Efficient synthesis of nucleoside diphosphate glycopyranoses

Article abstract of DOI:10.1002/anie.200703237

(Chemical Equation Presented) Short and sweet: A new, short synthetic pathway for the synthesis of the enormously important class of nucleoside diphosphate sugars (NDP sugars) was developed using cyclo-saligenyl (cycloSal) nucleosyl phosphate triesters as

Full text of DOI:10.1002/anie.200703237

Communications
DOI: 10.1002/anie.200703237
Carbohydrate Synthesis
Efficient Synthesis of Nucleoside Diphosphate Glycopyranoses**
Silke Wendicke, Svenja Warnecke, and Chris Meier*
Nucleoside diphosphate pyranoses 1 (NDP sugars; see
Scheme 1) play a key role as glycosyl donors in the synthesis
of oligo- and polysaccharides.[1,2] Moreover, they serve as
precursors of deoxysugars, aminodeoxysugars, chain-
branched sugars, uronic acids, as well as glycoconjugates. In
biosynthetic pathways the energy-rich linkage between the C1
atom and the b-phosphate is cleaved, and thus the glycosyl
part is enzymatically transferred to an oligosaccharide chain,
releasing the nucleoside diphosphate (NDP) moiety. For
biosynthesis studies of oligosaccharides (for example, of
lipopolysaccharides)[3] an efficient access to this important
class of compounds is needed. The classical method is the
coupling of glycosyl 1-phosphates to nucleotide morpholi-
dates (Moffat–Khorana method).[4,5] However, this reaction
normally takes days, and the chemical yields are often low (5–
25%). Attempts to improve the reaction yields by using 1H-
tetrazole as activator[6–9] are often not successful and have
failed also in our hands. Instead of the morpholidates, also
imidazolides have been used in the past but without improv-
ing the yields markedly.[10–12] Alternatively, Hindsgaul and
Jakeman published a procedure starting from nucleoside
diphosphates and glycopyranosyl bromides.[13,14] However,
yields were also found to be low, and often the stereochem-
istry at the anomeric center could neither be controlled nor
stereospecifically formed. Thiem and co-workers reported an
enzymatic procedure starting from unprotected sugars that
were first phosphorylated and then treated with a nucleoside
triphosphate.[15,16] However, this reaction sequence is based
on a three-enzyme pathway and depends on the availability of
the needed kinases and NDP sugar pyrophosphorylases and
on expensive nucleoside triphosphates. The yields obtained
have seldom exceeded 30%. Herein, we report on a concep-
tionally new chemical synthesis of NDP sugars that uses
cyclo-saligenyl (cycloSal) nucleosyl phosphate triesters II as
an active ester (Scheme 1).
a–c; Scheme 1). The technique has been applied successfully
to a variety of nucleoside analogues, providing superior
antiviral activity.[18–20] However, the same type of compounds
may also be used as an active ester for synthetic applications.
Here, cycloSal nucleotides like II were treated with glyco-
pyranosyl 1-phosphate salts III–VI as nucleophiles with
formation of the pyrophosphate bond in NDP sugars (path-
way d,e; Scheme 1).
Scheme 1. General principle of the cleavage of cycloSal phosphate
triesters using water/hydroxide or pyranose 1-phosphates. See text for
details.
As starting materials, 5-nitro-cycloSal-3’-O-acetylthymi-
dine 2 and peracetylated glycopyranosyl phosphates 3–6 were
prepared (see Scheme 3). Thus, thymidine was protected by
silylation with tert-butyldimethylsilylchloride (TBDMS-Cl) at
the 5-position. The product was treated with acetic anhydride
to yield fully protected thymidine in 92% yield. Finally, the
TBDMS group was cleaved by (nBu)4NF to give 3’-O-
acetylthymidine (3’-OAcT) in 96% yield. This material was
converted into target triester 2 by reaction with 5-nitro-
cycloSal phosphorochloridite 7, which was prepared as
reported before starting from 2-hydroxymethyl-4-nitrophenol
(8).[21] Alcohol 8 was prepared from the salicylic acid
derivative 9 by borane reduction in THF (86% yield). The
intermediate phosphite was subsequently oxidized using
oxone (2KHSO5·KHSO4·K2SO4) to the cycloSal phosphate
triester (> 90% yield; Scheme 2). The product was pure
enough to be used as a raw material in the following reactions
with glycosyl 1-phosphates. However, when oxone was
replaced by the originally used tert-butylhydroperoxide, the
Originally, the cycloSal technique was developed to
deliver biologically active nucleotides into cells.[17] The
cleavage relies on a nucleophilic attack of the neutral
phosphate triester by water or hydroxide and a subsequent
selective hydrolysis pathway to yield the nucleotide (pathway
[*] Dr. S. Wendicke, Dipl. Chem. S. Warnecke, Prof. Dr. C. Meier
Institute of Organic Chemistry
Department of Chemistry
Faculty of Science
University of Hamburg
Martin-Luther-King-Platz 6, 20146 Hamburg (Germany)
Fax: (+49)404-2838-2495
E-mail: chris.meier@chemie.uni-hamburg.de
Homepage: http://www.chemie.uni-hamburg.de/oc/meier/
[**] This work has been accomplishedin the network of the SFB 470/A8.
We are grateful for financial support from the Deutsche For-
schungsgemeinschaft.
1500
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 1500 –1502

Angewandte
Chemie
Scheme 2. Synthesis of 5-nitro-cycloSal-(3’-OAc)-thymidine monophos-
phate.
Scheme 3. Synthesis of the NDP sugars starting from the cycloSal
phosphate triester.
yields were found to be significantly lower, and sometimes the
reaction failed completely. Alternatively, also an iodine/
pyridine/THF/water solution as used in oligonucleotide
synthesizers can be used.
silica gel. Usually, two passages were sufficient to isolate the
product in high purity. Yields after purification were found to
be between 46% and 60% (Table 1). As expected, the
spectroscopic characterization proved that the anomeric
configuration was unchanged in the product compared to
the glycopyranosyl 1-phosphates.
d-Glucose, d-mannose, d-galactose and 6-deoxy-d-
gulose[22] were peracetylated by standard conditions and
then converted into the 1-phosphate 3–6 by two different
methods. In the case of the a-configured glycosyl 1-phos-
phates 3(a)–6(a), pentaacetylglycopyranoses were depro-
tected at the anomeric center by hydrazinium acetate[23] and
subsequently treated with dibenzyl(diisopropylamino)phos-
phine and 1H-tetrazole to give the phosphite intermediate,[24]
which was oxidized by m-chloroperbenzoic acid at 08C to give
dibenzylphosphate triesters in 31–70% overall yield. For
formation of the b-configured counterparts, pentaacetylgly-
copyranoses were converted into the glycopyranosyl bro-
mides (HBr, HOAc), which were then treated with dibenzyl-
phosphite in the presence of Ag2CO3 at 08C.[25] This reaction
led to the exclusive formation of the b-d-glycopyranosyl
phosphates in 81–98% yield. In case of d-mannose even this
reaction sequence led to the formation of the a-configured
phosphate triester. Subsequently, the phosphate moieties
were deprotected by hydrogenolysis using Pd/C, H2, and
NEt3, which gave a- or b-configured d-glucose- (3(a) and
3(b)), d-galactose- (4(a) and 4(b)), d-mannose- (5(a)), and d-
6-deoxygulose phosphates (6(a) and 6(b)) as their triethy-
lammonium salts in 73–95% yield.
Table 1: Yields of the prepared dTDP pyranoses.
d-Glc
d-Man
d-Gal
6d-d-Gul
1a(a) 1a(b) 1b(a) 1b(b) 1c(a) 1c(b) 1d(a) 1d(b)
yield[%] 60 40 49 58 46 57
–
–
Instead of using the triethylammonium salt of the sugar
phosphates, also more lipophilic counterions (e.g. (nBu)4N+)
were investigated. Although the solubility of the salts was
improved and the reaction works as well, the purification of
the NDP sugars was extremely difficult.
To compare our novel method with one of the classical
methods, the synthesis of TDP d-glucose 1a(a) and 1a(b) was
repeated using the morpholidate protocol with either pyridine
or DMF as solvent. Although some of the starting morpho-
lidate was still present, the reaction was stopped after five
days. After deprotection and purification the target com-
pound 1a(a,b) was isolated in only 10% yield. Thus, our
disclosed novel method allows a fast and efficient preparation
of anomerically defined NDP glycopyranoses.
Using this new procedure, hitherto unknown TDP d-6-
deoxygulopyranoses 1d(b) were prepared and will be used
now for biosynthetic studies of lipopolysaccharides. More-
over, the presented method has also been applied to the
synthesis of NDP sugars with uridine, 2’-deoxyguanosine, and
cytidine as nucleosides.[26]
The highest yields for the coupling of 5-nitro-cycloSal
nucleotide 2 and the glycosyl 1-phosphates 3–6 were obtained
using DMF as solvent. After 3–5 h at room temperature the
cycloSal phosphate triester was completely consumed, and a
single product was formed in this highly stereospecific
reaction in all cases (Scheme 3). In contrast, with pyridine
as solvent, very long reaction times were required and
cycloSal triester 2 slowly decomposed.
Prior to purification, the acetyl groups in the crude
product were cleaved by a 1:7:3 mixture of NEt3/MeOH/H2O.
This deprotection before purification made the subsequent
isolation by chromatography decisively easier. TDP sugars
1a–d were finally purified on a glass column filled with RP-18
In summary, this report confirms for the first time that the
cycloSal technique not only can be used as a nucleotide
delivery system (pronucleotides) but is also applicable as a
new approach for the synthesis of biomolecule conjugates
linked by a pyrophosphate group.
Angew. Chem. Int. Ed. 2008, 47, 1500 –1502
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1501

Communications
Experimental Section
[1] E. F. Neufeld, W. Z. Hassid, Adv. Carbohydr. Chem. 1963, 18,
309.
[2] N. K. Kochetkov, V. N. Shibaev, Adv. Carbohydr. Chem. Bio-
chem. 1973, 28, 307.
[3] M. Skurnik, L. Zhang, APMIS 1996, 104, 849.
[4] S. Roseman, J. J. Distler, J. G. Moffatt, H. G. Khorana, J. Am.
Chem. Soc. 1961, 83, 659.
[5] J. G. Moffatt, H. G. Khorana, J. Am. Chem. Soc. 1958, 80, 3756.
[6] E. S. Simon, S. Grabowski, G. M. Whitesides, J. Org. Chem. 1990,
55, 1834.
[7] V. Wittmann, C.-H. Wong, J. Org. Chem. 1997, 62, 2144.
[8] R. E. Campbell, M. E. Tanner, J. Org. Chem. 1999, 64, 9487.
[9] A. Schäfer, J. Thiem, J. Org. Chem. 2000, 65, 2 4.
[10] Q. Zhang, H.-W. Lui, J. Am. Chem. Soc. 2000, 122, 9065.
[11] R. R. Schmidt, B. Wegmann, K.-H. Jung, Liebigs Ann. Chem.
1991, 121.
[12] S. Zamyatina, C. Gronow, M. Oertelt, H. Puchberger, P. Brade, P.
Kosma, Angew. Chem. 2000, 112, 4322; Angew. Chem. Int. Ed.
2000, 39, 4150.
Compound 2 (100 mg, 0.2mmol) and the corresponding glycopyr-
anosyl 1-phosphate 3–6 (1.2equiv) were dissolved in dry DMF
(5 mL), and the solution was stirred for 3–5 h at room temperature
(conversion monitored by TLC). The solvent was evaporated under
reduced pressure, and the residue was dissolved in H2O and extracted
with CH2Cl2. The aqueous phase was lyophilized and subsequently
treated with NEt3 (0.7 mL), MeOH (5 mL), and H2O (2mL). The
mixture was maintained at room temperature for 12h. After
lyophilization, the crude product was purified on on RP-18 column
using H2O or a H2O/MeCN gradient as eluent. Spectroscopic
characterization is exemplified as given for compounds 1c(a) and
1c(b). 1c(a): 1H NMR (400 MHz, D2O): d = 1.29 (t, 18H, JHH
=
3
7.4 Hz, 2” CH 3-NEt3), 1.94 (d, 3H, 4JHH = 1.0 Hz, thymine-CH3),
3
2.37–2.44 (m, 2H, H-2’), 3.21 (q, 12H, JHH = 7.4 Hz, 2” CH 2-NEt3),
3.71–3.83 (m, 3H, H-5, H-6), 3.93 (dd, 1H, 3JHH = 10.3 Hz, JHH
=
3
3.0 Hz, H-3), 4.04 (d, 1H, 3JHH = 3.0 Hz, H-4), 4.19–4.20 (m, 4H, H-2,
H-4’, H-5’), 4.63 (m, 1H, H-3’), 5.65 (dd, 1H, 3JHP = 7.3 Hz, JHH
=
3
3.8 Hz, H-1), 6.36 (dd, 1H, 3JHH = 6.5 Hz, 3JHH = 6.5 Hz, H-1’),
4
7.76 ppm (d, 1H, JHH = 1.0 Hz, thymine-CH3); 31P NMR (162MHz,
[13] M. Arlt, O. Hindsgaul, J. Org. Chem. 1995, 60, 14.
[14] S. C. Timmons, D. L. Jakeman, Org. Lett. 2007, 9, 1227.
[15] R. Stiller, J. Thiem, Liebigs Ann. Chem. 1992, 467.
[16] U. Gambert, J. Thiem, Top. Curr. Chem. 1997, 186, 2 1.
[17] C. Meier, Eur. J. Org. Chem. 2006, 1081.
[18] C. Meier, M. Lorey, E. De Clercq, J. Balzarini, J. Med. Chem.
1998, 41, 1417.
[19] C. Meier, T. Knispel, V. E. Marquez, M. A. Siddiqui, E.
De Clercq, J. Balzarini, J. Med. Chem. 1999, 42, 1615.
[20] C. Meier, A. Lomp, A. Meerbach, P. Wutzler, J. Med. Chem.
2002, 45, 5157.
D2O): d = À11.20 (d, JPP = 20.9 Hz, Pb), À12.66 ppm (d, JPP = 20.9 Hz,
Pa); HR-MS (ESI) calcd for C16H25N2O16P2 [M+H]: 563.0685; found:
1
3
563.0671. 1c(b): H NMR (400 MHz, D2O): d = 1.23 (t, 18H, JHH
=
7.3 Hz, 2” CH 3-NEt3), 1.88 (s, 3H, 4JHH = 1.0 Hz, thymine-CH3), 2.27–
2.39 (m, 2H, H-2’), 3.15 (q, 12H, 3JHH = 7.3 Hz, 2” CH 2-NEt3), 3.57
(dd, 1H, 3JHH = 7.6 Hz, 2JHH = 10.3 Hz, H-6), 3.63–3.67 (m, 1H, H-4),
3.70 (dd, 1H, JHH = 3.0 Hz, JHH = 10.3 Hz, H-6), 3.74–3.80 (m, 2H,
H-3, H-5), 3.87 (d, 1H, 3JHH = 3.3 Hz, H-2), 4.13–4.14 (m, 3H, H-4’, H-
5’), 4.58–4.60 (m, 1H, H-3’), 4.91 (dd, 1H, 3JHP = 7.6 Hz, 3JHH = 7.6 Hz,
3
2
3
3
H-1), 6.31 (dd, 1H, JHH = 6.6 Hz, JHH = 6.6 Hz, H-1’), 7.71 ppm (d,
1H, 4JHH = 1.0 Hz, thymine-CH3); 31P NMR (162MHz, D 2O): d =
À11.75 (d, JPP = 20.7 Hz, Pb), À13.24 ppm (d, JPP = 20.7 Hz, Pa);
HR-MS (ESI) calcd for C16H25N2O16P2 [M+H]: 563.0685; found:
563.0673.
[21] C. Meier, E. De Clercq, J. Balzarini, Eur. J. Org. Chem. 1998,
837.
[22] First we tried to reproduce the synthesis of 6d-d-gulose from an
earlier report (L. M. Lerner, Carbohydr. Res. 1975, 44, 116), but
we encountered problems. Therefore, we developed a new
synthetic protocol, which will be published in due course.
[23] G. Excoffier, D. Gagnaire, J.-P. Utille, Carbohydr. Res. 1975, 39,
368.
Received: July 19, 2007
Revised: August 30, 2007
Published online: November 21, 2007
[24] M. M. Sim, H. Kondo, C.-H. Wong, J. Am. Chem. Soc. 1993, 115,
2260.
[25] G. Baisch, R. ꢀhrlein, Bioorg. Med. Chem. 1997, 5, 383.
[26] The data is not shown here but will be published in due course.
Keywords: carbohydrates · glycoconjugates ·
glycophosphotransferases · glycosylation · nucleotides
.
1502
www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 1500 –1502