Solid-Phase Enzymatic Synthesis of a Sialyl Lewis X Tetrasaccharide on a Sepharose Matrix

Article abstract of DOI:10.1021/jo980074h

Thiopyridyl Sepharoses with different linker arm lengths were prepared from epoxy Sepharose 6B by reaction first with 1,8-diamino-3,6-dioxaoctane and then with, sucessively, diethoxy-3-cyclobutene-1,2-dione (squaric acid diethyl ester) and 1,8-diamino-3,6-dioxaoctane in several cycles, followed by reaction of the obtained amino Sepharoses with, successively, thiobutyrolactone and 2,2′-dithiopyridine. The thiopyridyl Sepharoses were reacted with the glucosamine derivative 2-(3′-mercaptobutyrylamido)ethyl 2-acetamido-2-deoxy-β-D-glucopyranoside, giving GlcNAc Sepharoses with different linker lengths. Enzymatic galactosylation of these with β-(1-4)-galactosyltransferase and UDP-galactose gave yields varying between 70 and 98%, and there was a clear correlation between linker length and yield. A GlcNAc Sepharose with a long linker was then used in a solid-phase synthesis of a sialyl Le x tetrasaccharide. The three required enzymes (galactosyl-, sialyl, and fucosyltransferase) and nucleotide sugars were reacted consecutively with the GlcNAc Sepharose, giving, after cleavage from Sepharose with DTT, the free sialyl Le x tetrasaccharide derivative in a 57% total yield after purification.

Full text of DOI:10.1021/jo980074h

J . Org. Chem. 1998, 63, 2705-2710
2705
Solid -P h a se En zym a tic Syn th esis of a Sia lyl Lew is X
Tetr a sa cch a r id e on a Sep h a r ose Ma tr ix
O. Blixt and T. Norberg*
Department of Chemistry, Swedish University of Agricultural Sciences, P.O. Box 7015,
S-750 07 Uppsala, Sweden
Received J anuary 15, 1998
Thiopyridyl Sepharoses with different linker arm lengths were prepared from epoxy Sepharose 6B
by reaction first with 1,8-diamino-3,6-dioxaoctane and then with, sucessively, diethoxy-3-cy-
clobutene-1,2-dione (squaric acid diethyl ester) and 1,8-diamino-3,6-dioxaoctane in several cycles,
followed by reaction of the obtained amino Sepharoses with, successively, thiobutyrolactone and
2,2′-dithiopyridine. The thiopyridyl Sepharoses were reacted with the glucosamine derivative 2-(3′-
mercaptobutyrylamido)ethyl 2-acetamido-2-deoxy-â-D-glucopyranoside, giving GlcNAc Sepharoses
with different linker lengths. Enzymatic galactosylation of these with â-(1-4)-galactosyltransferase
and UDP-galactose gave yields varying between 70 and 98%, and there was a clear correlation
between linker length and yield. A GlcNAc Sepharose with a long linker was then used in a solid-
phase synthesis of a sialyl Le x tetrasaccharide. The three required enzymes (galactosyl-, sialyl,
and fucosyltransferase) and nucleotide sugars were reacted consecutively with the GlcNAc
Sepharose, giving, after cleavage from Sepharose with DTT, the free sialyl Le x tetrasaccharide
derivative in a 57% total yield after purification.
In tr od u ction
a sialyl Lewis x derivative in kilogram amounts, using
cloned glycosyltransferases for two crucial glycosylations.
Methods for the synthesis of oligosaccharides have
developed considerably during the last 20 years, but a
chemical oligosaccharide synthesis is still a relatively
complicated task. Some recent attempts toward simpli-
fication have been made by applying chemical solid-phase
techniques,1-4 which are routine in peptide and oligo-
nucleotide synthesis. However, chemical glycosylation
is seldom stereospecific, and consistent high-coupling
yields with respect to one anomer are therefore difficult
to achieve. Furthermore, the elaborate protecting group
patterns that have to be used for each monosaccharide
component make their preparation tedious.
Glycosyltransferase-catalyzed glycosylations in com-
bination with solid-phase techniques should offer a
particulary simple way to synthesize natural oligosac-
charides on a laboratory scale (Scheme 1). Some reports
on such syntheses, using different transferases and solid-
phase materials, have indeed appeared.9-17 Only few,9-10
however, report several consecutive enzymatic glycosy-
lations on the solid phase to produce more complicated
oligosaccharides. We have previously reported18 an
enzymatic solid-phase fucosylation of a disaccharide
reversibly linked to Sepharose via disulfide linkage to
produce a Le a trisaccharide. We now report on the
synthesis of a sialyl Le x tetrasaccharide by carrying out
three consecutive, high-yield enzymatic glycosylations on
a starting glucosamine monosaccharide attached to
Sepharose via a disulfide linkage. In connection with this
synthesis, we have also investigated how the yield of a
typical enzymatic glycosylation varies with the length of
the linker connecting the sugar acceptor to the Sepharose
matrix.
A radically different approach to oligosaccharide syn-
theses is to use enzymes (e.g., glycosyltransferases) as
glycosylating catalysts. These give regio- and stereospe-
cific glycosylations, without protecting groups on the
monosaccharides. Here, progress has been dramatic,5-7
mainly due to advances in enzyme production by genetic
engineering. An example of what can be done with
glycosyltransferases today is the reported8 synthesis of
* To whom correspondence should be addressed. Tel.: +46-18-
671578. Fax: +46-18-673477. E-mail: thomas.norberg@kemi.slu.se.
(1) Danishefsky, S. J .; Mcclure, K. F.; Randolph, J . T.; Ruggeri, R.
B. Science 1993, 260, 1307-1309.
(9) Halcomb, R. L.; Huang, H. M.; Wong, C. H. J . Am. Chem. Soc.
1994, 116, 11315-11322.
(10) Yamada, K.; Nishimura, S. I. Tetrahedron Lett. 1995, 36, 9493-
(2) Rademann, J .; Schmidt, R. R. Tetrahedron Lett 1996, 37, 3989-
9496.
3990.
(11) Meldal, M.; Auzanneau, F.; Hindsgaul, O.; Palcic, M. J . Chem.
Soc., Chem. Commun. 1994, 1849-50.
(3) Randolph, J . T.; Mcclure, K. F.; Danishefsky, S. J . J . Am. Chem.
Soc. 1995, 117, 5712-5719.
(12) Nunez, H. A.; Barker, R. Biochemistry 1980, 19, 489-495.
(13) Schuster, M.; Wang, P.; Paulson, J . C.; Wong, C. H. J . Am.
Chem. Soc. 1994, 116, 1135-1136.
(4) Yan, L.; Taylor, C. M.; Goodnow, R.; Kahne, D. J . Am. Chem.
Soc. 1994, 116, 6953-6954.
(5) Ichikawa, Y.; Look, G. C.; Wong, C. H. Anal. Biochem. 1992, 202,
215-238.
(14) Zehavi, U.; Herchman, M. Carbohydr. Res. 1986, 151, 371-
378.
(6) Wong, C. H.; Halcomb, R. L.; Ichikawa, Y.; Kajimoto, T. Angew.
Chem. Int. Ed. Engl. 1995, 34, 521-546.
(15) Zehavi, U.; Herchman, M.; Hakomori, S.; Ko¨pper, S. Glycocon-
jugate J . 1990, 7, 219-28.
(7) Wong, C. H.; Halcomb, R. L.; Ichikawa, Y.; Kajimoto, T. Angew.
Chem. Int. Ed. Engl. 1995, 34, 412-432.
(16) Zehavi, U.; Herchman, M.; Schmidt, R.; Ba¨r, T. Glycoconjugate
J . 1990, 7, 229-34.
(8) Bayer, R.; DeFrees, S.; Gaeta, F.; J iang, C.; Norberg, T.;
Ramphal, J .; Ratcliffe, M.; Walker, L. Presented at the 2nd CHI
Glycotechnology Conference, LA J olla, CA, 1994.
(17) Zehavi, U.; Herchman, M.; Hakomori, S.; Ko¨pper, S. Glycocon-
jugate J . 1990, 7, 219-28.
(18) Blixt, O.; Norberg, T. J . Carbohydr. Chem. 1997, 16, 143.
S0022-3263(98)00074-7 CCC: $15.00 © 1998 American Chemical Society
Published on Web 03/20/1998

2706 J . Org. Chem., Vol. 63, No. 8, 1998
Sch em e 1. Typ ica l Solid -P h a se En zym a tic Oligosa cch a r id e Syn th esis
Blixt and Norberg
Resu lts a n d Discu ssion
diethyl ester20) and 1,8-diamino-3,6-dioxaoctane. This
gave a second amino Sepharose (33 linker atoms).
Further repetition of this elongation cycle gave Sepharoses
with linker length of 45 and 57 atoms. The obtained
amino Sepharoses were then reacted with, successively,
thiobutyrolactone and 2,2′-dithiopyridine, giving thiopy-
ridyl Sepharoses with linker lengths of 26, 38, 50, and
62 atoms. Coupling (Scheme 3) of these with monosac-
charide 2 gave GlcNAc Sepharoses 4, 5, 6, and 7 with
linker lengths of 35, 47, 59, and 71 atoms, respectively.
For reference purposes, monosaccharide 2 was also
coupled (Scheme 3) with commercial thiopyridyl Sepharose
6B to give GlcNAc Sepharose 3 with 12 linker atoms; this
was the linker used in our previous work. All the
chemical reactions performed on Sepharose proceeded in
excellent yields, as measured by p-toluenesulfonic acid
titration of the gel amino groups after each step in the
elongation cycles. The degree of functionalization on the
gels was 7.5-12.4 µmol/ml drained gel, as measured by
NMR quantification of released sugar after coupling.
In our previous work18 on solid-phase enzymatic syn-
thesis, a disaccharide (Gal1-3â-GlcNAc) acceptor was
used, linked to Sepharose via a disulfide linkage and a
linker arm with a total length of 12 atoms. The yield in
the enzymatic fucosylation seldom exceeded 70%, despite
attempts with increases in enzyme and nucleotide sugar
amounts, or reaction time. Even washing of the gel after
coupling and repeated coupling with fresh enzyme and
nucleotide sugar (“double-coupling”) did not lead to better
yields. Obviously, there were unreactive acceptor sites
present on the gel. This is a well-known phenomenon
in chemical solid-phase synthesis, caused by, e.g., capping
by an undesired chemical group during the coupling
procedure or unfavorable conformational changes when
the polymer chain length increases (steric factors). In
our case, it was reasonable to assume that the low
reactivity was caused not by capping but rather by steric
factors. The relatively short length of the linker (12
atoms, approximately 15 Å in the most extended confor-
mation) connecting the acceptor disaccharide to the
Sepharose matrix and the size of the fucosyltransferase
enzyme (diameter approximately 60 Å, assuming a
globular form) should make at least some acceptor sites
“unapproachable” by the enzyme. A longer linker be-
tween the Sepharose matrix and the acceptor sugar
should give a better result because of less steric interfer-
ence between enzyme and solid phase and also because
of greater conformational flexibility. To test this hypoth-
esis, we prepared Sepharose gels with different lengths
of the linker. We chose a PEG-type linker and 1,4-â-
galactosylation with galactosyltransferase of an N-acetyl-
glucosamine acceptor as a model reaction. To prepare
the N-acetylglucosamine acceptor, 2-azidoethyl 3,4,6-tri-
O-acetyl-2-acetamido-2-deoxy-â-D-glucopyranoside19 was
deacetylated (sodium methoxide) and hydrogenated (Pd/
C) to give 2-aminoethyl 2-acetamido-2-deoxy-â-D-glucopy-
ranoside (1), which was further reacted with thiobuty-
rolactone to give a monosaccharide (2) with a thiol group
on the aglycon. This monosaccharide was then coupled
to several thiopyridyl Sepharoses having linkers differing
in length. The thiopyridyl Sepharoses were prepared in
the following way (Scheme 2):
The GlcNAc Sepharoses 3-7 were next subjected to
enzymatic galactosylation using a total of 3 equiv of
UDP-galactose and 1 U of â-1,4-galactosyltransferase/
10 µmol gel and a total reaction time of 90 h (Scheme 4).
After galactosylation, the gels were treated with dithio-
threitol (DTT) to release the bound carbohydrates into
solution, and the obtained solutions were mixed with a
1
known amount of internal standard and analyzed by H
NMR spectroscopy. The results are shown in Table 1.
In all cases, the recovery of material (monosaccharide
plus disaccharide) exceeded 95%. The gel with the
shortest linker (3, 12 atoms) gave a moderate galactosy-
lation yield (70%), in accordance with the results obtained
in our previous work18 using the same linker in an
enzymatic fucosylation reaction. The other gels gave
significantly better yields; thus, there was a significant
increase in yield (70-89%) on going from gel 3 (12 linker
atoms) to gel 4 (35 linker atoms). The best yield (98%)
was obtained with the gel (7) with the longest (71 linker
atoms) linker; however, the increase in yield is less
significant here, and the variations between gel 5-6 and
7 are within the limits of the estimated experimental
error ((5%). The reason for the better yield obtained on
changing the linker can only be speculated at this point.
Other laboratories have studied the influence on solid-
phase synthesis yields of PEG-type linkers grafted to
polyamide11 or polystyrene,22,23 and the general observa-
tion, at least in regard to the yields in chemical synthesis,
First, epoxy Sepharose 6B was reacted with 1,8-
diamino-3,6-dioxaoctane; this gave the first (21 linker
atoms) of a series of amino Sepharoses. Next, a portion
of this first amino Sepharose was reacted with, succes-
sively, 3,4-diethoxy-3-cyclobutene-1,2-dione (squaric acid
(19) Eklind, K.; Gustafsson, R.; Tiden, A. K.; Norberg, T.; Aberg, P.
M. Carbohydr. Chem. 1996, 15 (9), 1161-1178.
(20) Tietze, L.; Arlt, M.; Beller, M.; Glusenkamp, K. H.; J ahde, E.;
Rajewsky, M. F. Chem. Ber. 1991, 124, 1215.

Synthesis of a Sialyl Lewis X Tetrasaccharide
J . Org. Chem., Vol. 63, No. 8, 1998 2707
Sch em e 2. P r ep a r a tion of Dith iop yr id yl Sep h a r oses w ith Differ en t Lin k er Len gth s
is that the presence of such a linker is beneficial (which
is also evidenced by the fact that PEG-modified resins
have become commercially available).
Sch em e 3. Atta ch m en t of Mon osa cch a r id e 2 to
Sep h a r oses w ith Lin k er s of Differ en t Len gth
Whatever the cause is of the yield improvement in our
case, it can be concluded that the yields now achieved
(90-98%) are in the same high range as is common in
repetetive chemical solid-phase synthesis. With such
yields, it should be possible to construct more complicated
oligosaccharide structures by carrying out several con-
secutive enzymatic glycosylations on the solid phase,
without losing much material or ending up with insepa-
rable mixtures at the end of the synthesis. We decided
to use gel 7 to synthesize the sialyl Lewis x tetrasaccha-
ride 8, since this is a biologically interesting molecule,
and because of the fact that all three enzymes needed
are reasonably accessible. The synthesis was carried out
as shown in Scheme 5.
(21) Lin, Y. C.; Hummel, C. W.; Huang, D. H.; Ichikawa, Y.;
Nicolaou, K. C.; Wong, C. H. J . Am. Chem. Soc. 1992, 114, 5452-5454.
(22) Vagner, J .; Barany, G.; Lam, K. S.; Krchnak, V.; Sepetov, N.
F.; Ostrem, J . A.; Strop, P.; Lebl, M. Proc. Natl. Acad. Sci. U.S.A. 1996,
93, 8194-8199.
(23) McGuinness, B. F.; Kates, S. A.; Griffin, G. W.; Herman, L. W.;
N. A. Sole, Vagner, J .; Albericio, F.; Barany, G. Proc. Am. Pept. Symp.,
14th 1995.

2708 J . Org. Chem., Vol. 63, No. 8, 1998
Blixt and Norberg
Sch em e 4. Ga la ctosyla tion of N-Acetylglu cosa m in e Sep h a r oses 3-7 a n d Relea se of th e Atta ch ed Su ga r
Ta ble 1. En zym a tic Ga la ctosyla tion of Gels w ith
Differ en t Lin k er Len gth s
were carried out in small, silanized (by brief treatment with
5% dichlorodimethylsilane in hexane) columns (5-15 mL) with
a fritted glass filter and a Teflon stopcock at the bottom end.
After being charged with the appropriate reagents, the column
was sealed at the top with a ground glass stopper, the Teflon
stopcock was closed, and the column was slowly rotated in a
37 °C water bath for the specified time. Epoxy Sepharose 6B
and thiopyridyl Sepharose 6B were from Pharmacia Biotech-
nology AB (Uppsala, Sweden). â-(1-4)-Galactosyltransferase,
1,2-O-isopropylidene-R-D-glucofuranose, dithiothreitol (DTT),
uridine 5′-diphosphogalactose, calf intestine alkaline phos-
phatase (CIAP), R-lactalbumin, and cytidine monophospho-N-
acetylneraminic acid (CMP-NeuAc) were from Sigma Chemical
Co. (St. Louis, MO), and 2,2′-dithiopyridine, 3,4-diethoxy-3-
cyclobutene-1,2-dione, and 1.8-diamino-3.6-dioxaoctane were
from Acros Chimica (Geel, Belgium). Recombinant R-(2-3)-
sialyltransferase was a gift from Cytel Corp., San Diego, CA.
Milk R-(1-3/4)-fucosyltransferase and guanosine 5′-(â-L-fu-
copyranosyl)diphosphate were prepared as previously de-
scribed.18 Buffer A had the following composition: 25 mM
sodium cacodylate, 5 mM manganese(II) chloride, 0.1% Tri-
tonX, pH 7.5. Buffer B had the following composition: 10 mM
sodium cacodylate, 5 mM manganese(II) chloride, 2% glycerol,
pH 6.5.
2-(3′-Mer ca p tobu tyr yla m id o)eth yl 2-Aceta m id o-2-d e-
oxy-â-D-glu copyr an oside (2). 2-Azidoethyl 3,4,6-tri-O-acetyl-
2-acetamido-2-deoxy-â-D-glucopyranoside19 (360 mg, 0.87 mmol)
in methanolic sodium methoxide (3 mL, 0.1 M) was stirred at
room temperature until TLC indicated complete reaction, and
then the mixture was neutralized with Dowex-50 (H+) ion-
exchange resin and concentrated. The residue was taken up
in ethanol/water/acetic acid (5:5:1, 11 mL), Pd/C (10%, 0.3 g)
was added, and the mixture was hydrogenated at atmospheric
pressure. When ready (TLC: ethyl acetate/methanol/acetic
acid/water, 3:3:3:1), the mixture was filtered, concentrated, and
lyophilized to give 2-aminoethyl 2-acetamido-2-deoxy-â-D-
glucopyranoside (1, acetate form, 280 mg, 0.87 mmol, 98%).
This material was dissolved in a mixture of sodium bicarbonate
(0.5 M, 5 mL) and ethanol (4 mL), and then γ-thiobutyrolactone
(0.75 mL, 8.6 mmol) and DTT (0.67 g, 4.3 mmol) were added
and the mixture was stirred at 50 °C under a nitrogen
atmosphere overnight (TLC: ethyl acetate/methanol/acetic
acid/water, 3:3:3:1 by volume). The mixture was then neu-
tralizied to pH 7 with 1 M HCl and concentrated, and the
residue was dissolved in water (3 mL), filtered, and applied
onto a BioGel P-2 column (2.5 × 75 cm, packed and eluted
with 5% 1-butanol in water). The appropriate fractions were
pooled and lyophilized to give 2 (227 mg, 0.62 mmol, 71%).
NMR data: 1H, δ 1.88 (quintet, 2H, CH2CH2CH2SH), 2.04 (s,
3H, CH3CONH), 2.37 (t, 2H, CH2CH2CH2SH), 2.55 (t, 2H,
CH2CH2CH2SH), 3.37 (t, 2H, OCH2CH2NH), 3.33-3.50 (m, 2H,
H-4, H-5), 3.55 (t, 1H, H-3), 3.66-3.79, 3.87-3.96 (2 m, 1H
each, OCH2CH2NH), 3.69 (1H, H-2), 3.66-3.79, 3.87-3.96 (2
m, 1H each, H-6a,b), 4.53 (d, 1H, 8.3 Hz, H-1); 13C, δ 22.9,
23.5, 29.8, 34.8 (CH3CONH-, -CH2CH2CH2-), 39.7 (OCH2-
CH2NH-), 56.0 (C-2), 61.2 (C-6), 68.5 (OCH2CH2NH-), 74.2,
70.3, 76.3 (C-3, C-4, C-5), 101.5 (C-1), 175.0 (CH3CONH-),
176.6 (-NHCOCH2-); HRMS calcd for C14H26O7N2SNa
389.1358, found 389.1410 (M + Na).
linker
no. atoms
bound GlcNAc
(µmol/mL)
enzymatic
gel
gal. transfera (%)
3
4
5
6
7
12
35
47
59
71
7.5
12.4
12.4
12.4
9.0
70
89
89
90
98
By 1H NMR (Gal/GlcNAc H-1). All experiments at 1.0 mL
drained gel volume.
a
GlcNAc Sepharose 7 (24 µmol), was incubated with
â-(1-4)-galactosyltransferase and UDP galactose over an
incubation period of 4 days, with succesive additions at
intervals of UDP galactose (a total of 3 equiv). This,
according to NMR analysis of the mixture released from
an aliquot of the gel, gave a 90% galactosylation yield.
Washing of the gel and subsequent sialylation with R-(2-
3)-sialyltransferase and CMP-Neu5Ac using similar
amounts of enzyme and nucleotide sugar gave a 75%
glycosylation yield; the lower yield here was probably due
to nonoptimized conditions. Washing of the gel and
subsequent fucosylation with partially purified R-(1-3/
4)-fucosyltransferase from human milk and GDP-fucose
gave an approximately 95% fucosylation yield. Finally,
treatment of the gel with DTT and purification of the
released mixture by gel filtration gave the sialyl Le x
tetrasaccharide (8) in a 57% yield (calculated from the
amount of starting GlcNAc on the gel). The purity of the
material in terms of tetrasaccharide content was better
than 90%. The ratio of thiol to disulfide form of the
product varied, the latter dominating after exposure to
air for an extended time. The thiol form could be
recovered quantitatively from any thiol-disulfide mixture
by brief treatment with dithiothreitol.
In conclusion, using GlcNAc linked via a new linker
to Sepharose, the sialyl Le x tetrasaccharide 8 was
synthesized in good (57%) yield by sequential treatment
of the Sepharose gel with the three necessary enzymes
and their corresponding monosaccharide nucleotide sug-
ars. Further investigations with other sugars and trans-
ferases using the described solid-phase technique are in
progress.
Exp er im en ta l Section
Gen er a l Meth od s. Concentrations were performed under
reduced pressure at <40 °C bath temperature. NMR spectra
were recorded at 30 °C using Bruker 400 and 600 MHz
instruments. The following reference signals were used:
acetone δ 2.225 (1H in D2O), acetone δ 30.7 (13C in D2O). FAB
MS was recorded with a J EOL J MS-SX/SX-102A mass spec-
trometer using thioglycerol as matrix. Thin-layer chromatog-
raphy was performed on Kieselgel 60 F254 Fertigplatten
(Merck, Darmstadt, Germany). After elution with appropriate
eluants, spots were visualized by UV light and/or by dipping
in 5% sulfuric acid, followed by charring. Water for all
solutions was from a MilliQ water purification system (Milli-
pore Corp., Bedford, MA) and was degassed by vacuum
treatment before use. The solid-phase enzymatic reactions
Solid -P h a se P r ep a r a tion s. Dith iop yr id yl Sep h a r oses
6B w ith Exten d ed Lin k er Ar m : (1) Am in o Der iva tiza -
tion of Ep oxy Sep h a r ose 6B. Dry epoxy Sepharose 6B (7
g) was swollen in water, washed (water, 200 mL), transferred
to a glass filter column (2.5 × 40 cm), and drained. A solution
of 1.8-diamino-3,6-dioxaoctane (3.0 mL, 4.4 g, 30 mmol) in
aqueous sodium bicarbonate buffer (0.1 M, pH 11.0, 50 mL)

Synthesis of a Sialyl Lewis X Tetrasaccharide
J . Org. Chem., Vol. 63, No. 8, 1998 2709
Sch em e 5. Solid -P h a se En zym a tic Syn th esis of a Lew is x Tetr a sa cch a r id e
was added, and the column was sealed and rotated for 48 h at
room temperature. The resin was transferred to a glass filter
funnel, filtrated, and washed with water (4 L). An aliquot of
the gel (0.25 mL, drained) was reacted with aqueous p-
toluenesulfonic acid (1mM, 5.0 mL) for 10 min with gentle
stirring. The absorbance at 264 nm of the supernatant was
recorded before and after gel treatment, and from this differ-
ence the amount of amino groups on the gel was calculated to
be approximately 14 µmol/mL drained gel.
(2) Am in o-Lin k er Elon ga tion Cycle. A portion of the
amino-derivatized gel (2.0 mL drained) was set aside, and the
rest of the material (approximately 23 mL drained gel) was
reacted with 3,4-diethoxy-3-cyclobutene-1,2-dione (1.3 mL, 1.5
g, 8.8 mmol) dissolved in a mixture of ethanol (95%, 50 mL)
and bicarbonate buffer (0.1M, pH 9.5, 50 mL) and rotated for
6 h at room temperature. The gel was filtered off and washed
with ethanol (50 mL) and water (4 L). An aliquot of the gel
(0.25 mL, drained) was reacted with p-toluenesulfonic acid as
described above. The absorbance of the supernatant at 264
nm before and after reaction with resin did not differ notably,
indicating complete disappearance of amino groups on the gel.
The gel was mixed with a solution of 1.8-diamino 3,6-
dioxaoctane (3.0 mL) in carbonate buffer (0.1 M, 50 mL, pH
9.5) rotated for 24 h at room temperature, and then filtered
and washed with water (4 L). An aliquot of the gel (0.25 mL,
drained) was reacted with p-toluenesulfonic acid as described
above. The absorbance at 264 nm of the supernatant was
recorded before and after gel treatment; from this difference
the amount of amino groups on the gel was calculated to be
approximately 14 µmol/mL drained gel.
A portion of this amino-derivatized gel (2.0 mL drained) was
set aside, and the rest of the material was subjected to two
more elongation cycles as described above. For each elonga-
tion, the amount of amino groups on the gel, as measured by
p-toluenesulfonic acid titration, did not decrease significantly.
(3) Con ver sion of Gel Am in o Gr ou p s to Dith iop yr id yl
Gr ou p s. Amino-derivatized gels from the above-described
preparations were each subjected to the following procedure:
Drained gel (2 mL) was mixed, in a glass column (1.8 × 10
cm), with a solution of DTT (62 mg) and thiobutyrolactone (70
µL) in ethanol (95%, 4.5 mL) and aqueous sodium bicarbonate
(0.5 M, 5.0 mL). The column was rotated in a water bath at
50 °C for 20 h, and then the gel was washed with water (200
mL), and treated with phosphate buffer (0.1 M, pH 8.0, 10 mL),
followed by treatment with DTT (77 mg) in phospate buffer
(0.1 M, 10 mL) for 2 h. After being washed with water (200
mL), the gel was rotated overnight at room temperature with
a solution of 2,2′-dithiopyridine (35 mg) in phosphate buffer
(0.1 M, pH 8.0, 5.5 mL) and ethanol (2.5 mL). The absorbance
at 343 nm (proportional to released thiopyridone) was deter-
mined for the supernatant, giving indirect estimates of dithi-
opyridyl substitution degrees (in order of increasing linker
length) of 14, 13, 14, and 9 µmol/mL drained gel. The gels
were filtered off and washed with water (200 mL) and were
stored until used under 30% ethanol at 4 °C.
buffer (0.2 M, pH 4.0, 7 mL) was added to the Sepharose gels
obtained as described above (conditioned in citric acid buffer),
and the mixtures were rotated in columns overnight at room
temperature. The amount of released thiopyridone (as deter-
mined by supernatant absorbance at 343 nm) indicated
substitution degrees of 13, 12, 12, and 10 µmol/drained gels
for 4, 5, 6, and 7, respectively. The gels were filtered and
washed with water (200 mL), and then three bed volumes of
a 50 mM solution of mercaptoethanol in citric acid buffer (0.2
M, pH 4.0) was added to the gels and the columns were rotated
for 2 h at room temperature, followed by filtering and washing
the gels with water (200 mL). The amount of bound sugar
was determined by treating gel (0.5 mL) with DTT in phos-
phate buffer, as described previously, and quantifying the
released sugar with 1H NMR against an added internal
standard. This gave values of 12.4, 12.4, 12.4, and 9.0 for gels
4, 5, 6, and 7, respectively, in good agreement with the values
obtained above with spectrophotometry.
Using the same procedure as above, commercial thiopyridyl
Sepharose 6B was reacted with 2 to give gel 3, with a
substitution degree of 7.5 µmol/drained gel.
Gen er a l E n zym a t ic Ga la ct osyla t ion P r oced u r e.
Drained gel (3-7, 1 mL) conditioned in buffer (50 mM
cacodylate, 5mM MnCl2, 0.1% TritonX, pH 7.5, 5 mL) was
transferred to a column (1 × 5 cm), buffer (3 mL), â-(1-4)-
galactosyltransferase (1U), CIAP (4U), R-lactalbumin (0.6 mg),
and UDP-galactose (1 equiv) were added, and the column was
rotated in a water bath at 37 °C. After 6, 18, 24, and 48 h
more UDP-galactose (0.5 equiv) was added. Incubation was
continued for 90 h, and the gel was then filtered and washed
with water (200 mL). To release the carbohydrates, the gels
were treated for 2 h at room temperature with DTT (23 mg)
in phosphate buffer (0.1 M, pH 8.0, 3 mL) containing 5 µmol
of 1,2-O-isopropylidene-R-D-glucofuranose (as internal stan-
dard). After filtration and washing with water (10 mL), the
filtrate was lyophilized and the amount of galactosylation was
1
determined by H NMR (Table 1).
Sia lyl Lew is x Der iva tive 8 a n d Its Disu lfid e Isom er .
GlcNAc Sepharose derivative 7 (2.7 mL, 24 µmol) was trans-
ferred to a 15 mL glass column, washed with water (200 mL),
conditioned with buffer A, and drained. Buffer A (8.0 mL) was
added, followed by UDP-galactose (17 mg, 28 µmol, dissolved
in 0.4 mL of buffer A), â-(1-4)-galactosyltransferase (5 U,
dissolved in 0.8 mL of 1:1 buffer A/glycerol), R-lactalbumin (2.7
mg), and CIAP (25 mg, 30 U), and the pH was adjusted to
7.35. The mixture was rotated in a water bath at 37 °C. After
6, 24, and 48 h, new additions of UDP-galactose (17, 13, and
10 mg, respectively) were made, and the pH was adjusted to
to 7.35 with 1 M aqueous cacodylic acid. After 5 days, the gel
was filtered off and washed with water (300 mL). An aliquot
(0.50 mL, 4.5 µmol) was taken out, washed with phosphate
buffer (0.1 M, pH 8.0, 10 mL), drained, and mixed with DTT
(11 mg) in phosphate buffer (1.5 mL) containing 1,2-O-
isopropylidene-R-D-glucofuranose (as internal standard, 5.0
µmol). After rotation for 2 h at room temperature, the gel was
filtered off and washed with water (10 mL). The filtrate was
Rea ction of Am in o Gels w ith th e GlcNAc Der iva tive
2 To Give Gels 3-7. GlcNAc thiol 2 (3 equiv) in citric acid
1
lyophilized, and the H NMR spectrum was recorded. Integra-

2710 J . Org. Chem., Vol. 63, No. 8, 1998
Ta ble 2. NMR Da ta for 8 a n d Its Disu lfid e Isom er a
Blixt and Norberg
CH2 CH2
X
1
Neu5Ac
Gal
4.54
101.9
3.56
69.0
4.08
75.4
3.94
67.0
3.59
75.0
3.71
61.5
GlcNAc
Fuc
CH2
CH2
CH2
4.51
102.0
3.90
56.0
3.84
75.0
3.93
73.5
3.59
75.1
4.01
59.0
5.09
99.4
3.69
68.5
3.90
69.0
3.78
72.0
4.81
67.0
1.16
14.5
3.66-3.96
3.37
39.7
2.37
34.6
1.99 (1.88)
24.6 (29.6)
2.75 (2.55)
37.3 (22.2)
68.5
2
3
1.79, 2.75
39.8, 37.3
3.70
4
68.2
5
3.84
52.0
6
3.66
73.4
7
3.58
68.0
8
3.90
72.1
9
3.65
62.9
CH3
2.04
21.1
2.02
21.1
CdO
175.0
spacer-NHCOCH2- 176.6
a
CH2 groups in order of increasing distance from the GlcNAc C-1. Numbers in parentheses refer to the thiol form, if different from the
disulphide form.
tion of appropriate analyte and internal standard peaks gave
a recovery estimate of >95%, and a di-monosaccharide ratio
of 90%.
yield. The crude reaction mixture was dissolved in 3.0 mL of
20 mM ammonium formate buffer (pH 4.7) and loaded on a
BioGel P2 column (1.8 × 100 cm), conditioned in 20 mM
ammonium formate buffer (pH 4.7) containing 5% n-BuOH.
Elution with the same buffer gave, after pooling of appropriate
fractions and lyophilization, tetrasaccharide derivative 8 (8.2
mg, 8.5 µmol, 57% total yield) as a white fluffy solid. NMR
data are given in Table 2. For all assigned sugar atoms, there
was excellent agreement with previously published data on
sialyl Le x tetrasaccharide â methyl glycoside.21 The propor-
tion between 8 and its disulfide form in the NMR sample was
estimated to be 1:9, from the relative intensities of the
respective CH2S methylene signals. The FAB positive-ion
mass spectrum showed a peak at m/z 966, corresponding to
M + H for 8, and a peak at m/z 1929, corresponding to M + H
of the disulfide isomer of 8 (proportions varied between runs).
The bulk part of the gel was conditioned with buffer A and
drained, and then buffer A (5 mL), R-(2-3)-sialyltransferase
(3.5 U), CMP-Neu5Ac (15 mg, 1.2 equiv, 24 µmol, dissolved in
1 mL of buffer A), and CIAP (15 mg, 18 U) were added. The
pH was adjusted to 7.45. The mixture was rotated in a water
bath at 37 °C. After 6, 24, and 48 h new additions (0.5 equiv
each) of CMP-Neu5Ac were made, and pH was adjusted to 7.35
by addition of 1 M aqueous cacodylic acid. After 5 days, the
gel was washed with water (300 mL). An aliquot (0.50 mL)
was taken out and was treated and analyzed with NMR as
described above; this indicated a >95% recovery and a 75%
sialylation yield.
The bulk part of the gel was finally fucosylated with R-(1-
3/4)-fucosyltransferase from human milk (0.2 U in buffer B,
11 mL), GDP-fucose (15 mg, 1.4 equiv, 24 µmol), and CIAP
(4.0 mg, 5U); and the pH was 6.5. The mixture was rotated
in a water bath at 37 °C. After 15, 26, 42, 62, and 84 h, new
additions of GDP-fucose (15, 10, 8, 8, and 8 mg, respectively)
were made, and the pH was adjusted to 6.5 with 1 M aqueous
cacodylic acid. After 5 days, the gel was washed with water
(300 mL) and 1 M NaCl (50 mL), and then treated with DTT
(77 mg) in phosphate buffer (0.1 M, pH 8.0, 10 mL) containing
internal standard (10 µmol) and analyzed by NMR. This
analysis indicated a >95% recovery and a 95% fucosylation
Ack n ow led gm en t. We thank the Swedish Natural
Science Research Council and the Swedish Research
Council for Engineering Sciences for financial support.
We also thank Pharmacia Biotechnology (Uppsala,
Sweden) for supplying the Sepharose, and Cytel Cor-
poration (San Diego, CA) for a supplying the R-(2,3)-
sialyltransferase.
J O980074H