Scheme 3 Synthesis of GalNAcb1–3Gala1–3Galb1–4Glcb-5-azidopentamide 5 with in situ UDP-GlcNAc regeneration.
here the construction of LgtD/WbgU (UDP-GlcNAc C4
epimerase from P. shigelloides)8 fusion protein. WbgU cata-
lyzes the interconversion of UDP-GlcNAc to UDP-GalNAc;
meanwhile, LgtD catalyzes the introduction of GalNAc residue
to the acceptor substrates. The overall function of the re-
combinant protein is the transfer of the GlcNAc residue from
UDP-GlcNAc to the acceptors. Thus, the construction alters the
donor substrate requirement of the reaction from UDP-GalNAc
to UDP-GlcNAc This change is cost effective because the
commercial price of UDP-GlcNAc is 20 times less than that of
UDP-GalNAc. Moreover, it avoids the need to express and
purify two recombinant enzymes for reactions
transplantation. After two steps of purification, overall yields
from 60% to 85% were achieved in the reactions. It is a
significant improvement as compared to the chemical synthesis
of these tetrasaccharides, for which the total yields from
lactosides are normally below 40% due to multiple protection
and purification steps.4
The synthetic efficiency of the fusion protein was also tested
by coupling with a multiple-enzyme UDP-GlcNAc regenera-
tion cycle which consists of UDP-GlcNAc pyrophosphorylase
(GlmU), pyruvate kinase (PykF), and inorganic pyrophospha-
tase (PPase) from E. coli K12; GlcNAc phosphate mutase
(Agm1) from S. cerevisiae; and GlcNAc kinase (GlcNAcK)
from C. albicans (Scheme 3). This system further reduces the
cost of glycosylation reactions because it avoids the use of
expensive sugar nucleotides. An overall yield of 70% was
achieved in the reaction. We have previously used the accepter
4 in the synthesis of a-Gal conjugated polymers that sig-
nificantly inhibit the binding of human anti-Gal antibody to
mouse laminin glycoprotein and mammalian PK15 cells.9
Similarly, the oligosaccharide product 5 can be easily trans-
formed into other functional glycoconjugates that may serve as
tools in bacterial toxin inhibition, xenotransplantation and other
pharmaceutical studies.
The bifunctional LgtD-WbgU fusion enzyme was prepared
by in-frame linking of the P. shigelloides wbgU gene down-
stream to the H. influenzae lgtD gene through a five-residue
peptide linker (Thr-Gly-Gly-Gly-Gly). The C-terminus of the
fusion protein includes a 6 3 His tag for convenient purification
by nickel–nitrilotriacetic acid affinity chromatography (Fig. 1).
The enzyme was expressed in E. coli BL21(DE3) at a relatively
high level (22 U L21). The one-step purified enzyme has an
estimated molecular weight of 70 kDa and a specific activity of
1.68 U mg21
.
To investigate its acceptor specificity, the GalNAc transfer-
ase activity was tested with 12 carbohydrate compounds
including globotriose, Gala1,3Lac and some derivatives 1a–l.
As shown in Table 1, all these oligosaccharides are good
substrates though compounds with a terminal a(1?4)-linked
Gal-Gal structure are better acceptors than those with an
a(1?3) linkage (compare 1a with 1g, and 1f with 1l).
Apparently, the fusion protein has very broad acceptor substrate
specificity to the anomeric aglycon.
This work was supported by National Institutes of Health
Grant AI 44040.
Notes and references
1 M. J. Kuehn, J. Heuser, S. Normark and S. J. Hultgren, Nature, 1992, 356,
252–255.
2 M. A. Karmali, B. T. Steele, M. Petric and C. Lim, Lancet, 1983, 1,
619–620.
The fusion protein was then employed in coupled enzymatic
glycosylation reactions (100 mg scale). Globotetraose 3a and its
b-benzyl 3b, b-methyl 3c and b-azido 3d derivatives were
synthesized from the corresponding lactosides 2a–d in reactions
catalyzed sequentially by the a(1?4) galactosyltransferase
(LgtC) from N. meningitidis and the fusion enzyme (Scheme 1).
The UDP-Glc C4 epimerase (GalE) from E. coli K12 was also
used so that UDP-Glc and UDP-GlcNAc, instead of expensive
UDP-Gal and UDP-GalNAc, were supplied as donor substrates.
Compound 3d has a purposely-introduced azido group, which
makes it flexible in the solid-phase synthesis of glycopeptides
and glycopolymers. Meanwhile, isoglobotetraose 3e and its b-
benzyl 3f, b-methyl 3g and b-acetylamide 3h derivatives were
synthesized from the corresponding lactosides 2e–h in one-pot
reactions catalyzed by GalE from E. coli K12, a (1?3)
galactosyltransferase (a1,3GalT) from bovine and the LgtD-
WbgU fusion protein (Scheme 2). It should be noted that these
compounds also possess the a-Gal (Gala1,3Gal) structure,
which is believed to have potential applications in xeno-
3 P. R. Chipman, M. Agbandje-McKenna, S. Kajigaya, K. E. Brown, N. S.
Young, T. S. Baker and M. G. Rossman, Proc. Natl. Acad. Sci. USA,
1996, 93, 7502–7506.
4 (a) S. Bhattacharyya, B. G. Magnusson, U. Wellmar and U. J. Nilsson, J.
Chem. Soc., Perkin Trans. 1, 2001, 886–890; (b) U. Nilsson, A. K. Ray
and G. Magnusson, Carbohydr. Res., 1994, 252, 117–136; (c) H. Ishida,
R. Miyawaki, M. Kiso and A. Hasegawa, Carbohydr. Res., 1996, 284,
179–190.
5 K. M. Koeller and C. H. Wong, Chem. Rev., 2000, 100, 4465–4494.
6 (a) J. Fang, X. Chen, W. Zhang, A. Janczuk and P. G. Wang, Carbohydr.
Res., 2000, 329, 873–878; (b) O. Blixt, J. Brown, M. J. Schur, W.
Wakarchuk and J. C. Paulson, J. Org. Chem., 2001, 66, 2442–2448; (c)
M. Gilbert, R. Bayer, A. M. Cunningham, S. DeFrees, Y. Gao, D. C.
Watson, M. N. Young and W. W. Wakarchuk, Nat. Biotechnol., 1998, 16,
769–772.
7 J. Shao, J. Zhang, P. Kowal, Y. Lu and P. G. Wang, Biochem. Biophys.
Res. Commun., 2002, 295, 1–8.
8 P. Kowal and P. G. Wang, Biochemistry, 2002, 41, 5410–4.
9 J.-Q. Wang, X. Chen, W. Zhang, S. Zacharek, Y. Chen and P. G. Wang,
J. Am. Chem. Soc., 1999, 121, 8174–8181.
CHEM. COMMUN., 2003, 1422–1423
1423
Shao, Jun
