9614
J . Org. Chem. 1998, 63, 9614-9615
regio- and glycan-specific glycosylation of proteins. This
method is rapid, utilizes reagents that may be prepared in
a facile manner and, in principle, is unconstrained in the
scope of sites and glycans that may be conjugated. The
strategy involves the introduction of cysteine at preselected
positions and then reaction of its thiol residue with gly-
comethanethiosulfonate reagents (Scheme 1). Methaneth-
iosulfonate (MTS) reagents react specifically and quantita-
tively with thiols14 and allow the controlled formation of
neutral disulfide linkages. Recently, we have successfully
used the representative serine protease subtilisin Bacillus
lentus (SBL) as our vehicle and continue to do so here.15 SBL
is an ideal model protein for evaluating the validity of this
strategy as it does not contain a natural cysteine and is not
naturally glycosylated.
Con tr olled Site-Selective Glycosyla tion of
P r otein s by a Com bin ed Site-Dir ected
Mu ta gen esis a n d Ch em ica l Mod ifica tion
Ap p r oa ch
Benjamin G. Davis, Richard C. Lloyd, and
J . Bryan J ones*
Department of Chemistry, University of Toronto,
80 St. George Street, Toronto, Ontario M5S 3H6, Canada
Received August 17, 1998
Surface glycoproteins act as markers in cell-cell com-
munication events that determine microbial virulence,1
inflammation2 and host immune responses.3 In addition, the
correct glycosylation of proteins is critical to their expression
and folding4 and increases their thermal and proteolytic
stability.5 Glycoproteins occur naturally in a number of
forms (glycoforms)6 that possess the same peptide backbone
but differ in both the nature and site of glycosylation. The
differences exhibited6,7 by each component within these
microheterogeneous mixtures present regulatory difficulties8
and problems in determining exact function. To explore these
key properties, there is a pressing need for methods that
will not only allow the preparation of pure glycosylated
proteins but will also allow the preparation of nonnatural
variants for the determination of structure-activity rela-
tionships (SARs). The few studies that have compared single
glycoforms successfully have required abundant sources and
extensive chromatographic separation.9 Neoglycoproteins,10
formed via unnatural linkages between sugars and proteins,
provide an invaluable alternative source of carbohydrate-
protein conjugates.11 In particular, chemical glycosylation
allows control of the glycan structure and the nature of the
sugar-protein bond. However, despite these advantages,
existing methods for their preparation11a typically generate
mixtures. Advances in the site-specific glycosylation of
bovine serum albumin (BSA) have been made.12 However,
these methods rely upon modification of an existing cysteine
in BSA and, as such, allow no flexibility in the choice of
glycosylation site. We therefore set ourselves the goal of
developing a versatile route to neoglycoproteins without
these limitations.
Four SBL sites at different locations and of different
characteristics were selected for mutation to cysteine in
order to provide a broad test of the glycosylation methodol-
ogy. S156 of the S1-pocket is a surface-exposed residue that
permits the introduction of externally disposed glycans
mirroring those found naturally in glycoproteins.16 In con-
trast, N62 in the S2 pocket, S166 in the S1 pocket, and L217
in the S1′ pocket have side chains that are internally oriented
and test the applicability of the method for introducing
sugars at hindered locations. Broad applicability with
respect to the sugar moiety was evaluated by using the
representative series of protected and deprotected mono- and
disaccharide methanethiosulfonates 1a -k . These were pre-
pared from their parent carbohydrates in good to excellent
yields (Schemes 2 and 3). Two types of glycosylating re-
agents, the anomeric methanethiosulfonate 1a and the ethyl-
tethered methanethiosulfonates 1b,c,g,h , were prepared
from D-glucose (2a , Scheme 2). The preparation of these
reagents in fully protected 1a ,g,h and deprotected 1b,c
forms allowed the effects of increased steric bulk and hy-
drophobicity to be assessed. Parallel routes allowed sim-
ilarly efficient access to the R-D-manno-MTS reagents 1d and
1i, which are epimeric at C-2 relative to 1b and 1g, respec-
tively, and the â-D-galacto-MTS reagents 1e and 1j, epimeric
at C-4 relative to 1c and 1h , respectively (Scheme 3).
The glyco-MTS reagents 1a -k were then reacted with
SBL-N62C, -S156C, -S166C, and -L217C in aqueous buffer.15
These reactions were rapid and quantitative, as judged by
monitoring changes in specific activity and by titration of
residual free thiols with Ellman’s reagent.17 The glycosylated
chemically modified mutants (CMMs) were purified by size-
exclusion chromatography and dialysis, and their structures
were confirmed by rigorous ES-MS analysis. The CMMs
each appeared as a single band on nondenaturing gradient
PAGE, thereby establishing their high purities. In all cases,
modification with the fully deprotected reagents 1b-f led
to site-specific glycosylations and the formation of single
glycoforms. These are the first examples of homogeneous
neoglycoproteins in which both the site of glycosylation and
the strucure of the glycan introduced have been predeter-
Site-directed mutagenesis combined with chemical modi-
fication has permitted us to realize this goal and, for the
first time,13 provides a general method that allows both
(1) Sharon, N.; Lis, H. Essays Biochem. 1995, 30, 59-75.
(2) (a) Lasky, L. A. Annu. Rev. Biochem. 1995, 64, 113-139. (b) Weis,
W. I.; Drickamer, K. Annu. Rev. Biochem. 1996, 65, 441-473.
(3) (a) Varki, A. Glycobiology 1993, 3, 97-130. (b) Dwek, R. A. Chem.
Rev. 1996, 96, 683-720.
(4) Helenius, A. Mol. Biol. Cell 1994, 5, 253-265.
(5) Opdenakker, G.; Rudd, P. M.; Ponting, C. P.; Dwek, R. A. FASEB J .
1993, 7, 1330-1337.
(6) Rademacher, T. W.; Parekh, R. B.; Dwek, R. A. Glycobiology 1988,
57, 785-838.
(7) (a) Parekh, R. B.; Tiemeier, D. C. Biochemistry 1989, 28, 7670-7679.
(b) Knight, P. Biotechnology 1989, 7, 35-40.
(13) During the course of this work, Boons announced his intention to
use a similar method for the regioselective glycosylation of IgG (ref 12c)
but has published no details.
(14) For reviews see (a) Kenyon, G. L.; Bruice, T. W. Methods Enzymol.
1977, 47, 407-430. (b) Wynn, R.; Richards, F. M. Methods Enzymol. 1995,
251, 351-356.
(8) (a) Liu, D. T. Y. Trends Biotechnol. 1992, 10, 114-120. (b) Bill, R.
M.; Flitsch, S. F. Chem. Biol. 1996, 3, 145-149.
(9) Rudd, P. M.; Dwek, R. A. Biochemistry 1994, 33, 17-22.
(10) Krantz, M. J .; Holtzmann, N. A.; Stowell, C. P.; Lee, Y. C.
Biochemistry 1976, 15, 3963-3968.
(11) For reviews, see: (a) Stowell, C. P.; Lee, Y. C. Adv. Carbohydr. Chem.
Biochem. 1980, 37, 225-281. (b) Neoglycoconjugates: Preparation and
Applications; Lee, Y. C., Lee, R. T., Eds.; Academic: London, 1994. (c)
Abelson, J . N., Simon., M. I., Eds. Methods Enzymol. 1994, 242. (d) Lee, Y.
C.; Lee, R. T., Eds. Ibid. 1994, 247.
(12) (a) Davis, N. J .; Flitsch, S. L. Tetrahedron Lett. 1991, 32, 6793-
6796. (b) Wong, S. Y. C.; Guile, G. R.; Dwek, R. A.; Arsequell, G. Biochem.
J . 1994, 300, 843-850. (c) Macindoe, W. M.; van Oijen, A. H.; Boons, G.-J .
J . Chem. Soc., Chem. Commun. 1998, 847-848.
(15) (a) Stabile, M. R.; Lai, W. G.; DeSantis, G.; Gold, M.; J ones, J . B.;
Mitchinson, C.; Bott, R. R.; Graycar, T. P.; Liu, C.-C. Bioorg. Med. Chem.
Lett. 1996, 6, 2501-2512. (b) Berglund, P.; DeSantis, G.; Stabile, M. R.;
Shang, X.; Gold, M.; Bott, R. R.; Graycar, T. P.; Lau, T. H.; Mitchinson, C.;
J ones, J . B. J . Am. Chem. Soc. 1997, 119, 5265-5266.
(16) Molecular Glycobiology; Fukuda, M., Hindsgaul, O., Eds.; Oxford
University: Oxford, 1994.
(17) Ellman, G. L.; Courtney, K. D.; Andres, V.; Featherstone, R. M.
Biochem. Pharmacol. 1961, 7, 88-95.
10.1021/jo9816461 CCC: $15.00 © 1998 American Chemical Society
Published on Web 12/01/1998