following
sequence: (1) PdCl2, NaOAc, AcOH–H2O (9:1), 80°C, 2 h, 66%; (b) n-BuLi, THF, −78°C, 30 min, then
ClPO(NMe2)2, HMPA, −30°C, 2 h, 70%; (3) BF3·OEt2, MS4A, CH2Cl2, −23°C, 6 h, 80% (isomerization); (4)
K2CO3, MeOH, 6 h, 89%.
17. Nashed, M. A.; Anderson, L. Carbohydr. Res. 1976, 51, 65.
18. Murase, T.; Ishida, H.; Kiso, M.; Hasegawa, A. Carbohydr. Res. 1988, 184, C1. (b) Hasegawa, A.; Kiso, M. In
Preparative Carbohydrate Chemistry; Hanessian, S., Ed.; Marcel Dekker: New York, 1996; pp. 357–379.
19. Kondo, H.; Aoki, S.; Ichikawa, Y.; Halcomb, R. L.; Ritzen, H.; Wong, C.-H. J. Org. Chem. 1994, 59, 864.
20. Under these conditions, 6-O-benzoyl-a-D-galactopyranosyl N,N,N%,N%-tetramethylphosphorodiamidate also
remained intact, but the reaction with 2 gave a complex mixture of products.
21. These results together with sialylations in MeCN reported to date7d,18 suggest that triflate ion is the superior
choice from counterions associated with sialyl-nitrilium intermediates for allowing high order of a-selectivities
while its role is presently not clear.
22. Hasegawa and co-workers reported the dibenzoylation of C2 and C4 hydroxyl groups in a sialyl-galactoside
derivative, see: Kameyama, A.; Ishida, H.; Kiso, M.; Hasegawa, A. Carbohydr. Res. 1990, 200, 269.
23. For the use of (1%“4)-lactonized sialyl-a(2“3)-galactosyl donor, see: Gervay, J.; Peterson, J. M.; Oriyama, T.;
Danishefsky, S. J. J. Org. Chem. 1993, 58, 5465.
24. The coupling of 6a with 12 met with little success due mainly to the poor solubility of 12 in EtCN below −50°C,
the limit temperature allowing for high b-selectivity.
25. At this stage lactone formation occurred, but it was not determined which of (1%“2)- and (1%“4)-lactones was
produced.
26. Koerner Jr., T. A. W.; Prestegard, J. H.; Demou, P. C.; Yu, R. K. Biochemistry 1983, 22, 2676.
.