Chemoenzymatic synthesis and lectin array characterization of a class of N-glycan clusters

Article abstract of DOI:10.1021/ja9078539

N-Glycans are major components of many glycoproteins. These sugar moieties are frequently involved in important physiological and disease processes via their interactions with a variety of glycanbinding proteins (GBP). Clustering effect is an important feature in many glycan-lectin interactions. We describe in this paper a chemoenzymatic synthesis of novel N-glycan clusters using a tandem endoglycosidase-catalyzed transglycosylation. It was found that the internal β-1,2-linked GlcNAc moieties in the N-glycan core, once exposed in the nonreducing terminus, was able to serve as acceptors for transglycosylation catalyzed by Endo-A and EndoM-N175A. This efficient chemoenzymatic method allows a quick extension of the sugar chains to form a class of glycan clusters in which sugar residues are all connected by native glycosidic linkages found in natural N-glycans. In addition, a discriminative enzymatic reaction at the two GlcNAc residues could be fulfilled to afford novel hybrid clusters. Lectin microarray studies revealed unusual properties in glyco-epitope expression by this panel of structurally well-defined synthetic N-glycans. These new compounds are likely valuable for functional glycomics studies to unveil new functions of both glycans and carbohydrate-binding proteins.

Full text of DOI:10.1021/ja9078539

Published on Web 11/16/2009
Chemoenzymatic Synthesis and Lectin Array Characterization
of a Class of N-Glycan Clusters
†
‡
§
,†
Wei Huang, Denong Wang, Masao Yamada, and Lai-Xi Wang*
Institute of Human Virology and Department of Biochemistry and Molecular Biology, UniVersity
of Maryland School of Medicine, Baltimore, Maryland 21201, Stanford Tumor Glycome
Laboratory, Department of Genetics, Stanford UniVersity School of Medicine,
Stanford, California 94305, and GP Biosciences, Ltd., Yokohama 225-0012, Japan
Received September 16, 2009; E-mail: LWang@som.umaryland.edu
Abstract: N-Glycans are major components of many glycoproteins. These sugar moieties are frequently
involved in important physiological and disease processes via their interactions with a variety of glycan-
binding proteins (GBP). Clustering effect is an important feature in many glycan-lectin interactions. We
describe in this paper a chemoenzymatic synthesis of novel N-glycan clusters using a tandem endogly-
cosidase-catalyzed transglycosylation. It was found that the internal ꢀ-1,2-linked GlcNAc moieties in the
N-glycan core, once exposed in the nonreducing terminus, was able to serve as acceptors for
transglycosylation catalyzed by Endo-A and EndoM-N175A. This efficient chemoenzymatic method allows
a quick extension of the sugar chains to form a class of glycan clusters in which sugar residues are all
connected by native glycosidic linkages found in natural N-glycans. In addition, a discriminative enzymatic
reaction at the two GlcNAc residues could be fulfilled to afford novel hybrid clusters. Lectin microarray
studies revealed unusual properties in glyco-epitope expression by this panel of structurally well-defined
synthetic N-glycans. These new compounds are likely valuable for functional glycomics studies to unveil
new functions of both glycans and carbohydrate-binding proteins.
Introduction
provide the basis for molecular recognition, but the way of their
presentation, e.g., the monovalent vs multivalent format, is also
N-linked glycosylation is a predominant covalent modification
of proteins in eukaryotes. It is well documented that N-glycans
of glycoproteins are involved in many important biological
events including protein folding, ER-associate protein degrada-
tion, cell differentiation, cell adhesion, host-pathogen interaction,
of paramount importance in governing the specificity and
3
strength in carbohydrate-protein interactions. Recent advances
in glycan and glycan-binding protein microarray technology
have offered exciting new opportunities to unveil the mysteries
4
of glycans in various cellular processes and disease states.
1
cancer metastasis, and autoimmunity. Glycoproteins are often
Nevertheless, functional glycomics studies are still limited by
the availability of structurally well-defined N-glycans and related
glycoconjugates, which are difficult to obtain in homogeneous
characterized by their structural microheterogeneity in terms of
the components of the attached glycans. It becomes clear that
distinct N-glycans can confer significantly different effects on
the structure and function of a given glycoprotein. This was
exemplified by recent discoveries that the attachment of subtly
different N-glycans at the conserved N-glycosylation site of the
Fc domain could result in dramatically different impacts on the
ADCC function of monoclonal antibodies and on the anti2-
inflammatory activity of intravenous immunoglobulin (IVIG).
On the other hand, not only does the fine structure of the glycans
5
forms from natural source.
We and others have previously demonstrated that a class of
endo-ꢀ-N-acetylglucosaminidases (ENGases), including the
(
3) (a) Lee, R. T.; Lee, Y. C. Glycoconjugate J. 2000, 17, 543–551. (b)
Lundquist, J. J.; Toone, E. J. Chem. ReV. 2002, 102, 555–578. (c)
Brewer, C. F.; Miceli, M. C.; Baum, L. G. Curr. Opin. Struct. Biol.
2002, 12, 616–623.
(
4) (a) Blixt, O.; et al. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 17033–
1
7038. (b) Feizi, T.; Chai, W. Nat. ReV. Mol. Cell Biol. 2004, 5, 582–
†
University of Maryland School of Medicine.
Stanford University School of Medicine.
GP Biosciences, Ltd.
588. (c) Xia, B.; Kawar, Z. S.; Ju, T.; Alvarez, R. A.; Sachdev, G. P.;
Cummings, R. D. Nat. Methods 2005, 2, 845–850. (d) Oyelaran, O.;
Gildersleeve, J. C. Expert ReV. Vaccines 2007, 6, 957–969. (e)
Pilobello, K. T.; Mahal, L. K. Curr. Opin. Chem. Biol. 2007, 11, 300–
305. (f) Hirabayashi, J. J. Biochem. 2008, 144, 139–147. (g) Zhao, J.;
Patwa, T. H.; Lubman, D. M.; Simeone, D. M. Curr. Opin. Mol. Ther.
2008, 10, 602–610. (h) Liang, P. H.; Wu, C. Y.; Greenberg, W. A.;
Wong, C. H. Curr. Opin. Chem. Biol. 2008, 12, 86–92. (i) Hsu, K. L.;
Mahal, L. K. Curr. Opin. Chem. Biol. 2009, 13, 427–432. (j) Oyelaran,
O.; Gildersleeve, J. C. Curr. Opin. Chem. Biol. 2009, 13, 406–413.
(k) Song, X.; Xia, B.; Stowell, S. R.; Lasanajak, Y.; Smith, D. F.;
Cummings, R. D. Chem. Biol. 2009, 16, 36–47.
‡
§
(
1) (a) Varki, A. Glycobiology 1993, 3, 97–130. Dwek, R. A. Chem. ReV.
1
2
2
996, 96, 683–720. (b) Helenius, A.; Aebi, M. Science 2001, 291,
364–2369. (c) Haltiwanger, R. S.; Lowe, J. B. Annu. ReV. Biochem.
004, 73, 491–537. (d) Dube, D. H.; Bertozzi, C. R. Nat. ReV. Drug
DiscoV. 2005, 4, 477–488. (e) Arnold, J. N.; Wormald, M. R.; Sim,
R. B.; Rudd, P. M.; Dwek, R. A. Annu. ReV. Immunol. 2007, 25, 21–
5
1
0. (f) Gamblin, D. P.; Scanlan, E. M.; Davis, B. G. Chem. ReV. 2009,
09, 131–163.
(
2) (a) Jefferis, R. Biotechnol. Prog. 2005, 21, 11–16. (b) Kaneko, Y.;
Nimmerjahn, F.; Ravetch, J. V. Science 2006, 313, 670–673. (c)
Anthony, R. M.; Nimmerjahn, F.; Ashline, D. J.; Reinhold, V. N.;
Paulson, J. C.; Ravetch, J. V. Science 2008, 320, 373–376.
(5) (a) Raman, R.; Raguram, S.; Venkataraman, G.; Paulson, J. C.;
Sasisekharan, R. Nat. Methods 2005, 2, 817–824. (b) Paulson, J. C.;
Blixt, O.; Collins, B. E. Nat. Chem. Biol. 2006, 2, 238–248.
10.1021/ja9078539  2009 American Chemical Society
J. AM. CHEM. SOC. 2009, 131, 17963–17971 9 17963

A R T I C L E S
Huang et al.
Endo-A from Arthrobactor protophormiae and the Endo-M from
Mucor hiemali, were able to transfer an oligosaccharide en bloc
from either natural N-glycans or synthetic sugar oxazolines to
a GlcNAc-containing moiety to form a new ꢀ-1,4-glycosidic
linkage in a regio- and stereospecific manner, leading to the
synthesis of complex oligosaccharides, N-glycopeptides and
Scheme 1
6
N-glycoproteins. Endo-A is specific for high-mannose or
hybrid-type N-glycans and has been applied for the synthesis
7
of high-mannose-type oligosaccharides and N-glycopeptides,
whereas Endo-M is able to work on three major types (high-
mannose, hybrid, and complex type) of N-glycans and has been
particularly useful for synthesizing complex type N-glycopep-
8
tides. In particular, the recent findings that synthetic oligosac-
charide oxazolines (the mimics of the oxazolinium ion inter-
mediate of the enzymatic reaction) could be used for Endo-A-
catalyzed transglycosylation have significantly expanded the
scope of the chemoenzymatic method for glycopeptide and
refractory to enzymatic hydrolysis due to the slight structural
modification. Moreover, the discovery of several ENGase-based
glycosynthases, including EndoM-N175A; EndoA-N171A, and
EndoA-E173Q that could promote transglycosylation with sugar
oxazolines of natural N-glycans but lack the ability to hydrolyze
the product, has enabled the synthesis of homogeneous glyco-
9-11
glycoprotein synthesis.
It was found that the highly activated
sugar oxazolines corresponding to the truncated or modified
N-glycans could serve as substrates for the Endo-A-catalyzed
transglycosylation, but the ground-state products formed were
1
2-14
proteins carrying full-size natural N-glycans.
Subsequent
(
(
6) (a) Yamamoto, K. J. Biosci. Bioeng. 2001, 92, 493–501. (b) Wang,
L. X. Carbohydr. Res. 2008, 343, 1509–1522.
studies indicate that Endo-A and Endo-M could accommodate
diverse structures in the aglycon portions of GlcNAc- or Glc-
tagged acceptors for transglycosylation, permitting the introduc-
7) (a) Takegawa, K.; Yamaguchi, S.; Kondo, A.; Kato, I.; Iwahara, S.
Biochem. Int. 1991, 25, 829–835. (b) Takegawa, K.; Yamaguchi, S.;
Kondo, A.; Iwamoto, H.; Nakoshi, M.; Kato, I.; Iwahara, S. Biochem.
Int. 1991, 24, 849–855. (c) Fan, J. Q.; Takegawa, K.; Iwahara, S.;
Kondo, A.; Kato, I.; Abeygunawardana, C.; Lee, Y. C. J. Biol. Chem.
tion of N-glycans into a wide range of natural products, unnatural
15,16
peptides, and even polysaccharides.
The relaxed substrate
specificity of Endo-A and Endo-M, together with the powerful
transglycosylation potential of the glycosynthase mutants,
prompted us to examine the possibility to glue multiple
N-glycans to the complex-type GlcNAc2Man3GlcNAc2-Asn
core through tandem enzymatic transglycosylation. We report
in this contribution the chemoenzymatic synthesis of a class of
novel N-glycan clusters containing multiple N-glycan cores, in
which all monosaccharide residues are connected via defined
native glycosidic bonds found in natural N-glycans. Lectin
microarray analysis of the synthetic N-glycan clusters has
revealed unusual lectin-carbohydrate recognition patterns that
were not observed before.
1
995, 270, 17723–17729. (d) Takegawa, K.; Tabuchi, M.; Yamaguchi,
S.; Kondo, A.; Kato, I.; Iwahara, S. J. Biol. Chem. 1995, 270, 3094–
3
099. (e) Wang, L. X.; Fan, J. Q.; Lee, Y. C. Tetrahedron Lett. 1996,
7, 1975–1978. (f) Wang, L. X.; Tang, M.; Suzuki, T.; Kitajima, K.;
3
Inoue, Y.; Inoue, S.; Fan, J. Q.; Lee, Y. C. J. Am. Chem. Soc. 1997,
19, 11137–11146. (g) Deras, I. L.; Takegawa, K.; Kondo, A.; Kato,
1
I.; Lee, Y. C. Bioorg. Med. Chem. Lett. 1998, 8, 1763–1766. (h) Fujita,
K.; Takegawa, K. Biochem. Biophys. Res. Commun. 2001, 282, 678–
6
82. (i) Fujita, K.; Miyamura, T.; Sano, M.; Kato, I.; Takegawa, K.
J. Biosci. Bioeng. 2002, 93, 614–617. (j) Singh, S.; Ni, J.; Wang, L. X.
Bioorg. Med. Chem. Lett. 2003, 13, 327–330. (k) Li, H.; Singh, S.;
Zeng, Y.; Song, H.; Wang, L. X. Bioorg. Med. Chem. Lett. 2005, 15,
8
95–898. (l) Wang, L. X.; Song, H.; Liu, S.; Lu, H.; Jiang, S.; Ni, J.;
Li, H. ChemBioChem 2005, 6, 1068–1074.
(
8) (a) Yamamoto, K.; Kadowaki, S.; Watanabe, J.; Kumagai, H. Biochem.
Biophys. Res. Commun. 1994, 203, 244–252. (b) Haneda, K.; Inazu,
T.; Yamamoto, K.; Kumagai, H.; Nakahara, Y.; Kobata, A. Carbohydr.
Res. 1996, 292, 61–70. (c) Haneda, K.; Inazu, T.; Mizuno, M.; Iguchi,
R.; Yamamoto, K.; Kumagai, H.; Aimoto, S.; Suzuki, H.; Noda, T.
Bioorg. Med. Chem. Lett. 1998, 8, 1303–1306. (d) Mizuno, M.;
Haneda, K.; Iguchi, R.; Muramoto, I.; Kawakami, T.; Aimoto, S.;
Yamamoto, K.; Inazu, T. J. Am. Chem. Soc. 1999, 121, 284–290. (e)
Mori, T.; Sekine, Y.; Yamamoto, K.; Okahata, Y. Chem. Commun.
Results and Discussions
Enzymatic Transglycosylation onto the ꢀ-1,2-Linked GlcNAc
Residues in the Asn-Linked GlcNAc2Man3GlcNAc2 Core.
Construction of an array of N-glycan clusters started with the
preparation of a biotinylated biantennary complex type N-glycan
(
Cambodia) 2004, 2692–2693. (f) Yamanoi, T.; Yoshida, N.; Oda,
(Scheme 1). Coupling of glycosylamine 2 and an activated biotin
Y.; Akaike, E.; Tsutsumida, M.; Kobayashi, N.; Osumi, K.; Yamamoto,
tag (3) gave the GlcNAc-LC-biotin (4), which was then used
as an acceptor for enzymatic transglycosylation. We have
previously reported the synthesis of an asialoglycan oxazoline
K.; Fujita, K.; Takahashi, K.; Hattori, K. Bioorg. Med. Chem. Lett.
2005, 15, 1009–1013. (g) Fujita, K.; Yamamoto, K. Biochim. Biophys.
Acta 2006, 1760, 1631–1635. (h) Makimura, Y.; Watanabe, S.; Suzuki,
T.; Suzuki, Y.; Ishida, H.; Kiso, M.; Katayama, T.; Kumagai, H.;
Yamamoto, K. Carbohydr. Res. 2006, 341, 1803–1808. (i) Haneda,
K.; Takeuchi, M.; Tagashira, M.; Inazu, T.; Toma, K.; Isogai, Y.; Hori,
M.; Kobayashi, K.; Takegawa, K.; Yamamoto, K. Carbohydr. Res.
(11) (a) Li, B.; Song, H.; Hauser, S.; Wang, L. X. Org. Lett. 2006, 8, 3081–
3084. (b) Zeng, Y.; Wang, J.; Li, B.; Hauser, S.; Li, H.; Wang, L. X.
Chem.sEur. J. 2006, 12, 3355–3364. (c) Ochiai, H.; Huang, W.;
Wang, L. X. J. Am. Chem. Soc. 2008, 130, 13790–13803.
(12) Umekawa, M.; Huang, W.; Li, B.; Fujita, K.; Ashida, H.; Wang, L. X.;
Yamamoto, K. J. Biol. Chem. 2008, 283, 4469–4479.
2
006, 341, 181–190.
(
9) (a) Fujita, M.; Shoda, S.; Haneda, K.; Inazu, T.; Takegawa, K.;
Yamamoto, K. Biochim. Biophys. Acta 2001, 1528, 9–14. (b) Li, H.;
Li, B.; Song, H.; Breydo, L.; Baskakov, I. V.; Wang, L. X. J. Org.
Chem. 2005, 70, 9990–9996. (c) Wei, Y.; Li, C.; Huang, W.; Li, B.;
Strome, S.; Wang, L. X. Biochemistry 2008, 47, 10294–10304. (d)
Rising, T. W.; Claridge, T. D.; Davies, N.; Gamblin, D. P.; Moir,
J. W.; Fairbanks, A. J. Carbohydr. Res. 2006, 341, 1574–1596. (e)
Rising, T. W.; Claridge, T. D.; Moir, J. W.; Fairbanks, A. J.
ChemBioChem 2006, 7, 1177–1180. (f) Rising, T. W.; Heidecke, C. D.;
Moir, J. W.; Ling, Z.; Fairbanks, A. J. Chem.sEur. J. 2008, 14, 6444–
(13) Huang, W.; Li, C.; Li, B.; Umekawa, M.; Yamamoto, K.; Zhang, X.;
Wang, L. X. J. Am. Chem. Soc. 2009, 131, 2214–2223.
(14) Heidecke, C. D.; Ling, Z.; Bruce, N. C.; Moir, J. W.; Parsons, T. B.;
Fairbanks, A. J. ChemBioChem 2008, 9, 2045–2051.
(15) Huang, W.; Ochiai, H.; Zhang, X.; Wang, L. X. Carbohydr. Res. 2008,
343, 2903–2913.
(16) (a) Huang, W.; Groothuys, S.; Heredia, A.; Kuijpers, B. H.; Rutjes,
F. P.; van Delft, F. L.; Wang, L. X. ChemBioChem 2009, 10, 1234–
1242. (b) Ochiai, H.; Huang, W.; Wang, L. X. Carbohydr. Res. 2009,
344, 592–598.
6
464.
(
10) Li, B.; Zeng, Y.; Hauser, S.; Song, H.; Wang, L. X. J. Am. Chem.
Soc. 2005, 127, 9692–9693.
1
7964 J. AM. CHEM. SOC. 9 VOL. 131, NO. 49, 2009

Chemoenzymatic Synthesis of N-Glycan Clusters
A R T I C L E S
Scheme 2
(
5) and have shown that it was a good substrate of the
3032.14; found M + Na ) 3055.12). The newly formed
glycosidic bonds in the Man3GlcNAc2 cores were assumed to
be in the natural ꢀ-1,4-glycosidic linkage between the two
GlcNAc residues on the basis of all previously reported
glycosynthase mutant EndoM-N175A for glycoprotein synthe-
13
sis. Thus, the reaction of oxazoline (5) and acceptor (4) (donor/
acceptor, 3:1) under the catalysis of EndoM-N175A gave the
biotinylated N-glycan product (6) in 72% yield. The two terminal
galactose residues were then selectively removed by treatment
with ꢀ-1,4-galactosidase from Bacteroides fragilis to provide
10,11,15
stereospecificity of Endo-A-catalyzed transglycosylation.
To further substantiate this assumption, compound 9 was treated
with wild-type Endo-M that specifically hydrolyzes the ꢀ-1,4-
glycosidic bond in natural N-glycan. It was found that treatment
2 3 2
GlcNAc Man GlcNAc -LC-biotin (7), in which the two internal
GlcNAc residues were exposed at the nonreducing terminus.
Enzymatic extension of the sugar chain from the two GlcNAc
residues would enable the synthesis of novel glycan structures.
The advantage of introducing a biotin tag in the aglycon portion
of the glycan primer (7) was apparent: it would facilitate the
detection of the synthetic N-glycans bound in a glycan-binding
protein microarray platform, or the biotinylated glycans could
be directly immobilized on a streptavidin plate to provide a
glycan array platform.
of 9 with Endo-M gave GlcNAc-LC-biotin (1 equiv), GlcNAc
2
-
Man GlcNAc (1 equiv) and Man3GlcNAc (2 equiv) (data not
3
shown), suggesting that the three Man3GlcNAc2 cores all have
the natural GlcNAc-ꢀ-1,4-GlcNAc linkage in the core. The
successful simultaneous enzymatic transfer of two Man3GlcNAc
moieties to the GlcNAc2Man3GlcNAc2-LC-biotin indicates that
the two terminal ꢀ-1,2-linked GlcNAc residues, albeit in a rigid
and crowded environment on the N-glycan core, are accessible
to Endo-A-catalyzed glycosylation.
To test whether the two terminal ꢀ-1,2-linked GlcNAc
residues of 7 are able to serve as acceptors for the next round
of enzymatic transglycosylation, we first examined the Endo-
Next, we examined the glycosylation of 7 by the glycosyn-
thase mutant EndoM-N175A, using a large complex type glycan
oxazoline (5) as the donor substrate. Interestingly, when a
limited amount of oxazoline donor substrate was used, a
selective glycosylation on the two terminal GlcNAc residues
was observed. Thus, incubation of oxazoline 5 and acceptor 7
(donor/acceptor, 3:1) with EndoM-N175A at 30 °C for 4 h gave
three products: the two monoglycosylated products (10a and
10b) and the doubly transglycosylated product 10c in 25, 5,
and 8% yields, respectively (Scheme 2). The GlcNAc linked to
the mannose on the R-1,6-arm was much more favorable for
transglycosylation than the GlcNAc linked to the mannose at
the R-1,3-arm, with a selectivity of 5:1 (25% vs 5%). Although
both GlcNAc are ꢀ-1,2-linked to the mannose moiety, the
3
A-catalyzed transglycosylation of 7 using Man GlcNAc oxazo-
1
0
line (8) as the donor substrate (Scheme 2). It was found that
when oxazoline 8 and GlcNAc2Man3GlcNAc2-LC-biotin (7)
(
donor/acceptor, 6:1) were incubated with Endo-A in a phos-
phate buffer at 30 °C, the transglycosylation reaction proceeded
very efficiently, and after 2 h, almost all the starting material 7
was converted to the doubly glycosylated product 9 (as revealed
by HPLC). The product was readily isolated by HPLC in 88%
yield. MALDI-TOF MS analysis showed a single m/z species
at 3055.12, which matches well with the theoretical data of 9
that carries three Man3GlcNAc2 cores (calculated, M )
J. AM. CHEM. SOC. 9 VOL. 131, NO. 49, 2009 17965

A R T I C L E S
Huang et al.
regioselectivity might result from the difference in steric
hindrance, as the GlcNAc at the 6-arm would be spatially less
hindered than the GlcNAc located on the R-1,3-arm. Neverthe-
less, double glycosylation could be achieved by further chemoen-
zymatic transglycosylation with excess amount of donor sub-
strate for a longer incubation time. This was exemplified by
further glycosylation of the purified monoglycosylated glycan
Scheme 3
1
0a and oxazoline 5 (3.5 mol equiv) under the catalysis of
EndoM-N175A, giving the doubly glycosylated product (10c)
in 50% yield. When Man GlcNAc oxazoline (11) was used as
9
the donor substrate, an even higher regioselectivity in transg-
lycosylation was observed. Incubation of 11 and 7 (3:1) with
EndoA-N175A for 4 h gave 27% of 12a, 3% of 12b, and trace
amount of the doubly glycosylated product (12a:12b ) 9:1)
(
Scheme 2). The higher selectivity might be explained by the
bulky structure of triantennary glycan (11) in comparison with
the biantennary complex glycan 5. It should be mentioned that
the use of wild-type Endo-M for the reaction between acceptor
Scheme 4
7
and the oxazoline 5 or 11 failed to provide the transglyco-
sylation products due to the enzymatic hydrolysis of the acceptor
7
as well as the products that were formed (data not shown).
It should be mentioned that in the synthesis of large complex
N-glycan clusters, most of the transglycosylation reactions were
carried out on a relatively small scale. Nevertheless, the specific
enzymatic transglycosylation usually gave a very clear HPLC
profile, allowing easy quantification and isolation of the reaction
products by HPLC. To characterize the transglycosylation
products, the new compounds were purified by HPLC and were
subjected to MALDI-TOF MS analysis, which confirmed that
the products 10a and 10b were monotransglycosylated product
and the 10c was the doubly glycosylated product (see Experi-
mental Section). To discriminate the two monotransglycosylated
products 10a and 10b, specific enzymatic transformation coupled
with MS analysis was applied (Figure S1, Supporting Informa-
tion). Treatment of 10a and 10b with ꢀ-N-acetylglucosaminidase
from Xanthomonas manihotis resulted in the removal of the
remaining terminal GlcNAc, which exposed the R-1,3-Man
residue in 10a or the R-1,6-Man residue in 10b. When the two
intermediates, which had the same molecular mass, were then
treated further with an R-1,2/R-1,3-mannosidase from X. mani-
hotis, only the exposed R-1,3-man residue in 10a would be
removed, leading to the formation of a species with a loss of
162 Da in molecular mass, while the R-1,6-Man residue exposed
in 10b would not be hydrolyzed. Thus, MALDI-TOF MS
analysis of the resulting products led to an unambiguous
assignment of 10a and 10b as the clusters with the additional
glycan attached at the R-1,6-arm and the R-1,3-arm positions,
respectively. Discrimination between 12a and 12b was achieved
by similar enzymatic transformations of 12a with ꢀ-N-acetyl-
glucosaminidase and R-1,2/R-1,3-mannosidase coupled with MS
analysis (Figure S2, Supporting Information).
Further Extension of the Sugar Chains in the Glycan
Clusters by Tandem Enzymatic Transglycosylation. For those
glycan clusters that contain terminal LacNAc moieties such as
compounds 10a, 10b, 10c, and 13, further sugar chain extensions
could be readily achieved by unmasking the terminal GlcNAc
residues followed by repeat enzymatic transglycosylation. For
example, treatment of 10c with ꢀ-1,4-galactosidase from B.
fragilis resulted in selective removal of the four terminal
galactose residues to provide compound 14, in which the four
internal GlcNAc residues were exposed. Simultaneous glyco-
sylation of these GlcNAc residues in 14 was achieved in a single
step by its Endo-A-catalyzed transglycosylation with an excess
amount of Man3GlcNAc-oxazoline (8) (4 equiv per terminal
GlcNAc) to give N-glycan cluster 15 in essentially quantitative
yield (Scheme 4). The HPLC and MALDI-TOF MS profiles of
the purified N-glycan cluster 15 were shown in Figure 1. The
m/z species observed in the MS spectrum is in good agreement
with the calculated molecular mass (observed, m/z ) 6625.98;
calculated, [M + Na] ) 6625.41). Glycan cluster 15 consists
The regioselectivity observed in the EndoM-N175A-catalyzed
transglycosylation also provided an opportunity to introduce
distinct N-glycans at the R-1,6- and R-1,3-arms of the core,
permitting the synthesis of novel hybrid N-glycan clusters. Thus,
after the first selective transglycosylation of 7 with oxazoline
(
5) to produce compound 10a, another distinct glycan,
Man GlcNAc, was successfully introduced to the remaining
3
terminal GlcNAc at the R-1,3-arm by Endo-A to give the hybrid
N-glycan cluster 13 in essentially quantitative yield (Scheme
3). This result represents a remarkable example on the potential
of the enzymatic transglycosylation for expanding the diversity
of the glycan cluster structures.
3 2
of four terminal Man GlcNAc cores and three internal
1
7966 J. AM. CHEM. SOC. 9 VOL. 131, NO. 49, 2009

Chemoenzymatic Synthesis of N-Glycan Clusters
A R T I C L E S
N-glycan clusters were assayed at a concentration of 100 nM,
and the bound glycans were detected by a fluorescence(Cy3)-
labeled streptavidin. Figure 2b demonstrates a typical fluorescent
imaging of the microarray profiles for glycan clusters 9 and
1
5. The lectin-binding profiles of all the synthetic N-glycan
clusters are summarized in Figure 3. For each compound, its
responses to the respective lectins were quantitatively deter-
mined and expressed as mean fluorescence intensity (MFI) of
triplicate detections.
For the simple biotinylated N-glycans, 6 and 7, no apparent
lectin interactions were observed under the screening conditions
(
at 100 nM glycan concentration) except for a weak binding of
the GlcNAc-terminated glycan 7 to ConA, a lectin that is specific
for terminal GlcNAc, Man, and Glc residues. The known sugar-
binding specificity of the 45 lectins used in the present
microarray analysis is listed in Figure S3 (Supporting Informa-
tion). The cluster 9, which contains two terminal Man3GlcNAc2
cores, showed significant signals on spots for lectins Calsepa
and ConA. These results are consistent with the known
specificity of these two lectins, which recognize terminal
mannose/GlcNAc residues and particularly high-mannose gly-
Figure 1. HPLC and MALDI-TOF MS profiles of N-glycan cluster 15.
1
8
can. Glycans 10a and 10b picked up three lectins, Calsepa,
ConA, and RCA120. RCA120 is a lectin specific for terminal
19
ꢀ
-Gal residues. The strong interactions between lectin Calsepa
and 10a or 10b is an interesting observation. Calsepa is a
mannose-binding-type Jacalin-realated lectin (mJRL) with bind-
ing preference to small high-mannose N-glycan (Man2-6). It is
also reported that Calsepa interacts with sialo/asialo complex-
type N-glyans with terminal Gal or GlcNAc. Interestingly
bisecting GlcNAc dramatically enhances binding affinity of
18
complex-type N-glycans with Calsepa. The strong interaction
between Calsepa and 10a or 10b suggests that both the terminal
Man3GlcNAc core and the remaining terminal GlcNAc in 10a
and 10b might bind to two distinct binding sites in the lectin
simultaneously in a concerted manner, thus dramatically en-
hancing its affinity to the lectin. It should be noted that Calsepa
did not show detectable affinity to the complex-type N-glycans
6
and 7 at 100 nM, despite the presence of terminal Gal and
GlcNAc in the glycans.
Figure 2. A 45-lectin microarray platform for probing N-glycan clusters.
(
a) Lectin printing pattern; (b) fluorescent imaging of the lectin-recognition
The lectin microarray also revealed interesting properties for
glycan cluster 10c, which bears four terminal Galꢀ1-4GlcNAc
moieties. First, it showed significantly enhanced affinity to the
profiles of glycan clusters 9 and 15.
2
0
LacNAc-specific galectins RCA120 and ECA as compared
with the glycans 6, 10a, and 10b. The dramatic enhancement
of affinity could be explained by the clustering effect of the
Galꢀ1-4GlcNAc ligands in 10c. Second, the Calsepa binding
signals were also enhanced. More surprisingly, glycan cluster
3 2
Man GlcNAc cores. It would be worthwhile to emphasize that
all the sugar residues in these enzymatically synthetic N-glycan
clusters are connected to each other by the defined natural
glycosidic bonds found in natural N-glycans, which usually
demonstrate confined rotations of the chains than those linked
through flexible spacers. Thus, the N-glycan clusters described
in this work represent a new class of N-glycan clusters that are
different from other synthetic oligosaccharide clusters in which
mono- or oligosaccharides are linked with various spacers.
Lectin Array Characterization of the Synthetic N-Glycan
Clusters. The availability of a class of structurally well-defined
synthetic N-glycan clusters has now provided an opportunity
to investigate how the unusual configuration of multiple
N-glycan cores in a molecule would contribute to their recogni-
tion with various glycan-binding proteins such as lectins. For a
preliminary study, a lectin microarray consisting of 45 lectins
was used (Figure 2a), in which the lectins were immobilized
1
0c also demonstrated significant affinity to ConA, although it
does not contain terminal GlcNAc/Man residues. These results
suggest that the appropriate clustering arrangement of the
internal Man3GlcNAc2 cores in 10c may generate novel new
binding sites or interfaces for lectin ConA that hitherto favors
terminal mannose/GlcNAc residues. As to the Man9GlcNAc2-
containing cluster 12a, it was recognized by 5 lectins in the
microarray. In addition to its expected affinity to the high-
mannose recognizing lectins, including Calsepa, ConA, and
HHL, it also picked up two unexpected lectins, ABA and UDA.
(18) Nakamura-Tsuruta, S.; Uchiyama, N.; Peumans, W. J.; Van Damme,
1
7
E. J.; Totani, K.; Ito, Y.; Hirabayashi, J. FEBS J. 2008, 275, 1227–
on a glass slide as previously reported. All the biotinylated
1
239.
(
19) Baenziger, J. U.; Fiete, D. J. Biol. Chem. 1979, 254, 9795–9799.
(
17) Kuno, A.; Uchiyama, N.; Koseki-Kuno, S.; Ebe, Y.; Takashima, S.;
(20) De Boeck, H.; Loontiens, F. G.; Lis, H.; Sharon, N. Arch. Biochem.
Biophys. 1984, 234, 297–304.
Yamada, M.; Hirabayashi, J. Nat. Methods 2005, 2, 851–856.
J. AM. CHEM. SOC. 9 VOL. 131, NO. 49, 2009 17967

A R T I C L E S
Huang et al.
Figure 3. Glyco-epitope profiles of novel synthetic N-glycans detected by lectin arrays. Overlay plots were produced using JMP-Genomics 4.0 software
package (SAS Institute in Cary, North Carolina). Each plot was an overlay plot of GlcNAc-LC-biotin (4) that served as a control in the assay (marked with
a symbol “X”) and the respective N-glycan cluster (marked with a symbol “0”).
Lectin ABA has a known specificity for terminal Galꢀ1-
GalNAc structure and a very weak affinity to the terminal
GlcNAc residue on complex-type N-glycans (K at 100 µM
to UDA. The glycan cluster 14, which possesses multiple
terminal GlcNAcꢀ1,2-Man structures, demonstrated strong
affinity to lectins Calsepa, ConA, and ABA, and moderate
binding capacity to HHL and UDA. While its strong interactions
with Calsepa, ConA, and ABA are expected because of the
clustering effect of multiple terminal GlcNAc residues, the
interaction between glycan 14 and lectin HHL that has a known
specificity for mannose is a new observation. Compound 15,
which possesses 4 terminal Man3GlcNAc2 cores and 3 internal
Man3GlcNAc2 cores, is another novel N-glycan cluster that
demonstrates unusual lectin recognition properties. A compari-
son of this glycan cluster with the simpler glycan cluster 9 (2
terminal Man3GlcNAc2 cores and 1 internal Man3GlcNAc core)
revealed clear cluster effects in lectin recognition (Figure 3).
Under the same assay conditions, glycan 9 recognized only two
lectins, Calsepa and ConA, that are specific for terminal
mannose residues and high-mannose-type N-glycans. However,
glycan cluster 15 was recognized by seven different lectins in
the microarray. In addition to Calsepa and ConA, compound
15 showed high affinity to the other three high-mannose specific
3
d
2
1
level). The reason why 12a exhibits such a high affinity to
lectin ABA is yet to be elucidated, but perhaps both the Man9
structure and the terminal GlcNAc at the other arm are involved
in a simultaneous interaction with two binding sites in the lectin.
The glycan clusters 10a and 10b, which contain a terminal
GlcNAc but carry a complex glycan on the other arm, were not
recognized by ABA. Lectin UDA (Urtica dioica agglutinin) is
known to be specific for chitin oligomers (N,N′,N′′-triacetyl-
22,23
chitotriose and higher chitin oligosaccharides).
The signifi-
cant interaction between UDA and compound 12a may suggest
that the three internal GlcNAc residues may form a discontinu-
ous glyco-epitope that would fit to the chitotriose-binding
23
domain in the lectin. Alternatively, the Man9GlcNAc2 moiety
itself may also contribute directly to the binding, as compound
1
0a, which contains a similar three internal GlcNAc motif but
carries a complex N-glycan at the 6-arm, clearly did not bind
to UDA. The glycan cluster 13, which possesses both terminal
Galꢀ1-4GalNAc and high-mannose moieties, demonstrated the
expected binding specificity for lectins Calpesa, ConA, and
RCA120. But again, it also showed significant binding affinity
2
4
25
lectins, GNA, HHL, and NPA. The fact that glycan cluster
9 did not show sufficient affinity to these three lectins (negative
at 100 nM) strongly suggests that these three lectins recognize
particularly a high-density cluster of high-mannose type glycans,
as demonstrated by the glycan cluster 15. Moreover, glycan
(
(
21) Nakamura-Tsuruta, S.; Kominami, J.; Kuno, A.; Hirabayashi, J.
Biochem. Biophys. Res. Commun. 2006, 347, 215–220.
22) (a) Shibuya, N.; Goldstein, I. J.; Shafer, J. A.; Peumans, W. J.;
Broekaert, W. F. Arch. Biochem. Biophys. 1986, 249, 215–224. (b)
Broekaert, W. F.; Van Parijs, J.; Leyns, F.; Joos, H.; Peumans, W. J.
Science 1989, 245, 1100–1102.
(24) (a) Hester, G.; Wright, C. S. J. Mol. Biol. 1996, 262, 516–531. (b)
Wright, C. S.; Hester, G. Structure 1996, 4, 1339–1352.
(25) Kaku, H.; Van Damme, E. J.; Peumans, W. J.; Goldstein, I. J. Arch.
Biochem. Biophys. 1990, 279, 298–304.
(
23) Harata, K.; Muraki, M. J. Mol. Biol. 2000, 297, 673–681.
1
7968 J. AM. CHEM. SOC. 9 VOL. 131, NO. 49, 2009

Chemoenzymatic Synthesis of N-Glycan Clusters
A R T I C L E S
cluster 15 also picked up two chitin oligosaccharide-specific
250 mm) at 40 °C. The column was eluted at a flow rate of 1.0
mL/min using a linear gradient of 0-90% MeCN containing 0.1%
trifluoroacetic acid (TFA) for 30 min. The yields were calculated
on the basis of the HPLC quantification of both starting materials
and products [absorbance (abs) at 214 nm], using the following
formula: yield (%) ) [product abs/(starting material Abs + product
abs)] × 100. Preparative HPLC was performed on a Waters 600
HPLC instrument with a preparative C18 column (Symmetry300,
23
26
lectins, UDA and LEL, that recognize chitin oligosaccharides
with three or more GlcNAc moieties. A plausible explanation
is that the compact cluster of the multiple GlcNAcꢀ1-4GlcNAc
moieties in 15 forms a novel discontinuous epitope that mimics
the chitin oligosaccharide to provide a high-affinity structure
to interact with the binding domain in the lectin. Notably, among
the synthetic glycan clusters, only compound 15 showed
significant binding to lectin LEL that is known to be specific
for N-acetyl-chitooligosaccharide moieties (Figure 3). These
experimental data suggest that the unusual configuration of the
N-glycan cores in the defined N-glycan clusters creates novel
lectin recognition motifs or glyco-epitopes that may implicate
special roles of unusual N-glycans in a biological system.
Recently, Dennis and co-workers have reported that the number
and degree of branching (e.g., tetra-antennary vs biantennary)
of complex type N-glycans in cell-surface receptors are critical
factors to regulate cell proliferation and differentiation, due to
the distinct affinities of the differentially branched N-glycans
7.0 µm, 19 mm × 300 mm). The column was eluted with a suitable
gradient of water/acetonitrile containing 0.1% TFA. NMR spectra
were measured with JEOL ECX 400 MHz and/or Inova 500 MHz
NMR spectrometers. All chemical shifts were assigned in ppm. The
ESI-MS spectra were measured on a Waters Micromass ZQ-4000
single quadruple mass spectrometer. MALDI-TOF MS measure-
ment was performed on an Autoflex II MALDI-TOF mass
spectrometer (Bruker Daltonics). The instrument was calibrated by
using ProteoMass Peptide MALDI-MS calibration kit (MSCAL2,
Sigma/Aldrich). The matrix used for glycans is 2,5-dihydroxyben-
zoic acid (DHB) (10.0 mg/mL in 50% acetonitrile containing 0.1%
trifluoroacetic acid). The measuring conditions in detail: 337 nm
nitrogen laser with 100 µJ output; laser frequency 50.0 Hz; laser
power 30-45%; linear mode; positive polarity; detection range
1000-10000; pulsed ion extraction: 70 ns; high voltage: on;
realtime smooth: high; shots: 500-2000.
2
7
to galectins. On the other hand, the highly branched and
compactly packed N-glycan clusters described in this work,
which show unusually high affinity to some specific lectins, may
be used as specific inhibitors to decipher the functional roles
of N-glycans and lectins in a given biological process.
Synthesis of 1-(6-Biotinamidohexanoylamino)-1-deoxy-2-ac-
etamido-2-deoxy-ꢀ-D-glucopyranose (4). 1-ꢀ-Azido-GlcNAc (1)
(
10 mg, 38 µmol) was dissolved in MeOH (1.0 mL) containing
Conclusion
5% palladium on carbon (5.0 mg). The mixture was hydrogenated
at r.t. under atmospheric pressure overnight. The residue was filtered
via a Celite pad, and the filtrate was concentrated. The obtained
crude syrup of 1-ꢀ-amino-GlcNAc (2) was directly used without
additional purification. A solution of the crude compound 2 and
NHS-LC-biotin (3) (20 mg, 44 µmol) in a mixed solvent of
phosphate buffer (50 mM, pH 7.5, 1.0 mL), MeCN (1.0 mL) and
DMSO (1.0 mL) was shaken at r.t. overnight. The residue was
subject to preparative HPLC for purification. The fractions contain-
ing product were combined and lyophilized to give GlcNAc-LC-
A facile synthesis of a class of novel N-glycan clusters was
achieved via tandem chemoenzymatic tranglycosylation. The
unusual configurations and highly branched packing of the
subunit N-glycans in the clusters create new structural motifs,
demonstrating novel carbohydrate-lectin recognition patterns.
These synthetic N-glycan clusters should be valuable for
functional glycomics studies to unveil new functions for both
glycans and carbohydrate-binding proteins.
1
biotin (4) as white powder (17 mg, 79% for two steps). H NMR
(
4
D
2
O, 400 MHz): δ 4.93 (d, 1H, J ) 9.6 Hz, H-1 of GlcNAc),
Experimental Section
.47 (dd, 1H, J ) 4.8, 8.0 Hz, H-7 of biotin), 4.29 (dd, 1H, J )
Materials and Methods. 1-ꢀ-Azido-GlcNAc (1) was prepared
4.4, 8.0 Hz, H-8 of biotin), 3.74 (dd, 1H, J ) 2.0, 12.4 Hz, H-6a
of GlcNAc), 3.67 (t, 1H, J ) 10.0 Hz, H-2 of GlcNAc), 3.61 (dd,
1H, J ) 4.6, 12.4 Hz, H-6b of GlcNAc), 3.47 (t, 1H, J ) 10.0 Hz,
H-3 of GlcNAc), 3.38 (m, 2H, H-4 and H-5 of GlcNAc), 3.21 (m,
2
8
by the reported method. Succinimidyl-6-(biotinamido)hexanoate
(
NHS-LC-biotin, 3) was purchased from Pierce Biotechnology, Inc.
13 10
CT-GlcNAc oxazoline (5),
Man
Man
3
GlcNAc oxazoline,
and
1
2
9
GlcNAc oxazoline were synthesized following our previ-
1H, H-4 of biotin), 3.03 (t, 2H, J ) 6.8 Hz, CH NHCO), 2.87 (dd,
2
ously reported procedures. The ꢀ-1,4-galactosidase, ꢀ-N-acetylglu-
cosaminidase, and R-1,2/1,3-mannosidase were purchased from
New England Biolabs, Inc. Endo-A was overproduced, following
1H, J ) 4.8, 13.0 Hz, H-6a of biotin), 2.63 (d, 1H, J ) 13.0 Hz,
H-6b of biotin), 2.14 (m, 4H, CH
2
CONH), 1.86 (s, 3H, Ac of
), 1.16 (m,
2 2
H, CH ); C NMR (D O, 100 MHz) δ 177.7, 176.6, 174.5, 165.3,
8.3, 77.6, 74.1, 71.2, 69.4, 62.1, 60.5, 60.2, 55.4, 54.3, 39.7, 39.0,
5.7, 35.5, 28.0, 27.8, 27.7, 25.5, 25.2., 25.0, 22.0; analytical HPLC:
2 2
GlcNAc), 1.61-1.34 (m, 8H, CH ), 1.26 (m, 2H, CH
29
13
the literature. EndoM-N175A was overproduced according to the
2
7
3
1
2
previously reported method. The activity unit of Endo-A was
defined as the following: 1 unit of Endo-A is the amount of enzyme
required to hydrolyze 1 µmol Man
centration, 10 mM) in one minute at 30 °C in a phosphate buffer
50 mM, pH 6.5). The unit of glycosynthase activity of EndoM-
N175A was defined as the following: 1 unit of EndoM-N175A is
the amount of enzyme required to transfer 1 µmol Man GlcNAc
9 2
GlcNAc Asn (substrate con-
t ) 12.2 min; ESI-MS: calculated for C24H N O S, M ) 559.27
R
41 5 8
+
Da; found, 560.67 [M + H] .
(
Synthesis of Biotinylated Complex Type N-Glycan (6). A
solution of GlcNAc-LC-biotin (4) (1.0 mg, 1.8 µmol) and CT-
GlcNAc oxazoline (5) (7.5 mg, 5.3 µmol) in a phosphate buffer
9
oxazoline to GlcNAc-pNP at 1:2 donor/acceptor ratio in one minute.
All the other reagents were purchased from Sigma/Aldrich and used
as received.
Analytical RP-HPLC was performed on a Waters 626 HPLC
instrument with a Symmetry300 C18 column (5.0 µm, 4.6 mm ×
(
50 mM, pH 7.5, 100 µL) was incubated with EndoM-N175A (125
mU) at 30 °C for 8 h. The residue was subject to preparative HPLC
for purification. The fractions containing product were combined
and lyophilized to give CT-GlcNAc
2
-LC-biotin (6) as white powder
1
(
2.5 mg, 72%). H NMR (D O, 400 MHz): δ 4.97 (s, 1H, H-1 of
2
4
1
Man ), 4.92 (d, 1H, J ) 9.6 Hz, H-1 of GlcNAc ), 4.78 (s, 1H,
4
3
(
(
26) Kilpatrick, D. C. Biochem. J. 1980, 185, 269–272.
H-1 of Man ′), 4.62 (s, 1H, H-1 of Man ), 4.46 (m, 4H, H-1 of
2 5 5
27) Lau, K. S.; Partridge, E. A.; Grigorian, A.; Silvescu, C. I.; Reinhold,
V. N.; Demetriou, M.; Dennis, J. W. Cell 2007, 129, 123–134.
28) O’Reilly, M. K.; Collins, B. E.; Han, S.; Liao, L.; Rillahan, C.; Kitov,
P. I.; Bundle, D. R.; Paulson, J. C. J. Am. Chem. Soc. 2008, 130,
GlcNAc , H-1 of GlcNAc , H-1 of GlcNAc ′, H-7 of biotin), 4.33
6 6
(
)
dd, 2H, J ) 2.4, 7.8 Hz, H-1 of Gal and Gal ′), 4.28 (dd, 1H, J
4.6, 7.8 Hz, H-8 of biotin), 4.11 (m, 1H), 4.06 (m, 1H), 3.97
m, 1H), 3.19 (quintet, 1H, J ) 4.8 Hz, H-4 of biotin), 3.03 (t, 2H,
(
(
7
736–7745.
J ) 6.8 Hz, CH NHCO), 2.86 (dd, 1H, J ) 4.8, 12.8 Hz, H-6a of
biotin), 2.65 (d, 1H, J ) 12.8 Hz, H-6b of biotin), 2.12 (m, 4H,
2
(
29) Fujita, K.; Tanaka, N.; Sano, M.; Kato, I.; Asada, Y.; Takegawa, K.
Biochem. Biophys. Res. Commun. 2000, 267, 134–138.
J. AM. CHEM. SOC. 9 VOL. 131, NO. 49, 2009 17969

A R T I C L E S
Huang et al.
Man(1,6)-[GlcNAcMan(1,3)]-ManGlcNAc -LC-
CH
2
CONH), 1.94 (s, 3H, Ac), 1.91 (s, 3H, Ac), 1.90 (s, 3H, Ac),
.85 (s, 3H, Ac), 1.61-1.34 (m, 8H, CH ), 1.28 (m, 2H, CH ),
O, 100 MHz) δ 177.7, 176.6,
CT-GlcNAc
2
2
1
1
1
1
9
2
2
biotin (10a): white powder, analytical HPLC: tR ) 10.8 min;
1
3
.16 (m, 2H, CH
2
); C NMR (D
2
MALDI-TOF-MS: calculated for C120
Da; found, 3097.78 [M + Na] .
H
199
N
11
O
78S, M ) 3074.17
6
6
+
74.7, 174.6, 174.5, 165.3 (7 × CO), 102.9 (C-1 of Gal and Gal ′),
2
3
4
01.5 (C-1 of GlcNAc ), 100.4 (C-1 of Man ), 99.7 (C-1 of Man ),
CT-GlcNAc
2
Man(1,3)-[GlcNAcMan(1,6)]-ManGlcNAc
2
-LC-
) 11.1 min;
78S, M ) 3074.17
5
5
4
9,6 (C-1 of GlcNAc and GlcNAc ′), 97.0 (C-1 of Man ′), 78.6
biotin (10b): white powder, analytical HPLC: t
R
1
(
C-1 of GlcNAc ), 62.8 (C-8 of biotin), 60.5 (C-7 of biotin), 55.8
C-4 of biotin), 39.7 (C-6 of biotin), 39.0 (2 × CH NHCO), 35.7,
5.5 (2 × CH CONH), 27.9, 27.8, 27.6, 25.4, 25.2., 24.8 (6 ×
CH ) 11.2 min;
48S, M ) 1978.77 Da;
MALDI-TOF-MS: calculated for C120
Da; found, 3098.36 [M + Na] .
H
199
N
11
O
(
3
2
+
2
CT-GlcNAc
2
Man(1,6)-[CT-GlcNAc
LC-biotin (10c): white powder, analytical HPLC: t
2
Man(1,3)]-ManGlcNAc
) 10.2 min;
118S, M ) 4496.24
2
-
2
), 22.3, 22.2, 22.0 (4 × Ac); analytical HPLC: t
R
R
MALDI-TOF-MS: calculated for C78
H
130
N
8
O
MALDI-TOF-MS: calculated for C174
Da; found, 4519.69 [M + Na] .
H
288
N
14
O
+
found, 2002.68 [M + Na] .
+
Synthesis of the Biotinylated Complex Type N-Glycan (7).
A solution of CT-GlcNAc -LC-biotin (6) (2.5 mg, 1.3 µmol) in a
phosphate buffer (50 mM, pH 5.5, 300 µL) was incubated with
-1,4-galactosidase (100 U) from Bacteroides fragilis (New England
Synthesis of N-Glycan Cluster 10c from the Monotransgly-
cosylated Compound 10a. A solution of 10a (0.30 mg, 0.10 µmol)
and CT-GlcNAc oxazoline (5) (0.50 mg, 0.35 µmol) in a phosphate
buffer (50 mM, pH 7.5, 5.0 µL) was incubated with EndoM-N175A
30 mU) at 30 °C for 8 h. The reaction mixture was subject to
HPLC purification to give the doubly glycosylated product 10c
2
ꢀ
Biolabs) at 37 °C for 24 h. The reaction was monitored by analytical
HPLC until complete removal of terminal galactose. The residue
was subject to preparative HPLC. The fractions containing product
(
(
50% yield).
were combined and lyophilized to give GlcNAc
LC-biotin (7) as white powder (2.1 mg, quantitative yield). H NMR
2 3 2
Man GlcNAc -
Transglycosylation with the Man9GlcNAc Oxazoline (11) by
1
EndoM-N175A. A solution of GlcNAc
2 3 2
Man GlcNAc -LC-biotin
4
2
(D O, 400 MHz): δ 4.98 (s, 1H, H-1 of Man ), 4.93 (d, 1H, J )
1 4
(
7) (0.20 mg, 0.12 µmol) and Man GlcNAc oxazoline (11) (0.60
9
9
.6 Hz, H-1 of GlcNAc ), 4.78 (s, 1H, H-1 of Man ′), 4.64 (s, 1H,
3 2
mg, 0.36 µmol) in a phosphate buffer (50 mM, pH 7.5, 15 µL)
was incubated with EndoM-N175A (10 mU) at 30 °C. The reaction
was monitored by analytical HPLC. After 4 h, HPLC indicated the
formation of the 6-arm glycosylated product 12a in 27% yield and
the 3-arm glycosylated product 12b in 3% yield. The products were
purified by HPLC. The doubly glycosylated product was formed
in only trace amount (as indicated by MS analysis), which was not
isolated for further analysis.
H-1 of Man ), 4.48 (m, 2H, H-1 of GlcNAc , H-7 of biotin), 4.43
5
5
(
d, 2H, J ) 8.6 Hz, H-1 of GlcNAc and GlcNAc ′), 4.29 (dd, 1H,
J ) 4.4, 8.0 Hz, H-8 of biotin), 4.12 (m, 1H), 4.05 (m, 1H), 3.97
m, 1H), 3.20 (quintet, 1H, J ) 4.8 Hz, H-4 of biotin), 3.03 (t, 2H,
J ) 6.8 Hz, CH NHCO), 2.87 (dd, 1H, J ) 4.8, 13.2 Hz, H-6a of
biotin), 2.63 (d, 1H, J ) 12.8 Hz, H-6b of biotin), 2.13 (m, 4H,
CH
CONH), 1.94 (s, 3H, Ac), 1.91 (s, 6H, 2 × Ac), 1.86 (s, 3H,
Ac), 1.62-1.34 (m, 8H, CH ), 1.27 (m, 2H, CH ), 1.15 (m, 2H,
O, 100 MHz) δ 177.7, 176.6, 174.7, 174.6,
(
2
2
2
2
Man
9
GlcNAc
2
Man(1,6)-[GlcNAcMan(1,3)]-ManGlcNAc
2
-LC-
) 10.7 min;
88S, M ) 3318.13
13
CH
2 2
); C NMR (D
biotin (12a): white powder, analytical HPLC: t
MALDI-TOF-MS: calculated for C128
Da; found, 3341.78 [M + Na] .
R
2
1
74.5, 165.3 (7 × CO), 101.3 (C-1 of GlcNAc ), 100.4 (C-1 of
H
213
N
9
O
3
4
5
5
Man ), 99.6 (C-1 of Man , C-1 of GlcNAc and GlcNAc ′), 97.0
+
4
1
(
7
6
C-1 of Man ′), 80.4, 79.5, 78.6 (C-1 of GlcNAc ), 78.2, 76.4, 76.2,
Man
9
GlcNAc
2
Man(1,3)-[GlcNAcMan(1,3)]-ManGlcNAc
2
-LC-
) 11.0 min;
88S, M ) 3318.13
5.8, 74.4, 74.3, 73.5, 73.3, 73.2, 72.8, 72.7, 72.0, 70.2, 69.9, 69.5,
9.4, 67.3, 65.8, 65.7, 62.1 (C-8 of biotin), 61.7, 61.6, 60.6 (C-7
biotin (12b): white powder, analytical HPLC: t
MALDI-TOF-MS: calculated for C128
Da; found, 3341.57 [M + Na] .
Synthesis of N-Glycan Cluster 13. A solution of 10a (0.10 mg,
R
H
213
N
9
O
of biotin), 60.2, 59.8, 55.4, 55.3 (C-4 of biotin), 39.7 (C-6 of biotin),
+
3
2
9.0 (2 × CH
7.6, 25.4, 25.2., 24.8 (6 × CH
) 11.4 min; MALDI-TOF-MS: calculated for
2
NHCO), 35.7, 35.5 (2 × CH
2
CONH), 27.9, 27.8,
2
), 22.3, 22.2, 22.0 (4 × Ac);
3
3
3 nmol) and Man GlcNAc oxazoline (5) (0.10 mg, 0.14 µmol) in
analytical HPLC: t
38S, M ) 1654.66 Da; found, 1678.07 [M + Na] .
Synthesis of the N-Glycan Cluster (9). A solution of
GlcNAc Man GlcNAc -LC-biotin (7) (0.20 mg, 0.12 µmol) and
Man GlcNAc oxazoline (8) (0.50 mg, 0.72 µmol) in phosphate
R
a phosphate buffer (50 mM, pH 7.5, 5.0 µL) was incubated with
Endo-A (1.0 mU) at 30 °C for 2 h. Analytical HPLC monitoring
indicated the formation of the hybrid glycosylated product 13 in a
quantitative yield. The product was purified by HPLC to give CT-
+
66 110 8
C H N O
2
3
2
3
GlcNAc
tin (13). Analytical HPLC: t
calculated for C146 98S, M ) 3763.40 Da; found, 3786.87
2
Man(1,6)-[Man
3
GlcNAc
2 2
Man(1,3)]-ManGlcNAc -LC-bio-
buffer (50 mM, pH 7.5, 15 µL) was incubated with Endo-A (2.0
mU) at 30 °C for 1 h. The reaction was monitored by analytical
HPLC until the completion of the reaction. The residue was subject
to preparative HPLC purification. The fractions containing product
R
) 10.4 min; MALDI-TOF-MS:
242 12
H N O
+
[
M + Na] .
Synthesis of N-Glycan Cluster 14. A solution of 10c (0.15 mg,
were combined and lyophilized to give Man
Man GlcNAc Man(1,3)]-ManGlcNAc -LC-biotin (9) as a white
powder (88% yield). Analytical HPLC: t ) 11.1 min; MALDI-
78S, M ) 3033.14 Da, found,
3
GlcNAc
2
Man(1,6)-
33 nmol) in a phosphate buffer (50 mM, pH 5.5, 30 µL) was
incubated with ꢀ-1,4-galactosidase (10 U) at 37 °C for overnight.
The reaction was monitored by analytical HPLC until the complete
removal of all the terminal galactose moieties. The residue was
subject to preparative HPLC purification. The fractions containing
[
3
2
2
R
TOF-MS: calculated for C118
H N
196 10
O
+
3
055.12 [M + Na] .
product were combined and lyophilized to give GlcNAc
GlcNAc Man(1,6)-[GlcNAc Man GlcNAc Man(1,3)]-ManGlcNAc
LC-biotin (14) (quantitative yield) as a white powder. Analytical
HPLC: 10.7 min; MALDI-TOF-MS: calculated for
2 3
Man -
Transglycosylation of Glycan 7 with the Complex Type
Glycan Oxazoline 5. Synthesis of Glycan Clusters 10a and
2
2
3
2
2
-
10b. A solution of GlcNAc
2 3 2
Man GlcNAc -LC-biotin (7) (0.50 mg,
t
N
R
)
O
0
.30 µmol) and CT-GlcNAc oxazoline (5) (1.3 mg, 0.91 µmol) in
+
C
150
H
248
14
98S, M ) 3847.67 Da; found, 3870.97 [M + Na] .
Synthesis of N-Glycan Cluster 15. A solution of 14 (0.10 mg,
6 nmol) and Man GlcNAc oxazoline (5) (0.30 mg, 0.44 µmol) in
a phosphate buffer (50 mM, pH 7.5, 20 µL) was incubated with
EndoM-N175A (30 mU) at 30 °C. The reaction was monitored by
analytical HPLC. After 4 h, HPLC showed the formation of three
new products, which were readily purified by HPLC to give the
2
3
a phosphate buffer (50 mM, pH 7.5, 15 µL) was incubated with
Endo-A (2.0 mU) at 30 °C. HPLC monitoring indicated the
complete transglycosylation after 2 h incubation. The reaction
mixture was subject to preparative HPLC. The fractions containing
product were combined and lyophilized to give the N-glycan cluster
6
-arm glycosylated product CT-GlcNAc
Man(1,3)]-LC-biotin (10a) in 25% yield; the 3-arm glycosylated
product CT-GlcNAc Man(1,3)-[GlcNAcMan(1,6)]-LC-biotin (10b)
in 5% yield, and the doubly glycosylated product 10c in 8% yield.
2
Man(1,6)-[GlcNAc-
2
(15) (quantitative yield). Analytical HPLC: t ) 9.6 min; MALDI-
R
1
7970 J. AM. CHEM. SOC. 9 VOL. 131, NO. 49, 2009

Chemoenzymatic Synthesis of N-Glycan Clusters
A R T I C L E S
TOF-MS: calculated for C254
420 18
H N O178S, M ) 6602.41 Da; found,
plots were produced using JMP-Genomics 4.0 software package
(SAS Institute in Cary, North Carolina).
+
6
625.98 [M + Na] .
Lectin Microarray Analysis. Lectins were immobilized on glass
Acknowledgment. We thank Prof. Kaoru Takegawa for provid-
ing the pGEX-2T/Endo-A plasmid, Prof. Kenji Yamamoto and Ms.
Midori Umekawa for providing the pET23b-Endo-M plasmid, and
Dr. Cishan Li for expressing the endoenzymes. This work was
supported by the National Institutes of Health (NIH) Grants R01
GM080374 (to L.X.W.) and U01 CA128416 (to D.W.).
1
7
slides according to the previously reported method. The lectin
microarray profilings of the glycan clusters were carried out
17
following the literature method, with some modifications. Briefly,
the biotinylated N-glycan clusters were applied on lectin arrays at
a concentration of 100 nM and incubated at 4 °C for 5 h. After
washing to remove the unbound N-glycans, the lectin arrays were
incubated with a fluorescence (Cy3)-labeled streptavidin at a
concentration of 1 µg/mL at r.t. for 30 min. The array-captured
N-glycans were then visualized and quantified by scanning the
arrays with a specialized GlycoStation (GP Biosciences Ltd.,
Yokohama, Japan). For each compound, the binding responses with
the lectin arrays were quantitatively determined and expressed as
mean fluorescence intensity (MFI) of triplicate detections. In all
the assays, GlcNAc-LC-biotin (4) was used as the control. Overlay
Supporting Information Available: Complete ref 4a; the
enzymatic transformation coupled with MS analysis for the
characterization of the regioselective transglycosylation products
1
13
1
0a, 10b, and 12a; the H NMR, C NMR, MS spectra, and
HPLC profiles of key compounds. This material is available
free of charge via the Internet at http://pubs.acs.org.
JA9078539
J. AM. CHEM. SOC. 9 VOL. 131, NO. 49, 2009 17971