A Simple synthesis of sugar nucleoside diphosphates by chemical coupling in water

Article abstract of DOI:10.1002/anie.201205433

Sugar nucleotides made easy: The new reagent ImIm , which is formed in-situ in water, is shown to activate nucleoside 5'-phosphates to their imidazolides, these can subsequently couple with sugar-1-phosphates; the whole procedure takes place in water. This truly simple method yields a crude product mixture that can be used directly as a source of donors for glycosyltransferase- mediated oligsaccharide synthesis. In the scheme, B stands for the nucleobases U, A, or G. Copyright

Full text of DOI:10.1002/anie.201205433

Angewandte
Chemie
DOI: 10.1002/anie.201205433
Synthetic Methods
A Simple Synthesis of Sugar Nucleoside Diphosphates by Chemical
Coupling in Water**
Hidenori Tanaka, Yayoi Yoshimura, Malene R. Jørgensen, Jose A. Cuesta-Seijo, and
Ole Hindsgaul*
The use of glycosyltransferases in chemoenzymatic oligosac-
charide synthesis is attractive, provided that the enzymes are
available, since it eliminates the tedious multistep protection-
deprotection and chemical glycosylation procedure that
characterizes classic chemical synthesis.[1,2] The sugar nucleo-
side diphosphates (sugar-NDPs) that most glycosyltransfer-
ases require as glycosyl donors (e.g. UDP-Glc, UDP-Gal,
ADP-Glc etc.) can be prepared either chemically or through
enzymatic methods including recycling systems. The enzy-
matic preparation of sugar-NDPs has the advantages of being
simple and compatible with glycosyltransferase reactions but
requires access to the necessary enzymes. In contrast,
chemical synthesis is more complex but is especially powerful
for the preparation of analogues that cannot easily be
prepared using enzymes owing to a low tolerance of the
enzyme for substrate modifications.
We report herein a simple method for the chemical
synthesis of sugar-NDPs in a one-pot reaction in water. The
procedure involves only the sequential addition of commer-
cially available compounds, and the crude product solution
can be used directly as a glycosyl donor source in glycosyl-
transferase-mediated oligosaccharide synthesis.
We found that the reaction of 2-chloro-1,3-dimethylimi-
dazolinium chloride (DMC, 1)[8,9] with imidazole (2) in D2O at
room temperature gave 2-imidazolyl-1,3-dimethylimidazoli-
nium chloride (ImIm, 3), within five minutes, along with
imidazole hydrochloride (Im-HCl; Scheme 1). The formation
of 3 was confirmed by 1H NMR spectroscopy and ESI–MS.[10]
We then examined whether 3 would convert phosphate
groups into reactive phosphorimidazolide groups in aqueous
solution. Phosphorimidazolides have been extensively used
for pyrophosphate bond formation in anhydrous organic
solvents.[3]
Chemical methods for the synthesis of sugar-NDPs have
recently been comprehensively reviewed and eloquently
discussed.[3] The most commonly used procedures involve
the formation of the pyrophosphate linkage by coupling of an
activated nucleoside-5’-monophosphate (NMP, typically
a morpholidate or imidazolide) with the sugar-1-phosphate;
in these reactions the addition of catalysts like nitrogen-
containing heterocycles can effect substantial improvements
in both rate and yield.[4,5] Newer methods include the
“cycloSal” phosphotriester coupling[6] and mixed phosphite–
phosphate formation followed by oxidation.[7] Direct coupling
of anomerically activated monosaccharides with nucleoside
diphosphates, such as uridine 5’-diphosphate, represents
a useful alternative approach (reviewed in reference [3]).
These chemical procedures are normally performed in
anhydrous organic solvents and usually employ per-O-acety-
lated sugars and phosphates, such as their tetraalkylammo-
nium salts, that are soluble in organic solvents. The syntheses
are notoriously difficult to reproduce from laboratory to
laboratory, especially when researchers have little previous
experience with chemical sugar nucleotide synthesis.
Scheme 1. Reaction of DMC (1) with imidazole (2) gives ImIm (3) as
a reactive intermediate.
Uridine monophosphate (UMP) disodium salt (4), DMC
(1), and imidazole (2) were dissolved in D2O at 378C, and the
reaction was monitored by 1H NMR spectroscopy (Figure 1).
The signal corresponding to H-5 of uracil in 4 (d =
8.01 ppm) decreased and simultaneously a peak assigned to
H-5 of UMP-imidazolide (UMP-Im, 5; d = 7.62 ppm)
increased for up to one hour where the maximum conversion
to 5 was approximately 70%. The reaction was also moni-
tored using 31P NMR spectroscopy, where the intensity of the
characteristic[5,11] signal for UMP-imidazolide 5 (UMP-Im,
d = ꢀ8.2 ppm) was observed. ESI–MS on the crude reaction
mixture further confirmed the formation of UMP-Im (m/z
373.3; see the Supporting Information). An identical mass
spectrum was obtained if DMC and imidazole were first
reacted for five minutes and subsequently UMP was added.
Reaction times longer than one hour led to hydrolysis of 5
and subsequent increased dimerization (4 + 5) to give UMP-
dimer (7; Figure 1). If the reaction was allowed to proceed for
16 h, only hydrolysis product (5!4) and dimer 7 were seen: at
that stage the reactive species 3 and 5 were fully hydrolyzed,
thereby rendering the solution chemically benign.
[*] H. Tanaka, Y. Yoshimura, M. R. Jørgensen, J. A. Cuesta-Seijo,
O. Hindsgaul
Carlsberg Laboratory
Gamle Carlsberg Vej 10, DK-1799, Copenhagen-V (Denmark)
E-mail: ole.hindsgaul@carlsberglab.dk
[**] We acknowledge Dr. Morten Munch Nielsen for his contribution
with the initial cloning and expression of starch synthase I, and Prof.
Monica M. Palcic for her guidance in all of the enzymology aspects
of this work.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201205433.
Angew. Chem. Int. Ed. 2012, 51, 11531 –11534
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
11531

.
Angewandte
Communications
spectroscopy. However, activation of 4 was also examined in
H2O, where the yield of UMP-Im (5) was consistently lower,
only 60–70% of that obtained in D2O. Both 3 and 5 were
hydrolyzed much faster in H2O than in D2O (verified by
kinetic NMR monitoring, see the Supporting Information);
this result is attributed to the known solvent isotope effect of
water in hydrolysis reactions.[12] The use of D2O is therefore
advantageous in preserving a high concentration of the active
species (3 and 5) in aqueous solution.
We next examined conditions for coupling the in situ
formed UMP-Im (5) with glucose-1-phosphate (Glc-1-P, 9) to
give UDP-glucose (UDP-Glc, 10) as the target sugar-NDP.
The products were quantitated by 31P NMR spectroscopy by
using the beneficial D2O as solvent (Table 1). The activation
Table 1: Optimization of reaction conditions for UDP-Glc synthesis
using ImIm activation.
Entry
D2O
Ratio
(4/1/2/9)
Additive
Yield
[%][a]
[Lmolꢀ1 of UMP]
1
2
3
4
5
6
7
1.0
1.0
1.0
1.0
0.5
0.5
0.5
1:2:4:1
1:2:4:1
1:2:4:1
2:4:8:1
2:4:8:1
4:8:16:1
4:8:16:1
none
20
23
21
34
43
57
43[b]
TEAC
[15]crown-5
none
none
none
none
[a] Yields of 10 based on Glc-1-P and calculated from the corresponding
signal integrations in the 31P NMR spectrum of the crude mixture.
[b] Yield of isolated 10, which was purified by ion-exchange chromatog-
raphy on DEAE-Sephacel after treatment with alkaline phosphatase for
24 h at 308C; the yield is based on Glc-1-P and was determined by UV
spectrosopy at lmax 260 nm.
step to produce 5 (1 + 2 + 4) was run for one hour (at which
time the majority of the ImIm (3) was consumed) before
adding Glc-1-P (dipotassium salt) 9. It is important to
minimize the activation of Glc-1-P by unhydrolyzed ImIm,
because this activation leads to the formation of glucose-1,2-
cyclic-phosphate. The rate of formation of UMP-Im (5) is,
however, dependent on the concentration of ImIm, so
a compromise in reaction time and corresponding yield
must be made to provide the highest possible concentration of
reactive 5 for coupling to Glc-1-P.
Coupling reactions were examined under a variety of
conditions at 378C for 18 h (Table 1). Addition of tetraethyl-
ammonium chloride (TEAC) or [15]-crown-5 was not bene-
ficial (Table 1, entries 1–3). Furthermore, use of dimethyl
sulfoxide (DMSO) as a co-solvent led to a decrease in
coupling yield (data not shown). An increase in the concen-
tration of 5 would be predicted to improve the efficiency of
coupling with 9. Indeed, higher yields were obtained when the
amounts of 1, 2, and 4 were increased and the volume of D2O
Figure 1. Activation of UMP (4) in D2O with ImIm monitored by
1H NMR spectroscopy.
A plausible reaction mechanism is presented in Figure 1,
where the rapidly formed ImIm activates the phosphate via an
intermediate isourea that is then captured by imidazole to
produce UMP-Im and urea 6. In an independent experiment
where DMC alone was incubated with UMP for 24 h at 378C,
the only products observed were urea 6 (40%) and UMP-
dimer (< 5%) along with 60% of unreacted DMC and more
than 95% unreacted UMP (see the Supporting Information).
The reactions described above were optimized in D2O,
1
because its use facilitates reaction monitoring by H NMR
11532 www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 11531 –11534

Angewandte
Chemie
decreased (Table 1, entries 4–6). The highest yield of crude
product 10 was 57%, estimated by NMR spectroscopy
(Table 1, entry 6). To facilitate isolation of 10, we used
alkaline phosphatase to cleave the phosphate group in
unreacted 4 and then purified the product by ion exchange
chromatography on DEAE-Sephacel leading to isolation of
10 in 43% yield (Table 1, entry 7).
The scope of the reaction with respect to different NMPs
and sugar-1-phosphates was briefly explored as summarized
in Table 2. The optimized conditions[13] led to the formation of
donor in a glycosyltransferase-catalyzed synthesis of oligo-
saccharides.
To this end, the unprocessed reaction mixtures containing
UDP-Gal (Table 2, entry 1) and ADP-Glc (Table 2, entry 2),
were evaluated as donors for two galactosyltransferases and
one glucosyltransferase, respectively (Scheme 2). The mix-
ture containing crude UDP-Gal synthesized by the present
method was added to either bovine b-1,4-galactosyltransfer-
ase[14] (an inverting galactosyltransferase) or human-blood-
group-B galactosyltransferase[15] (a retaining galactosyltrans-
ferase, GTB) in the presence of their known glycosyl accept-
ors GlcNAc-TMR[16] and FucGal-TMR,[17] respectively. After
incubation for 12 h, both galactosyltransferase incubations
resulted in complete conversion of the acceptor to the
galactosylated product when monitored by MALDI-TOF
MS and capillary electrophoresis (see the Supporting Infor-
mation). When compared to using pure commercial UDP-
Gal, the synthetic mixture gave a 2–4 times slower reaction in
the initial rate (see the Supporting Information). This result
was attributed to inhibition of the enzymes by both UMP and
UMP-dimer (7), which had inhibition constants (Ki) in the
same range as the Michaelis–Menten constant (Km) for UDP-
Gal.[18] However, the synthetic UDP-Gal mixture remained
completely effective for milligram-scale product synthesis in
an overnight incubation.
The crude product mixture containing synthetic ADP-Glc
(Table 2, entry 2) was incubated with recombinant barley
starch synthase I in the presence of maltopentaose as an
acceptor (Scheme 2c). After the incubation, the resultant
mixture was labeled with a fluorescent tag at the reducing
end.[19] Ultra-performance liquid chromatography (UPLC)–
MS analysis revealed that 1, 2, or 3 glucose residues had been
transferred to the acceptor, thereby giving maltooligosac-
charides 12–14 as products. In this case, inhibition of the
enzyme was not observed: in fact the reaction proceeded
more rapidly than with commercial ADP-Glc (see the
Supporting Information).
Table 2: Substrate scope in sugar-NDP synthesis using ImIm activation.
Entry
NMP
Sugar-1-P
NDP-Sugar
Yield [%][b,c]
1
UMP (4)
AMP
GMP
Gal-1-P
Glc-1-P
Man-1-P
UDP-Gal
ADP-Glc
GDP-Man
59 (35)
41 (32)
<50 (23)
2
3[a]
[a] DMC (16 equiv), imidazole (32 equiv), and D2O (1 Lmolꢀ1 of GMP)
were used and the activation step was performed for 2 h. [b] Yields of the
coupling product were based on sugar-1-P and calculated from the
corresponding signal integrations in the 31P NMR spectrum of the crude
mixture. [c] In parentheses, the yields of isolated coupling products are
given, which were purified by ion-exchange chromatography on DEAE-
Sephacel after alkaline phosphatase treatment for 24 h at 308C; yields
were based on sugar-1-P and determined by UV spectroscopy at lmax
260 nm (uracil and adenine) and lmax 252 nm (guanine). In the scheme,
B stands for the nucleobases U, A, or G.
UDP-Gal in a reasonable yield (59%). Although adenosine
monophosphate (AMP) possesses a free amino group avail-
able for potential reaction, ADP-Glc was successfully
obtained in a 41% yield. However,
guanosine monophosphate (GMP) has
limited solubility in water and required
a larger volume of D2O (1 L/mol of
GMP) resulting in a diminished yield. To
activate GMP efficiently, an increase in
both the amount of 1 and 2 and in the
time for the activation step resulted in
the formation of GDP-Man in a yield
close to 50% (23% isolated).
One objective of this work was to
provide a method for the sugar-NDP
preparation that could be performed
without any skill in organic synthesis,
thus eliminating the need for sugar
protection and anhydrous organic sol-
vents while employing only commer-
cially available reagents. However, as
alluded to above, the second objective
was that the final sugar-NDP be present
in a chemically benign solution that
Scheme 2. Enzymatic reactions: a) UDP-Gal with bovine b-1,4-galactosyltransferase (b-1,4-GalT);
could directly serve as the source of b) UDP-Gal with GTB; c) ADP-Glc with barley starch synthase I followed by fluorescent labeling.
Angew. Chem. Int. Ed. 2012, 51, 11531 –11534
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 11533

.
Angewandte
Communications
c) T. Tanaka, H. Nagai, M. Noguchi, A. Kobayashi, S. Shoda,
In conclusion, the new ImIm reagent formed in situ can
activate nucleoside 5’-monophosphates in D2O to give
a reactive phosphorimidazolide intermediate. Coupling of
this intermediate with sugar-1-phosphates in D2O produced
NDP-sugars in useful yields based on the sugar-1-phosphate
as the limiting reagent. All reactions occur in one pot, and the
procedure involves only the timed sequential addition of
commercial chemicals. Furthermore, all chemically reactive
species are hydrolyzed by the end of the coupling reaction,
thus the crude NDP-sugar solutions can be used directly
without any purification for preparative oligosaccharide
synthesis by glycosyltransferases.
Chem. Commun. 2009, 3378 – 3379.
1
[10] ImIm (3): H NMR (400 MHz, D2O): d = 8.11 (s, 1H), 7.48 (s,
1H), 7.32 (s, 1H), 4.08 (s, 4H), 3.01 ppm (s, 6H); m/z (ESI)
found [MꢀCl]+ 165.1, C8H13N4Cl calcd for [MꢀCl]+ 165.1.
[11] a) A. Zatorski, B. M. Goldstein, T. D. Colby, J. P. Jones, K. W.
Pankiewicz, J. Med. Chem. 1995, 38, 1098 – 1105; b) G. Baisch, R.
ꢁhrlein, Bioorg. Med. Chem. 1997, 5, 383 – 391; c) Y. Zhao, J. S.
Thorson, J. Org. Chem. 1998, 63, 7568 – 7572; d) A. L. Marlow,
L. L. Kiessling, Org. Lett. 2001, 3, 2517 – 2519.
[12] M. Komiyama, M. L. Bender, Bioorg. Chem. 1978, 7, 133 – 139.
[13] Typical procedure for the one-pot sugar-NDP synthesis using
ImIm activation: DMC (200 mmol), imidazole (400 mmol), and
nucleoside monophosphate disodium salt (100 mmol) were
placed in an Eppendorf tube with a stir bar. D2O (50 mL) was
added. The mixture was stirred at 378C for 1 h, and subsequently
sugar phosphate dipotassium salt (25 mmol) was added. After
further stirring at 378C for 18 h, the crude mixture can be used as
a source of glycosyl donor as described. When purification is
desired, the crude mixture was diluted with tris(hydroxymethyl)-
aminomethane–HCl (Tris-HCl) buffer (50 mm, pH 8.0, 5.0 mL)
and alkaline phosphatase (50 U) was added. After incubation at
308C for 24 h, the mixture was purified by DEAE Sephacel ion
exchange chromatography (flow rate: 2.0 mLminꢀ1, Eluting
solvent: 30 mm to 400 mm aq. ammonium acetate). The fractions
containing the product were collected, concentrated, and
lyophilized to give the corresponding coupling product at
> 90% of purity level.
Received: July 10, 2012
Revised: September 1, 2012
Published online: October 12, 2012
Keywords: glycosyltransferases · oligosaccharides ·
.
pyrophosphate formation · sugar nucleotide synthesis ·
synthetic methods
[1] R. M. Schmaltz, S. R. Hanson, C.-H. Wong, Chem. Rev. 2011,
111, 4259 – 4307.
[2] M. M. Palcic, Curr. Opin. Chem. Biol. 2011, 15, 226 – 233.
[3] G. K. Wagner, T. Pesnot, R. A. Field, Nat. Prod. Rep. 2009, 26,
1172 – 1194.
[14] B. Ramakrishnan, P. S. Shah, P. K. Qasba, J. Biol. Chem. 2001,
276, 37665 – 37671.
[4] V. Wittmann, C.-H. Wong, J. Org. Chem. 1997, 62, 2144 – 2147.
[5] H. Tsukamoto, D. Kahne, Bioorg. Med. Chem. Lett. 2011, 21,
5050 – 5053.
[15] S. L. Marcus, R. Polakowski, N. O. L. Seto, E. Leinala, S.
Borisova, A. Blancher, F. Roubinet, S. V. Evans, M. M. Palcic,
J. Biol. Chem. 2003, 278, 12403 – 12405.
[6] S. Wendicke, S. Warnecke, C. Meier, Angew. Chem. 2008, 120,
1523 – 1525; Angew. Chem. Int. Ed. 2008, 47, 1500 – 1502.
[7] H. Gold, P. van Delft, N. Meeuwenoord, J. D. C. Codꢀe, D. V.
Filippov, G. Eggink, H. S. Overkleeft, G. A. van der Marel, J.
Org. Chem. 2008, 73, 9458 – 9460.
[16] Y. Zhang, X. Le, N. J. Dovichi, C. A. Compston, M. M. Palcic, P.
Diedrich, O. Hindsgaul, Anal. Biochem. 1994, 227, 368 – 376.
[17] S. Lafertꢀ, N. W. C. Chan, K. Sujino, T. L. Lowary, M. M. Palcic,
Eur. J. Biochem. 2000, 267, 4840 – 4849.
[18] Kinetic parameters determined as described in the Supporting
Information: Km (UDP-Gal)/Ki (UMP)/Ki (UMP-dimer). For b-
1,4-GalT: 22:149:36 mm. For GTB: 27:9.9:34 mm.
[19] A modification of a reported solid-phase labeling protocol was
used: A. Lohse, R. Martins, M. R. Jørgensen, O. Hindsgaul,
Angew. Chem. 2006, 118, 4273 – 4278; Angew. Chem. Int. Ed.
2006, 45, 4167 – 4172.
[8] T. Isobe, J. Org. Chem. 1999, 64, 6984 – 6988.
[9] DMC (1) has recently been applied to some novel transforma-
tions of unprotected sugars in water: a) M. Noguchi, T. Tanaka,
H. Gyakushi, A. Kobayashi, S. Shoda, J. Org. Chem. 2009, 74,
2210 – 2212; b) T. Tanaka, W. C. Huang, M. Noguchi, A.
Kobayashi, S. Shoda, Tetrahedron Lett. 2009, 50, 2154 – 2157;
11534 www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 11531 –11534