A very simple synthesis of GlcNAc-α-pyrophosphoryl-decanol: A substrate for the assay of a bacterial galactosyltransferase

Article abstract of DOI:10.1016/j.bmcl.2007.11.031

Lipid-linked sugar pyrophosphates, such as GlcNAc-pyrophosphoryl undecaprenol, are important intermediates in the biosynthesis of cell-surface bacterial polysaccharides. It was recently demonstrated that much simpler lipids could substitute for undecaprenol while retaining biological activity, thus making efficient synthetic access to this class of compounds highly desirable. In order to facilitate the synthesis of pure substrates for bacterial glycosyltransferases, we have developed a simple 'two-pot' synthesis which we demonstrate here for GlcNAc-α-pyrophosphoryl-decanol (4). GlcNAc pyrophosphate, produced by mild periodate oxidation/β-elimination of commercial UDP-GlcNAc, is alkylated using 1-iododecane to yield the target compound 4 in 39% yield. Compound 4 is shown to be an efficient acceptor for a bacterial galactosyltransferase.

Full text of DOI:10.1016/j.bmcl.2007.11.031

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry Letters 18 (2008) 804–807
A very simple synthesis of GlcNAc-a-pyrophosphoryl-decanol:
A substrate for the assay of a bacterial galactosyltransferase
Inka Brockhausen,a E. Andreas Larssonb and Ole Hindsgaulb,*
aDepartment of Medicine, Department of Biochemistry, Queens University, Kingston, Ont., Canada K7L 3N6
bCarlsberg Laboratory, Gamle Carlsberg Vej 10, DK-2500 Valby, Copenhagen, Denmark
Received 15 October 2007; revised 7 November 2007; accepted 9 November 2007
Available online 17 November 2007
Abstract—Lipid-linked sugar pyrophosphates, such as GlcNAc-pyrophosphoryl undecaprenol, are important intermediates in the
biosynthesis of cell-surface bacterial polysaccharides. It was recently demonstrated that much simpler lipids could substitute for
undecaprenol while retaining biological activity, thus making efficient synthetic access to this class of compounds highly desirable.
In order to facilitate the synthesis of pure substrates for bacterial glycosyltransferases, we have developed a simple ‘two-pot’ syn-
thesis which we demonstrate here for GlcNAc-a-pyrophosphoryl-decanol (4). GlcNAc pyrophosphate, produced by mild periodate
oxidation/b-elimination of commercial UDP-GlcNAc, is alkylated using 1-iododecane to yield the target compound 4 in 39% yield.
Compound 4 is shown to be an efficient acceptor for a bacterial galactosyltransferase.
ꢀ 2007 Elsevier Ltd. All rights reserved.
Sugar pyrophosphates are ubiquitous in nature where
they are used as building blocks for the assembly of oli-
gosaccharides and polysaccharides. The Leloir1 sugar
nucleotides, such as UDP-GlcNAc (1), have been the
most intensively studied as they are critical to the bio-
synthesis of mammalian glycoconjugates. Numerous
chemical syntheses have been reported for this class of
compounds where the key step has been the coupling
of two monophosphates to form the pyrophosphate
linkage. As an example, GlcNAc-1-phosphate and uri-
dine monophosphate (usually activated as either the
morpholidate2 or imidazolide3) are condensed to form
UDP-GlcNAc (Fig. 1). In a few cases,4 a nucleotide
pyrophosphate has been used as the nucleophile in the
displacement of an anomeric leaving group on the sugar.
Despite being laborious, many analogs have also been
prepared by these chemical methods.5 More recently,
the use of enzymes for both the anomeric phosphoryla-
tion of sugars and their condensation to form sugar
nucleotides has become increasingly applied.6
acceptors and donors. Prominent examples include the
oligosaccharyl dolicholpyrophosphates involved in the
biosynthesis of N-linked glycans7 and the sugar-pyro-
phosphoryl polyprenols involved in the assembly of bac-
terial cell-surface polysaccharides8.
Less synthetic attention has been paid to lipid-linked su-
gar pyrophosphates which can serve both as glycosyl
Keywords: Glycosyl acceptor; Galactosyltransferase; Sugar pyrophos-
phate; GlcNAc-pyrophosphoryl-lipid.
*
Figure 1. Traditional chemical synthesis of a sugar-pyrophosphoryl
derivative: UDP-GlcNAc as example.
Corresponding author. Tel.: +45 33 21 53 82; fax: +45 33 21 47
08; e-mail: hindsgaul@crc.dk
0960-894X/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.11.031

I. Brockhausen et al. / Bioorg. Med. Chem. Lett. 18 (2008) 804–807
805
Figure 2. Reported acceptors for a bacterial galactosyltransferase and the synthetic target 4.
Montoya-Peleaz et al.9 recently reported the identifica-
tion of a UDP-Gal: GlcNAc-R galactosyltransferase
activity in Escherichia coli that used GlcNAc-a-pyro-
phosphoryl-undecaprenol (2) as the natural acceptor
substrate for the second step in the synthesis of the O7
antigen repeat unit (Fig. 2). Since compound 2 is very
difficult to access synthetically and from natural sources,
they prepared compound 3 which was demonstrated to
be an excellent acceptor for the target galactosyltrans-
ferase (Gal-T) that transfers to the 3-OH group as
shown.10 The simple lipophilic undecane group was
therefore demonstrated to functionally substitute for
the much more complex branched undecaprenol. The
transfer reaction was quantitated using a radiochemical
assay measuring the incorporation of [3H]-labeled Gal,
from UDP-[3H]Gal, into lipophilic acceptor 3. The ter-
minal phenoxy group was included to permit UV-detec-
tion in HPLC. It was not reported if this group
contributed to the observed enzyme activity. Yi et al.11
later used the same synthetic strategy to prepare the cor-
responding GalNAc-a-pyrophosphoryl-lipid which was
shown to be an acceptor for an a-GalNAc-transferase
in a different E. coli strain. In their case, they found that
the phenyl group at the end of an aliphatic chain of 11
carbons was not essential for activity.
utilizing sugar-pyrophosphate-lipids as natural sub-
strates,8,11 including the functional characterization of
glycosyltransferases of many different bacterial strains
that have until now been described as ‘putative glycosyl-
transferase’ based solely on DNA and protein sequence
information.
Since Yi et al.11 found that the phenoxy group in 3 was
not essential for GalNAc-transferase activity, we
decided to devise a simple rapid synthesis of a Gal-trans-
ferase acceptor functionally equivalent to 3, that is, 4,
which would not involve either protecting group manip-
ulations or coupling of phosphates to give the pyrophos-
phate linkage. The synthesis developed is shown in
Scheme 2 and has two key steps. We reasoned that if
we could provide ready access to GlcNAc-a-pyrophos-
phate (6), which is not commercially available, it could
be alkylated using a long-chain alkyl halide (or tosylate)
as demonstrated for other pyrophosphates by Poulter.12
The final step in the synthesis would therefore be the
conversion of 6 to the target 4 (Scheme 2). There re-
mained to devise a simple preparation of 6 from a com-
mercial starting material and for this we took advantage
of the known preference of periodate to cleave cyclic cis-
diols over cyclic trans-diols.13 If UDP-GlcNAc (1) were
the starting material, limited periodate oxidation should
More GlcNAc-a-pyrophosphoryl-lipid acceptors were
required to continue the studies of Montoya-Peleaz
et al.,9 and we contemplated repeating the reported syn-
thesis of 3. Though very efficient (Scheme 1), this would
have required 11 reactions with multiple extensive puri-
fication steps without contributing new methodology to
facilitate the preparation of other potential substrates
with different sugars and different lipids. Such substrates
would be of value in the assay of other glycosyltransfer-
ases involved in bacterial polysaccharide synthesis and
therefore produce the dialdehyde intermediate 5
(Scheme 2) which, by analogy with other nucleosides
and nucleotides like ATP, should readily undergo b-
elimination14 to the desired 6.
UDP-GlcNAc (1, ca. 1 mg/USD, from Sigma) was re-
acted in the cold with less than one equivalent of perio-
date to produce in situ the dialdehyde 5 formed by
selective cleavage of the ribose ring. Addition of base
then led to a b-elimination to directly produce 6. After

806
I. Brockhausen et al. / Bioorg. Med. Chem. Lett. 18 (2008) 804–807
Scheme 1. Reported synthesis of 3, adapted from Ref. 9. Reagents: (1) BnBr, NaH, DMF; (2) Ac2O/Pyr; (3) H2, Pd/C; (4) iPr2NP(OBn)2, tetrazole;
i
(5) mCPBA; (6) H2, Pd/C; (7) Pr2NP(OBn)2, tetrazole; (8) mCPBA; (9) H2, Pd/C; (10) carbonyldiimidazole; (11) NaOMe/MeOH.
Scheme 2. ‘Two-pot’ synthesis of target compound 4.
ion exchange to produce the nucleophilically reactive
tris-tetrabutylammonium salt, reaction with 1-iodode-
cane gave the target 4, which was purified by ion ex-
change and reverse-phase adsorption/elution from
solid-phase extraction cartridges.15 The overall yield
from 1 to 4 was 39%. HPLC analysis of 4 using an ana-
lytical C18 column and acetonitrile/water (9:91) as the
mobile phase at 1 mL/min showed a single peak eluting
at 12 min with the expected mass (MꢀH+ = 520). That
the decyl and GlcNAc groups were attached to different
phosphate groups was evident from the two-dimensional
1H/31P correlations in NMR.15
shows that the decyl chain as a lipid moiety in the accep-
tor is sufficient for sugar transfer, and that a phenyl
group may enhance activity but is not essential.
In contrast, 3 was a poor substrate for bovine b4Gal-
transferase. At 0.5 mM concentration in the assay, 3
showed only 3% of the activity obtained with 0.5 mM
GlcNAcb-Bn substrate. This indicates that the bacterial
and mammalian Gal-transferases differ significantly in
binding to their acceptor substrates although they both
bind GlcNAc-R and UDP-Gal, and require Mn2+ as a
cofactor. Thus, in contrast to the bacterial Gal-transfer-
ase, the mammalian enzyme cannot easily accommodate
the pyrophosphate group in the acceptor binding site
and has a distinct specificity for GlcNAc in b-
configuration.17
The activity of 4 as an acceptor substrate for WbbD,
UDP-Gal: GlcNAc-R b3-galactosyltransferase, activity
in E. coli VW18710 was compared with that of known
3. The new compound 4 was a very good acceptor for
WbbD Gal-transferase.16 At 0.5 mM concentration in
the assay, an activity of 115 nmol/h/mg was obtained,
corresponding to 70% of the activity obtained with 3.
The nano-electrospray-MS (negative ion mode) showed
m/z of 520 for 4 and m/z of 682 for the enzyme product
of WbbD, indicating the addition of Gal to 4. This
In summary, we have reported a very simple synthesis of
target compound 4 which was demonstrated to be an
acceptor for a bacterial galactosyltransferase. The pro-
cedure involved the facile generation of GlcNAc-pyro-
phosphate 6 from UDP-GlcNAc and its subsequent
alkylation by 1-iododecane. Since a large number of pri-

I. Brockhausen et al. / Bioorg. Med. Chem. Lett. 18 (2008) 804–807
807
11. Yi, W.; Yao, Q.; Zhang, Y.; Motari, E.; Lin, S.; Wang, P.
G. Biochem. Biophys. Res. Commun. 2006, 344, 631.
12. Davisson, V. J.; Davis, D. R.; Dixit, V. M.; Poulter, C. D.
J. Org. Chem. 1987, 52, 1794.
13. Perlin, A. S. Adv. Carbohydr. Chem. Biochem 2006, 60,
183.
14. (a) Steinschneider, A. Biochemistry 1971, 10, 173; (b)
Grant, A. J.; Lerner, L. M. J. Med. Chem. 1980, 23, 795;
(c) Ho, H.-T.; Frey, P. A. Biochemistry 1984, 23, 1978; (d)
Muthukumaran, T.; KrishnaMurthy, N. V.; Sudhaharan,
T.; Muralidharan, B. J. Appl. Rad. Isotopes 2005, 63, 63.
15. To a solution of UDP-GlcNAc di-sodium salt (1, 50 mg,
77 lmol), at 4 ꢁC, was added NaIO4 (402 lL of a 0.187 M
solution, 0.98 equiv) and the reaction tube was vortexed
for 1 min and then placed at 4 ꢁC for 1 h. The reaction was
quenched by the addition of glycerol (20 lL). After
15 min, 1 M aq NaOH (80 lL) was added, the sample
vortexed for 1 min, and placed at 4 ꢁC for 24 h.
mary alkyl halides and tosylates (or their precursor alco-
hols), as well as nucleotide sugars, are commercially
available, this method should provide rapid access to a
panel of lipid-linked sugar pyrophosphates for the func-
tional characterization of bacterial glycosyltransferases.
The key step of the synthesis relies on the facile perio-
date oxidation of the cis-diol of a ribofuranose over
the trans-diol of a glucopyranose. Beyond this selectiv-
ity, we have not yet defined the scope and limitations
of the method, in particular whether it can be extended
to sugar nucleotides containing pyranoses with cis-diols,
such as UDP-Gal and GDP-Man.
Acknowledgment
I.B. acknowledges support from the Canadian Cystic
Fibrosis Foundation.
The reaction mixture was diluted to 10 mL with cold H2O
and applied to a column of Dowex-H+ (4 · 2 cm id) and
eluted with cold H2O (80 mL) into a beaker containing
1 M tetrabutylammonium hydroxide (305 lL) in H2O
(10 mL). The pH of the resulting solution was found to be
ca. 6–7 and was increased up to ca.7–8 by addition of 1 M
tetrabutylammonium hydroxide (40 lL). The solution was
lyophilized, re-dissolved in H2O (1.5 mL), transferred to a
microfuge tube, and re-lyophilized. The brownish oily
residue was again re-dissolved in dry acetonitrile (300 lL)
and split into three 100 lL portions.
References and notes
1. (a) Cardini, C. E.; Paladini, A. C.; Caputto, R.; Leloir, L.
F. Nature 1950, 165, 191; (b) Leloir, L. F. Science 1971,
172, 1299.
2. (a) Roseman, S.; Distler, J. J.; Moffatt, J. G.; Khorana, H.
G. J. Am. Chem. Soc. 1961, 83, 659; (b) Whittmann, V.;
Wong, C.-H. J. Org. Chem. 1997, 62, 2144.
To one of the 100 lL portions was added 1-iododecane,
(13.6 lL, 64 lmol, 2.5 equiv) and the reaction tube was
vortexed for 1 min and placed on a shaker-table. After
48 h the reaction was diluted with H2O (1 mL), theþn
loaded onto a 10 · 0.5 cm column of Dowex-50 (NH4
form). The column was washed with H2O. A total of
50 mL was collected, and analyzed by ESI-MS which
showed 4 as the major product. In addition, the mass
spectra verified that all tetrabutylammonium salts had
been removed which otherwise hampered the following
separation. The water fraction was lyophilized, re-dis-
solved in H2O (3 mL), and added onto 2 preactivated C18
Sep-Pak cartridges connected in series. After the pass-
through, the cartridges were washed and fractions
(1.5 mL) were collected: H2O · 7, 5% aq MeOH · 7, 10%
aq MeOH · 6. The fractions were analyzed by ESI-MS
and the pure product was present from fraction 4 of the
H2O washes to fraction 3 of 10% MeOH washes, which
were pooled and concentrated. The residue was re-
dissolved in H2O (1.5 mL) and lyophilized in a microfuge
tube to yield 4 (5.18 mg, 9.93 lmol, 39%). ESI-MS;
[M ꢀ H+]ꢀ: expected: 520.26 found: 520.0; selected
3. (a) Staab, H. A.; Schaller, H.; Cramer, F. Angew. Chem.
1959, 71, 736; (b) Barker, R.; Olsen, K. W.; Shaper, J. H.;
Hill, R. L. J. Biol. Chem. 1972, 247, 7135; (c) Chen, L.;
Men, H.; Ha, S.; Ye, X.-Y.; Brunner, L.; Hu, Y.; Walker,
S. Biochemistry 2002, 41, 6824.
4. (a) Salam, M. A.; Behrman, E. J. Carbohydr. Res. 1982,
101, 339; (b) Arlt, M.; Hindsgaul, O. J. Org. Chem. 1995,
60, 14; (c) Uchiyama, T.; Hindsgaul, O. J. Carbohydr.
Chem. 1998, 17, 1181; (d) Hanessian, S.; Lu, P.-P.; Ishida,
H. J. Am. Chem. Soc. 1998, 120, 13296; (e) Sujino, K.;
Uchiyama, T.; Hindsgaul, O.; Seto, N. O. L.; Wakarchuk,
W.; Palcic, M. M. J. Am. Chem. Soc. 2000, 122, 1261; (f)
Ernst, C.; Klaffke, W. J. Org. Chem. 2003, 68, 5780; (g)
Timmons, S. C.; Jakeman, D. L. Org. Lett. 2007, 9, 1227.
5. (a) Qian, X.; Sujino, K.; Palcic, M. M.; Ratcliffe, R. M. J.
Carbohydr. Chem. 2002, 21, 911; (b) Khaled, A.; Ivannik-
´
ova, A.; Auge, C. Carbohydr. Res. 2004, 339, 2641.
6. (a) Ichikawa, Y.; Look, G. C.; Wong, C.-H. Anal.
Biochem. 1992, 202, 215; (b) Mizanur, R. M.; Zea, C. J.;
Pohl, N. L. J. Am. Chem. Soc. 2004, 126, 15993; (c) Errey,
J. C.; Mukhopadhyay, B.; Kartha, K. P. R.; Field, R. A.
1
NMR data supporting the structural assignment of 4: H
Chem. Commun. 2004, 2706; (d) Bulter, T.; Elling, L.
Glycoconj. J. 1999, 16, 147.
¨
NMR
(250 MHz,
D2O)
d
5.45
(H1,
dd,
J1,2 = 3.2,J1,P = 7.2 Hz), 2.04 (s, NCOCH3), 0.82 (t, dec-
ane-CH3); 13C: d 94.5 (C1): 31P: d ꢀ10.7 (P2, P-O-decane),
ꢀ13.2 (P1, GlcNAc-O-P); 1H/31P HMBC: 5.45/ꢀ13.2
(GlcNAc-H1/P2), 3.94/ꢀ13.2 (GlcNAc-H2/P2), 3.89/ꢀ10.7
(decane-CH2/P2).
7. (a) Brockhausen, I.; Schutzbach, J.; Kuhns, W. Acta Anat.
1998, 161, 36; (b) Schenk, B.; Fernandez, F.; Waechter, C.
J. Glycobiol. 2001, 11, 61R; (c) Helenius, J.; Aebi, M.
Semin. Cell Dev. Biol. 2002, 13, 171; (d) Lehle, L.; Strahl,
S.; Tanner, W. Angew. Chem., Int. Ed. 2006, 45, 6802.
8. (a) Raetz, R. H.; Whitfield, C. Ann. Rev. Biochem. 2002,
71, 635; (b) Guan, S.; Clarke, A. J.; Whitfield, C. J.
Bacteriol. 2001, 183, 3318; (c) Price, N. P.; Momany, F. A.
Glycobiology 2005, 15, 29R.
16. Bacterial cultures of Escherichia coli VW187 (O7:K1)
were grown and used as the enzyme source in Gal-
transferase assays as described.9,10 Enzyme products
using acceptor 3 or newly synthesized acceptor 4 were
isolated by C18 Sep-Pak columns. Mammalian b4-Gal-
transferase I (lactose synthase, Sigma) was assayed as
described.17
9. Montoya-Peleaz, P. J.; Riley, J. G.; Szarek, W. A.;
Valvano, M. A.; Schutzbach, J. S.; Brockhausen, I.
Bioorg. Med. Chem. Lett. 2005, 15, 1205.
17. Brockhausen, I.; Benn, M.; Bhat, S.; Marone, S.; Riley, J.
G.; Montoya-Peleaz, P.; Vlahakis, J. Z.; Paulsen, H.;
Schutzbach, J. S.; Szarek, W. Glycoconj. J. 2006, 23, 523.
10. Riley, J. G.; Menggad, M.; Montoya-Peleaz, P.; Szarek,
W. A.; Marolda, C. L.; Valvano, M. A.; Schutzbach, J. S.;
Brockhausen, I. Glycoconj. J. 2006, 23, 525.