Conversion of N-acetylneuraminic acid glycosyl chloride into dibenzyl glycosyl phosphate: O-glycosylation in the absence of a promoter

Article abstract of DOI:10.1023/B:RUCB.0000035663.15439.84

Readily accessible N-acetylneuraminic acid (Neu5Ac) glycosyl chloride, which was regarded to be a poor glycosyl donor, was shown to react with dibenzyl phosphoric acid salts in the absence of glycosylation promoters to give the corresponding β-Neu5Ac dibenzyl glycosyl phosphate in high yield.

Full text of DOI:10.1023/B:RUCB.0000035663.15439.84

Russian Chemical Bulletin, International Edition, Vol. 53, No. 3, pp. 717—719, March, 2004
717
Conversion of Nꢀacetylneuraminic acid glycosyl chloride into
dibenzyl glycosyl phosphate: Oꢀglycosylation in the absence of a promoter
A. M. Shpirt, L. O. Kononov,ꢀ V. I. Torgov, and V. N. Shibaev
N. D. Zelinsky Institute of Organic Chemistry, Russian Academy of Sciences,
4
7 Leninsky prosp., 119991 Moscow, Russian Federation.
Fax: +7 (095) 135 5328. Eꢀmail: kononov@ioc.ac.ru
Readily accessible Nꢀacetylneuraminic acid (Neu5Ac) glycosyl chloride, which was reꢀ
garded to be a poor glycosyl donor, was shown to react with dibenzyl phosphoric acid salts in
the absence of glycosylation promoters to give the corresponding βꢀNeu5Ac dibenzyl glycosyl
phosphate in high yield.
Key words: sialic acids, Nꢀacetylneuraminic acid, glycosyl chloride, glycosyl phosphates,
monosaccharides, NMR spectroscopy.
The development of new efficient procedures for the
synthesis of oligosaccharides containing sialic acid residues
and, in particular, Nꢀacetylneuraminic acid (Neu5Ac)
residue, which are responsible for various immunological,
neurobiological, oncological, and other biological proꢀ
the present study, we examined the reactions of this comꢀ
pound with dibenzyl phosphate (2a) and its salts 2b and 2c
(Scheme 1). The expected reaction product, viz., dibenzyl
glycosyl phosphate 3, can subsequently be transformed
into cytidine 5´ꢀmonophosphate Nꢀacetylneuraminic
1
4,5
cesses, is an important problem of modern synthetic carꢀ
bohydrate chemistry.
acid (CMPꢀNeu5Ac), which serves as a glycosyl donor
2
in enzymatic sialylation.
The preparation of various glycosyl donors proposed2
for this purpose is often very laborious. Presently, readily
accessible glycosyl chloride2,3 1 (Scheme 1) is little used
because of its low reactivity. However, it has recently
been found that this compound can react with MeOH in
the absence of an added promoter to give the correspondꢀ
ing acetylated glycoside in high yield.
It is of interest to search for other Oꢀglycosylation
reactions in which chloride 1 can efficiently be used. In
The synthesis of Neu5Ac βꢀglycosyl phosphate 3 has
4
,5
been described in connection with the preparation of
CMPꢀNeu5Ac and investigation of its glycosylating poꢀ
4
tential. Compound 3 was prepared by phosphitylation of
i
protected Neu5Ac hemiacetal (5a) with (BnO) PNPr
2
2
4
followed by oxidation of triester 6a as well as by
glycosylation of dibenzyl phosphoric acid (2a) with
Neu5Ac glycosyl phosphite 6b. An attempt to glycosylate
3
a
5
dibenzyl phosphoric acid (2a) with glycosyl chloride 1 in
Scheme 1
Published in Russian in Izvestiya Akademii Nauk. Seriya Khimicheskaya, No. 3, pp. 684—686, March, 2004.
066ꢀ5285/04/5303ꢀ0717 © 2004 Plenum Publishing Corporation
1

7
18
Russ.Chem.Bull., Int.Ed., Vol. 53, No. 3, March, 2004
Shpirt et al.
the presence of AgOTf as a promoter was unsuccessful. In
this case, only the elimination product, glycal 4, and
hydrolysis product 5a were isolated from the reaction
uct 3 was confirmed by the 31P NMR spectrum, which has
31
one signal (δ –6.0). Generally, signals in the P NMR
P
spectra of dibenzyl hexosyl phosphates are observed at δ
from –2.0 to –3.0. The unusually highꢀfield position of
P
4
7
mixture.
Our experiments demonstrated that acid 2a did not
react (TLC data) with chloride 1 in the absence of a
promoter under the conditions used (MeCN, 20 °C).
However, chloride 1 was smoothly transformed into
βꢀglycosyl dibenzyl phosphate 3 in the reaction with more
nucleophilic ammonium dibenzyl phosphates (2b,c). In
all cases, the reactions afforded glycal 4 as the only byꢀ
product. Separation of this product from phosphate 3 is a
rather laborious procedure because of similar chromatoꢀ
graphic mobilities of these compounds.
the signal for the phosphorus atom in the spectrum of
phosphate 3 is apparently associated with the electronꢀ
withdrawing effect of the methoxycarbonyl group. The
13
C NMR spectrum of phosphate 3 shows doublets correꢀ
2
sponding to the C(2) (δ 100.0, J
= 7.5 Hz) and C(3)
C,P
3
(δ 37.2, J
= 5.3 Hz) atoms with the characteristic
C,P
spinꢀspin coupling constants between the carbon atoms
and the phosphorus atom.
The β configuration of the anomeric C(2) center of
phosphate 3 was additionally confirmed by the small spinꢀ
spin coupling constant between the C(1) atom (the
Salt 2b was synthesized in situ by the reaction of
i
Pr NEt with dibenzyl phosphoric acid (2a). Salt 2c, which
methoxycarbonyl group) and the axial proton at the C(3)
2
3
is prepared by the reaction of an aqueous solution of
atom ( J
< 1 Hz). In the case of the α configuraꢀ
C(1),Hax(3)
3
Bu NOH with acid 2a in MeCN, was isolated and dried
tion of the anomeric center in Neu5Ac, J
be equal to ca. 5—6 Hz.
would
C(1),Hax(3)
4
8
before the reaction with chloride 1. In all cases, the reacꢀ
tions were carried out with samples of salt 2b or 2c conꢀ
taining a small excess (10%) of dibenzyl phosphoric acid
Stereoselectivity of the formation of βꢀphosphate 3,
i.e., the overall retention of the configuration of the
anomeric center upon the nucleophilic substitution of the
Cl atom in glycosyl chloride 1, is apparently associated
with anomerization of the initially formed αꢀphosphate
due to the known fact that β isomers of Neu5Ac are therꢀ
(
2a) to prevent basic conditions (which can happen in the
case of the inaccurate measurement of the amount of a
base in the step of preparation of the salt), which favor the
formation of glycal 4 from chloride 1.6
2
The difference in the reaction rate in going from 2b to
modynamically favorable. It should be noted that it is
2
c (Table 1) can be attributable to the fact that the salt of
generally possible to isolate the kinetically controlled reꢀ
action product in the synthesis of other glycosyl phosꢀ
phates with the use of dibenzyl phosphoric acid salts.9
To summarize, we demonstrated that the reactions of
glycosyl chloride 1 with dibenzyl phosphoric acid salts
2b,c afford Neu5Ac βꢀphosphate 3 in high yield (72%;
80% in the reaction mixture according to NMR analysis).
Thus, this method is much more efficient than the
procedure for the synthesis of compound 3 based on
glycosylation of dibenzyl phosphoric acid with glycosyl
+
–
a quaternary ammonium base [Bu N] [OP(O)(OBn) ]
4
2
(
[
2c) is more ionic than the salt of tertiary amine
Pr NHEt] [OP(O)(OBn) ] (2b). Apparently, in MeCN
i
+
–
2
2
salt 2c forms a less tight ion pair than salt 2b and the
nucleophilicity of the phosphate anion in 2c can be exꢀ
pected to be higher.
Phosphate 3 was characterized by 1H, 13C, and
31
1
P NMR spectroscopy. The H NMR spectrum of βꢀglyꢀ
cosyl phosphate 3 synthesized in the present study is idenꢀ
4
,5
5
tical to that described in the literature, which is eviꢀ
dence that the configuration of compound 3 is β. The
presence of the phosphoric acid residue in reaction prodꢀ
phosphite 6b (53% yield). The approach to the synthesis
of Neu5Ac βꢀphosphate 3 developed in the present study
is experimentally simple, and the starting reagents are
readily accessible.
Table 1. Influence of the nature of the phosphate reagent
–
+
(
BnO) P(O) X (2a—c) on the results of the reaction with glyꢀ
2
Experimental
cosyl chloride 1
The reactions were performed with the use of commercial
reagents (BnO) P(O)OH (Aldrich), a 20% aqueous Bu NOH
Reꢀ
agent
Counterꢀ Amount of 2 τa/h
Yield
of 3 (%)
Ratio
3 : 4b
2
4
ion
(equiv.)
solution (Merck), 85% H PO4 (Fluka), and AcCl (Fluka).
3
Diisopropylethylamine (Aldrich) was distilled over CaH2 unꢀ
der Ar. The solvents were distilled and purified before use acꢀ
cording to standard procedures. Acetonitrile used for the synꢀ
2
2
2
a
b
c
Н+
Pr EtNH
+
Bu N
4
1.5
4
4
12
24
12
0c
—
i
+
d
e
72 (80 ) 4 : 1
63 (75 ) 3 : 1
2
d
e
thesis of compound 3 was freshly distilled over CaH under Ar.
2
a
The time of complete conversion of chloride 1 (TLC).
H NMR spectroscopic data.
No reaction.
After column chromatography on silica gel.
Thinꢀlayer chromatography was carried out on silica gel plates
Kieselgel 60 F254 (Merck) using AcOEt as the eluent; spots
were visualized by heating plates after immersion in a 1 : 10
b 1
c
d
mixture of 85% aqueous H PO4 and 95% EtOH or under
3
e 1
H NMR spectroscopic data for the reaction mixture after
workup.
UV light. Column chromatography was performed on silica gel
L 100—250 µm (Chemapol, Czech Republic) using AcOEt as

Neu5Ac dibenzyl phosphate
Russ.Chem.Bull., Int.Ed., Vol. 53, No. 3, March, 2004
719
the eluent. The NMR spectra were recorded on a Bruker ACꢀ200
instrument. The 1H chemical shifts are given relative to the
solution of chloride 1, which was prepared from acetate 5b
(56 mg, 0.1 mmol), in MeCN (2 mL). The flask in which salt 2c
was prepared was additionally rinsed with MeCN (3×0.5 mL).
The reaction mixture was stirred at 22 °C. After completion of
the reaction (12 h), the products were isolated analogously to
the method A. After chromatography, a mixture of 3 and 4 was
13
residual signal of CHCl (δ 7.27), the C chemical shifts were
3
measured relative to the signal of CDCl (δ 77.0), and the
3
31
P chemical shifts are given relative to 75% H PO in D O as
3
4
2
the external standard (δ 0.0). The assignment of the signals in
13
1
the C NMR spectra was made based on the DEPT135 exꢀ
periment. Phosphates were synthesized at room temperature
obtained in a yield of 50 mg (3 : 1, H NMR data), i.e., the yield
of phosphate 3 was 63%.
1
(
20—25 °C) in anhydrous solvents under dry argon. All reacꢀ
Compound 3. H NMR (CDCl ), δ: 1.87 (s, 3 H, AcN);
3
tions were carried out with the use of glycosyl chloride 1 freshly
prepared from peracetate 5b according to a modified procedure.3a
The ratio between phosphate 3 and glycal 4 in a mixture of
1.99, 2.00, 2.05, and 2.10 (all s, 3 H each, AcO); 1.99—2.10 (m,
1 H, H (3)); 2.60 (dd, 1 H, H (3), J = 5.0 Hz,
ax
eq
Heq(3),H(4)
JHeq(3),Hax(3) = 13.6 Hz); 3.65 (s, 3 H, OMe); 4.20—4.30 (m, 3 H,
H(5), H(6), H (9)); 4.60 (dd, 1 H, H (9), J = 10.1 Hz,
the reaction products was determined by integration of the sigꢀ
a
b
Hb(9),Ha(9)
1
nals of H(3) and C(1)O CH in the H NMR spectra (3: δ 2.62
J
= 2.2 Hz); 4.98—5.09 (m, H(4)); 5.03 (d, OCH Ph,
2
3
Hb(9),H(8)
2
3
3
(
H (3)), δ 3.65 (OCH ); 4: δ 5.95 (H(3)), δ 3.78. (OCH )). The
JH,P = 2.2 Hz); 5.08 (d, OCH Ph, JH,P = 3.1 Hz) (a total
eq
3
3
2
yield of phosphate 3 was calculated taking into account the
molar ratios between phosphate 3 and glycal 4.
of 5 H); 5.35—5.40 (m, 2 H, H(7), H(8)); 5.63 (d, 1 H,
C(5)NH, J
= 9.5 Hz); 7.30—7.41 (m, 10 H, Ph).
H(5),NH
13
Methyl (5ꢀacetamidoꢀ4,7,8,9ꢀtetraꢀOꢀacetylꢀ3,5ꢀdideoxyꢀβꢀ
DꢀglyceroꢀDꢀgalactoꢀnonꢀ2ꢀulopyranosyl chloride)onate (1). Anꢀ
hydrous MeOH (1.8 mL, 0.04 mol) was slowly added dropwise
to AcCl (4.5 mL, 0.06 mol) with cooling in an ice water bath.
The reaction mixture was added to a cold solution of Neu5Ac
acetate 5b (102 mg, 0.19 mmol) in a mixture of anhydrous
CH Cl (4.5 mL) and AcCl (4.5 mL, 0.06 mol) and the reaction
C NMR (CDCl ), δ: 20.7 (AcO); 22.9 (NAc); 37.2 (d,
3
C(3), 3JC,P = 6.6 Hz); 48.2 (C(5)); 55.5 (MeO); 62.5 (C(9));
68.0 (C(8)); 69.5 (C(7)); 70.1 (d, OCH Ar, JC,P = 6.0 Hz);
71.9 (C(6)); 73.4 (C(4)); 100.0 (d, C(2), J
2
2
2
= 6.6 Hz); 128.3,
= 6.6 Hz); 165.8
C,P
C,P
3
128.4, 128.6, 128.7 (Ph); 135.2 (OCH C, J
2
(C(1)); 170.0 (CH CONH); 170.3, 170.6, 171.0 (CO).
3
31
P NMR (CDCl ), δ: –6.4.
2
2
3
1
13
mixture was kept at +4 °C for 12 h (TLC control, R 0.50 (1),
The H and C NMR spectra of glycal 4 are identical to
f
R 0.27 (5b)). Volatile components were evaporated, CCl was
those published in the literature.6b
f
4
added, and volatile components were again evaporated (5×5 mL).
The residue was dried in vacuo (oil pump) to give glycosyl chloꢀ
ride 1 (114 mg), which was used without additional purification.
The H and C NMR spectra of the product are identical to
those described in the literature.3
Methyl 5ꢀacetamidoꢀ4,7,8,9ꢀtetraꢀOꢀacetylꢀ2ꢀOꢀdibenzylꢀ
oxyphosphorylꢀ3,5ꢀdideoxyꢀβꢀDꢀglyceroꢀDꢀgalactoꢀnonꢀ2ꢀuloꢀ
pyranosonate (3). A. Diisopropylethylamine (65 µL, 0.37 mmol)
was added to a solution of dibenzyl phosphoric acid (2a) (115 mg,
This study was financially supported by the Russian
Foundation for Basic Research (Project Nos. 02ꢀ03ꢀ
32271, 03ꢀ03ꢀ06315ꢀmas, and 04ꢀ03ꢀ32854) and the
Grant from the President of the Russian Federation
1
13
(
the Federal Program for the Support of Leading Sciꢀ
entific Schools of the Russian Federation, Grant
NShꢀ1557.2003.3).
0
.41 mmol) in MeCN (1 mL). The solution of salt 2b thus
References
prepared was added dropwise to a solution of glycosyl chloꢀ
ride 1, which was prepared from acetate 5b (50 mg, 0.09 mmol),
in MeCN (2 mL). The vessel in which salt 2b was prepared was
additionally rinsed with MeCN (2×1 mL). The course of the
reaction was monitored by TLC (R 0.50 (1), R 0.48 (3)). After
completion of the reaction (24 h), the reaction mixture was
cooled in an ice water bath and then a cold saturated NaHCO3
solution (10 mL) and CHCl (10 mL) were added. The aqueous
1
2
3
. Glycosciences. Status and Perspectives, Eds H. J. Gabius and
S. Gabius, Chapman and Hall, Weinheim, 1997, 631 pp.
. G.ꢀJ. Boons and A. V. Demchenko, Chem. Rev., 2000,
00, 4539.
. (a) L. O. Kononov and G. Magnusson, Acta Chem. Scand.,
998, 52, 141; (b) M. N. Sharma and R. Eby, Carbohydr.
f
f
1
1
3
Res., 1984, 127, 201.
phase was extracted with CHCl (3×5 mL). The organic phase
3
4
5
. M. M. Sim, H. Kondo, and C.ꢀH. Wong, J. Am. Chem. Soc.,
was concentrated in vacuo (without heating). The residue was
1
993, 115, 2260.
. T. J. Martin and R. R. Schmidt, Tetrahedron Lett., 1993,
4, 1765.
6. (a) P.Meindl and H. Tuppy, Monatsh. Chem., 1969, 100, 1295;
b) K. Okamoto, T. Kondo, and T. Goto, Bull. Chem. Soc.
dissolved in CH Cl and applied onto a silica gel column packed
2
2
in light petroleum. The products were eluted with AcOEt and a
3
mixture of 3 and 4 was obtained in a yield of 57 mg (4 : 1,
1
H NMR data, R 0.48 (3), R 0.49 (4)), i.e., the yield of phosꢀ
f
f
(
phate 3 was 72% (with respect to acetate 5b).
Jpn, 1987, 60, 631.
B. A suspension of acid 2a (120 mg, 0.43 mmol) in MeCN
7
8
. R. Lederkremer, J. Org. Chem., 1994, 59, 690.
. H. Hori, T. Nakajima, Y. Nishida, H. Ohrui, and H. Meguro,
Tetrahedron Lett., 1988, 29, 6317.
. V. N. Shibaev and L. L. Danilov, in Glycopeptides and Related
Compounds, Eds D. G. Large and C. D. Warren, Marcel
Dekker, New York, 1997, p. 427.
(
3 mL) was stirred until complete dissolution (ca. 3 min). Then
a 20% aqueous solution of Bu NOH (172 µL, 0.39 mmol) was
4
added, the reaction mixture was stirred for 15 min, volatile
components were evaporated, MeCN was added, and volatile
components were again evaporated (4×2 mL). The residue was
dried in vacuo (oil pump) for 2 h to give crude salt 2c in a yield of
9
31
2
19 mg ( P NMR (CDCl ): δ –0.5). Then MeCN (1 mL) was
3
added and the resulting suspension was added dropwise to a
Received January 14, 2004