Chemo-enzymatic synthesis of conformationally constrained oligosaccharides.

Article abstract of DOI:10.1039/b308559g

N-Acetyllactosamine derivative 4, which has a methylene amide tether between C-6 and C-2', was enzymatically glycosylated using rat liver alpha-2,6-sialyltransferase (ST6GalI) or recombinant human fucosyltransferase V (FucT-V) to give conformationally constrained trisaccharides 5 and 6, respectively. The methylene amide linker of 4 was installed by a two-step procedure, which involved acylation of a C-6 amino function of a LacNAc derivative with chloroacetic anhydride followed by macrocyclization by nucleophilic displacement of the chloride by a C-2' hydroxyl. The conformational properties of 4 were determined by a combination of NOE and trans-glycosidic heteronuclear coupling constant measurements and molecular mechanics simulations and these studies established that the glycosidic linkage of 4 is conformationally constrained and resides in only one of the several energy minima accessible to LacNAc. The apparent kinetic parameters of transfer to LacNAc and conformationally constrained saccharides 3 and 4 indicates that fucosyltransferase V recognize LacNAc in its A-conformer whereas alpha-2,6-sialyltransferase recognizes the B-conformer of LacNAc.

Full text of DOI:10.1039/b308559g

View Article Online / Journal Homepage / Table of Contents for this issue
Chemo-enzymatic synthesis of conformationally constrained
oligosaccharides†
M. Carmen Galan,a Andre P. Venot,a John Glushka,a Anne Imbertyb and Geert-Jan Boons*a
a Complex Carbohydrate Research Center, University of Georgia, 220 Riverbend Road, Athens,
GA 30602, USA. E-mail: gjboons@ccrc.uga.edu; Fax: ϩ1 706/5424412;
Tel: ϩ1 706-542-9161
b CERMAV-CNRS (affiliated to Université Joseph Fourier), BP 53, 38041 Grenoble, cedex 9,
France
Received 24th July 2003, Accepted 18th September 2003
First published as an Advance Article on the web 14th October 2003
N-Acetyllactosamine derivative 4, which has a methylene amide tether between C-6 and C-2Ј, was enzymatically
glycosylated using rat liver α-2,6-sialyltransferase (ST6GalI) or recombinant human fucosyltansferase V (FucT-V)
to give conformationally constrained trisaccharides 5 and 6, respectively. The methylene amide linker of 4 was
installed by a two-step procedure, which involved acylation of a C-6 amino function of a LacNAc derivative
with chloroacetic anhydride followed by macrocyclization by nucleophilic displacement of the chloride by a C-2Ј
hydroxyl. The conformational properties of 4 were determined by a combination of NOE and trans-glycosidic
heteronuclear coupling constant measurements and molecular mechanics simulations and these studies established
that the glycosidic linkage of 4 is conformationally constrained and resides in only one of the several energy
minima accessible to LacNAc. The apparent kinetic parameters of transfer to LacNAc and conformationally
constrained saccharides 3 and 4 indicates that fucosyltransferase V recognize LacNAc in its A-conformer
whereas α-2,6-sialyltransferase recongizes the B-conformer of LacNAc.
Little is known about the effect of inter-residual flexibility of
Introduction
glycosyl acceptors on kinetic parameters of glycosyl transfer.
Furthermore, structural studies have provided little information
about the conformation of a glycosyl acceptor that is recog-
nized by a glycosyltransferase. To address these issues, we
recently reported the chemical synthesis of conformationally
constrained N-acetyl lactosamine (LacNAc) derivatives 1–3
and determined apparent kinetic parameters of sialylation by
rat liver α(2–6)-sialyltransferase (Fig. 1).8 The conformational
properties of 1–3 were determined by a combination of molecu-
lar mechanics calculations and NMR spectroscopy. It was
found that compound 3 adopts a low energy conformations
that are centered about the A- and B-conformers of LacNAc
whereas 1 and 2 adopt conformational spaces outside these
regions (Fig. 1). Compound 3 was by far the best acceptor
tested indicating that the enzyme recognizes LacNAc in one of
the two low energy conformations.
In order to determine in more detail the conformational
requirements of glycosyltransferases, conformational con-
strained analogs of LacNAc are required that are more rigid
than 3 but still can adopt a low energy conformation. Further-
more, it is important to investigate whether such derivatives can
be used as substrates for the enzymatic synthesis of more com-
plex conformationally constrained oligosaccharides. It is to be
expected that the resulting compounds will be important
probes to determine in which conformation lectins or other
glycosyltransferases recognize their substrates. In this respect,
very few conformationally constrained saccharides have
been reported,8–16 which has made it difficult to establish the
effect of conformationally flexibility on protein-carbohydrate
recognition.
The cornerstone of biological advances has been the under-
standing of the biological interactions of nucleic acids and pro-
teins between either themselves or each other. However, in
recent years, the recognition roles played by carbohydrates and
glycoconjugates have received considerable attention. Carb-
ohydrate recognition has been implicated in fertilization,
embryogenesis, neuronal development, hormonal activities, cell
proliferation and their organization into specific tissues.1,2 As
many of these processes involve the interaction between pro-
teins and oligosaccharides, an understanding of the factors that
control these interactions at the molecular level is of prime
importance.
It is well established now that the majority of oligo-
saccharides have some degree of flexibility around their glyco-
sidic linkages and as a result can adopt several distinct low
energy minima on their conformational surface.3–5 Further-
more, there is evidence to suggest that the conformational
properties of a glycosidic linkage can be modulated by the
macromolecular structure it is part of and in particular flanking
oligosaccharide residues or attachment to proteins and lipids
may result in conformational changes.4,6 In this respect, the
examination of structural properties of a wide range of glyco-
proteins has revealed that glycosidic torsional angles may devi-
ate as much as 20–30Њ from minimum energy conformations.7 In
the binding site of proteins, not only global but also secondary
minima with very different conformations can be complexed.4
† Electronic supplementary information (ESI) available: Region of
1
500 MHz H{13C}-HSQC spectrum of compound 4 in D2O at 25 ЊC.
Here we report the chemical synthesis of a LacNAc deriv-
ative (4), which is preorganized by a methylene amide linker
between the C-6 and C-2Ј. A combination of NOE and trans-
glycosidic heteronuclear coupling constant measurements and
molecular mechanics simulations established that the glycosidic
linkage of compound 4 is highly conformationally constrained
and resides in the B-conformer on the energy map of LacNAc.
The apparent kinetic parameters for the α-(2,6)-sialyltrans-
600 MHz proton spectra of compound 4 in water and in 21% pyridine.
Regions of 800 MHz ROESY spectrum of compound 4 in 10%
DMSO–pyridine, showing some observed NOEs. Characteristics of the
conformational families that can be adopted by compound 4. Traces of
3
the 2D HSQMBC spectrum for the determination of JC1Ј,H4 for com-
pound 4. Nomenclature and detailed geometrical description of the low
energy conformation of compound 4. 1D NMR spectra for compounds
4–6. Procedures for enzyme kinetics. See http://www.rsc.org/suppdata/
ob/b3/b308559g/
T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 3
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 8 9 1 – 3 8 9 9
3891

View Article Online
Fig. 1 Energy map of LacNAc glycosidic linkage17 calculated with MM3. Conformations accessible to compounds 1–3 have been superimposed as
᭿ = 1, ꢀ = 2, ᭹ = 38
ferase and fucosyltransferase V catalyzed transformations indi-
cate that the two enzymes respond differently to imposed con-
formational constraints of the acceptor and it appears that the
fuocylstransferase recognizes LacNAc in the A conformation
whereas the sialyltransferase prefers the B-conformer. Com-
pound 4 was used as a substrate in the enzymatic synthesis
of trisaccharides 5 and 6 (Fig. 2) These two compounds
are analogs of 6Ј-sialyllactosamine (αNeu5Ac-(2–6)βGal(1–4)-
βGlcNAc-OMe) and Lewisx (βGal(1–4)[αFuc-(1–3)]βGlcNAc-
OMe), which are involved in a wide range of biological
processes and for example 6Ј-sialyllactosamine is a ligand
for influenza virus18 whereas Lewisx is a tumor-associated
antigen.19
linker of 4 was installed by a two-step procedure, which
involved acylation of the C-6 amino function of a LacNAc
derivative 10 with chloroacetic anhydride followed by macro-
cylization by nucleophilic displacement of the chloride by a
C-2Ј hydroxyl (Scheme 1). Enzymatic sialylation or fucosyl-
ation of 4 using rat liver α-2,6-sialyltransferase (ST6GalI) or
recombinant α-1,3-human fucosyltansferase V (FucT-V) pro-
vided preorganized 5 and 6, respectively. Previous studies have
shown that these transferases allow modifications of C-6 and
C-2Ј hydroxyls of LacNAc and therefore these positions are
appropriate for installing a linker.20,21
Fig. 2 Preorganized compounds 4–6.
Results and discussion
Scheme 1 Reagents and conditions: (i) NaN3, DMF; (ii) NaOMe,
MeOH; (iii) HS(CH2)3SH, H2O–pyridine; (iv) (ClCH2CO)2O, MeOH–
DCM; (v) NaH, DMF; (vi) Pd/C, H2.
Synthesis
It was envisaged that compound 4, which is modified by the
methylene amide linker, should be more rigid than the corre-
sponding ethylene analogue 3, which is conformationally con-
straint by a more flexible ethylene tether. The methylene amide
Key intermediated 10 could easily be prepared starting
from previously reported LacNAc derivative 7,8 which has a
methanesulfonyl ester at C-6 and an acetyl ester at C-2Ј
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 8 9 1 – 3 8 9 9
3892

View Article Online
Table 1 Comparison of experimentala and theoretical NMR values of optimized LacNAc and LacNAc-derivative 4
ΦH/Њ
JH1Ј,C4/Hz
ΨH/Њ
JH4,C1Ј/Hz
H1Ј–H4/Å
H1Ј–H6R/Å
H1Ј–H6S/Å
H1Ј–H3/Å
LacNAc-OMe
conf_A
conf_B
conf_C
conf_D
av. Conf
Exp.
45
28
43
2.9
4.4
3.1
6.8
3.7
3.8b
Ϫ5
Ϫ54
Ϫ174
Ϫ2.6
5.5
2.1
6.7
5.6
3.9
4.6b
2.2
2.4
3.7
3.6
2.3
2.4
2.8
4.1
5.2
4.5
3.1
2.9
3.2
4.5
4.1
5.4
3.5
3.0
176
35
Ϫ25
Compound 4
Trans_ap
Trans_p
Trans_p_2
Cis_p
47
43
54
28
Ϫ1
27
28
2.8
3.1
2.1
4.4
5.6
4.5
4.4
Ϫ19
Ϫ23
Ϫ5
5.0
4.8
5.6
5.5
4.1
1.4
1.5
2.3
2.3
2.4
2.1
2.1
2.4
2.3
4.7
4.3
4.4
4.4
3.8
3.1
3.0
Ϫ6
Cis_ap
Ϫ32
Ϫ63
Ϫ61
Trans_pc
Exp.b
a Exp. J values 15%; exp. NOE values 0.3 Å. b Obtained in 21% pyridine–D2O. c Distance optimized.
(Scheme 1). The key disaccharide 78 was converted into the
azido derivative 8 by displacement of the methanesulfonyl ester
with sodium azide in DMF. After purification by silica gel
column chromatography, 8 was isolated in a yield of 98%.
Deacetylation of 8 under standard conditions gave the desired
alcohol 9 in almost quantitative yield. Reduction of the azido
group of 9 with 1,3-propanedithiol in a mixture of pyridine and
water22 gave the desired free amine 10. The methylene amide
bridge of compound 12 was installed by a two-step procedure,
whereby 10 was first N-acylated with chloroacetic chloride to
provide 11, which was cyclized by base mediated ether form-
ation. Thus, compound 10 was treated with chloroacetic
anhydride in a mixture of DCM and MeOH to give 11 in a yield
of 90%. A significant amount of acylation of the C2 hydroxyl
was observed when the reaction was performed in the absence
of methanol as a co-solvent. Treatment of 11 with NaH in
DMF give a mixture of the required cyclized 12 and a dimer-
ized side product. Fortunately, a significantly reduction of the
amount of side product could be achieved by performing the
reaction under dilute conditions. The two compounds could be
separated by size exclusion chromatography over Sephadex
LH-20 to give pure 12 in a yield of 56%. Catalytic hydro-
genation of 12 over Pd/C gave the required derivative 4. The
presence of the tether in 4 was confirmed by a NOEs between
(i) H2Ј and H8RЈ; (ii) NH and H1Ј, H4 and H6S and by the
relatively downfield chemical shift of the C2Ј (δ 87.1),
characteristic of hydroxyl alkylation.
and in each case, the NH vector, which is approximately parallel
to C1Ј–H1Ј and C4–H4, can point in the same direction as these
two CH vectors (parallel orientation, p) or in the opposite direc-
tion (antiparallel orientation, ap). These five possible conform-
ations, i.e. Trans_ap, Trans_p, Trans_p_2, Cis_p and Cis_ap
correspond to the conformational families that can be adopted
by compound 4 as described in Table 1. The Trans_p_2 con-
formation displays a different orientation around O2Ј as com-
pared to Trans_p and is higher in energy. The lowest energy
conformations of the two possible trans shapes (Trans_ap,
Trans_p) have almost equal potential energies, whereas the cis
conformations (Cis_p and Cis_ap) are calculated to have poten-
tial energies about 3 kcal molϪ1 above the trans ones. (Table S2
and Fig. S6 of ESI†).
When analyzing the dihedral angles of the glycosidic linkage,
it appears that each conformational family of the cyclic
compound 4 can adopt only a narrow range of Φ and Ψ angles
(Fig. 3). The possible conformations occupy a conformational
plateau, with no significant energy barriers, that encompasses
Conformational analysis
The conformational properties of compounds 4 were deter-
mined by a combination of molecular mechanics simulations
and NMR spectroscopy and the results compared with similar
data for N-acetyllactosamine.
A complete conformational search of 4 was performed by
systematically varying Φ and Ψ angles of the glycosidic link-
ages and the torsion angles of the methylene amide tether using
the SEARCH procedure of the SYBYL software package
(Tripos Inc., St Louis). Starting conformations were built by
graphically editing the cyclic derivative of N-acetyllactosamine
that were the subject of an earlier study8 and conformational
families of the cyclic derivative were obtained using a clustering
procedure. Eight conformational families were identified and
the characteristics of the five lowest energy ones are in Table 1.
The tether of 4 can adopt different conformations that can be
described by the conformations and orientations of the –CO–
NH– segment. In particular, the amide bond can be cis or trans,
Fig. 3 Energy map of LacNAc glycosidic linkage17 calculated with
MM3. Conformations accessible to compound 4 have been super-
imposed
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 8 9 1 – 3 8 9 9
3893

View Article Online
the two low energy conformations of LacNAc that have been
previously described as conf_A (Φ = 40Њ, Ψ = 0Њ) and conf_B (Φ
= 40Њ, Ψ = Ϫ60Њ).8 The secondary energy minima corresponding
to trans orientations of Φ or Ψ cannot be reached by compound
4. The glycosidic linkage of the Cis families are limited to very
narrow range of conformation at the limit of the low energy
region. The trans families are somewhat more flexible, but
Trans_p, the most populated family can adopt only conform-
ations of the Conf_B region, whereas conformations of the
Trans_ap family occupy a region in between Conf_A and
Conf_B.
was observed indicating that the cis conformations are not
significantly populated in solution. An NOE between H3Ј and
H8SЈ proposed for the trans-ap structure could not be con-
firmed because of a strong NOE between H8SЈ and H8RЈ, and
3
H8RЈ is very close to H3Ј. The estimated JH5,H6R = 5 Hz, and
3JH5,H6S = 10 Hz couplings were consistent with a tg rotamer
form, and the Trans-p structure.
The experimental and computer modeling data indicate that
the conformation of compound 4 is consistent with the Trans-p
conformational family. This locates the amide proton on the
side of H1Ј and H4, and the carbonyl oxygen near H6R and
H5. However, the calculated Φ and Ψ angles of these con-
formers are not in perfect agreement with experimental data
and therefore a final geometry optimization was performed on
the lowest energy conformation of the Trans_p family, with the
inclusion of a distance constraint between H1Ј and H3, since
this proton pair distance is more dependant on the glycosidic
linkage conformation than H1Ј/H4.28 The refined structure
NMR data and comparison with molecular modeling
1
The H and 13C spectra of compound 4 were assigned by a
combination of one-dimensional proton, and two-dimensional
HSQC and COSY experiments. (Fig. S1 of ESI†) Due to signal
overlap and strong coupling, most of the data for conform-
ational analysis (3JCH and NOEs) were obtained with a sample
in 21% pyridine–D2O and 10% DMSO–pyridine. (Fig. S2 of
ESI†) Heteronuclear 3JCH coupling constants for LacNAc were
also measured in a 30% pyridine–D2O solution at 40 ЊC to per-
turb the chemical shifts of H1, H1Ј, H3 and H4. Earlier values
reported8 were underestimated due to overlap of H1 and H1Ј,
and strong coupling between H3 and H4.
3
provided a structure that has Φ angle and JH1Ј,C4 value very
close to the experimentally observed value (Table 1). The
resulting conformation is displayed in Fig. 4. The glycosidic
linkage adopts a conformation with (ΦH = 27Њ and ΨH = Ϫ64Њ),
which corresponds to the low energy conformation Conf_B of
LacNAc. The calculated heteronuclear coupling constants are
in agreement with the experimentally determined ones (Table
1). Comparison of the several observed strong NOEs with the
theoretical distances indicate that the conformation displayed
in Fig. 4 is the major one in solution. Some fluctuations
may occur around this equilibrium value, especially in the
tethering bridge. Furthermore, it can be concluded that 4
adopts a low minimum energy conformation on the energy map
of LacNAc and furthermore, it is significantly more rigid that
the corresponding ethylene tethered analogue 3.
Assignment of pro-R and pro-S hydroxymethylene protons
of the GlcNAc moiety of LacNAc and compound 4 was based
3
on relative JHH coupling constants and H5–H6 NOEs. For
LacNAc, it is assumed that predominantly gg and gt rotamers
are populated,23 and it is to be expected that H5–H6pro-R is
larger due to the trans relationship in the gt conformer, 3JH5,H6R
= 5.5 Hz, while 3JH5,H6S = 2.0 Hz. On the contrary, the tether of
compound 4 restricts the rotamers to tg or possibly eclipsed
forms, and therefore the H5–H6pro-S coupling is expected to be
similar or larger than H5–H6pro-R. Although the line widths
were large for these signals, estimates of the couplings could be
made and 3JH5,H6R = 5 Hz, while 3JH5,H6S = 10 Hz. A weak NOE
between the assigned H6pro-S and H5, and the fact that no NOEs
were observed between H1Ј and H6pro-R or H1Ј and H6pro-S is
consistent with a tg rotamer.
trans-Glycosidic heteronuclear coupling constants, which are
sensitive to orientations around glycosidic linkages, were meas-
ured using one- and two-dimensional HSQMBC experiments.24
The resulting values are listed in Table 1 together with the calcu-
lated values for each low energy conformer using an empirical
3
3
Karplus-type equation.25 For LacNAc, the JH1Ј,C4 and JH4,C1Ј
values of 3.8 and 4.6 Hz, respectively, support the mixture of
conformers as determined by the modeling studies. Compound
3
4 provided a value of JH1Ј,C4 = 4.4 Hz corresponding to Φ of
approximately 48Њ, and a small value of 3JH4,C1Ј = 1.5 Hz corre-
sponding to Ψ of approximately Ϫ61Њ. Although these values
were obtained in solvent systems other than pure water, it is
reasonable to assume that the conformations are not greatly
altered.26
Fig. 4 Low energy conformation of optimized compound 4. The
structure has been colored according to atom types. Arrows indicate
predicted NOEs.
Enzymatic glycosylations
Distance constraints were obtained from relative NOE inten-
sities using the isolated spin-pair approximation.27 A strong
H1Ј–H4 NOE (2.3 Å) was obtained for compound 4, which was
expected for a compound restricted to syn conformations. Fur-
thermore a medium strong H1Ј,H3 (3.0 Å) NOE was detected
which is consistent with a Ψ value of Ϫ61Њ since that positions
H3 closer to H1Ј. Additional qualitative NOE data were
obtained in a solution of 10% dmso–pyridine that support the
presence of the trans_p conformer. (Table S2 of ESI†) In par-
ticular, NOEs were observed between the NH and linker H8SЈ,
NH and H4, H6S, H6R and H1Ј. The NOE between H2Ј and
H8RЈ proposed for trans-p structure was also observed,
whereas in the other solvents (21% pyridine–D2O), it could not
be measured due to signal overlap. Moreover, neither the NOE
between H6S and H8SЈ expected for cis-ap structure nor the
one between H6R and H8RЈ proposed for the cis-p structure
After having established that compound 4 can adopt a low
energy conformation on the energy surface of LacNAc, atten-
tion was focused on the enzymatic glycosylation by α-2,6-sialyl-
transferase and α-1,3-fucosyltransferase V (Scheme 2). The
transfer of sialic acid was accomplished utilizing an excess of
CMP-NeuAc as glycosyl donor in the presence of rat liver
α-2,6-sialyltransferase as the catalyst and calf alkaline phos-
phatase to ensure low levels of CMP released during the reac-
tion which is known to inhibit the enzyme.29 The reaction was
carried out for two days after which the crude reaction mixture
was purified by ion exchange chromatography (Dowex 1 ×
8–200, PO42Ϫ, 100–200 mesh) followed by biogel P2 size exclu-
sion column chromatography to give 5 in quantitative yield.
In order to obtain fucosylated derivative 6, compound 4
was treated with an excess of GDP-fucose in the presence of
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 8 9 1 – 3 8 9 9
3894

View Article Online
Table 2 Apparent kinetic parameters for the transfer of N-acetylneuraminic acid by α-2,6-sialyltransferase and relative rates of transfer of fucose
by fucosyltransferase V.
α-2,6-ST
Km/mM
FucT V
Rel. rate transfer (%)
Rel. Vmax
(Rel. Vmax/Km)/mMϪ1
LacNAc-OMe
3
4
1.7 0.2
1.1 0.1
1.0 0.1
1.0
1.3
1.1
0.6
1.2
1.1
100
39
9
derivative 3 was the substrate that was employed more effi-
ciently. Thus, this finding indicates that the fucosyltransferase
recognize LacNAc in its A-conformer. Further studies with a
wider range of glycosyltransferases and modified LacNAc
acceptors will be reported elsewere to support the notion that
different transferases may recognize LacNAc in different
conformations.
Conclusion
Compound 4, which is preorganized by a methylene amide
linker between the C-6 and C-2Ј, was obtained by first syn-
thesizing a selectively protected 6-amino LacNAc derivative,
which was N-acylated with chloroacetic anhydride followed
by macro-cyclization and finally deprotection. A combination
of NOE and trans-glycosidic heteronuclear coupling constant
measurements and molecular mechanics simulations estab-
lished that the glycosidic linkage of compound 4 is highly con-
formationally constrained and resides in the B-conformer
on the energy map of LacNAc. Previously, we synthesized a
LacNAc derivative (3) that is conformationally constraint by an
ethylene tether and this compound could adopt both A and
B conformations. Thus, LacNAc derivative 4 is more rigid
than the ethylene analogue and provides an important comple-
mentary derivative to study conformational requirements of a
range of transferases and lectins. Compound 4 proved to be a
substrate for rat liver α-2,6-sialyltransferase and fucosyl-
transferase V and this finding was exploited for the preparative
enzymatic synthesis of conformationally constrained sialyl-
lactosamine 5 and Lewisx 6, respectively. Although several
studies have shown that glycosyltransferases can accept small
changes in their glycosyl acceptors,8 it is shown here for the first
time that these enzymes can also be employed for the chemo-
enzymatic synthesis of oligosaccharides that are locked in a
particular conformation. The apparent kinetic parameters for
the α-(2,6)-sialyltransferase and fucosyltransferase V catalyzed
transformations indicate that the two enzymes respond differ-
ently to imposed conformational constraints of the accep-
tor opening an exciting route for the design of selective
glycosyltransferase substrates and inhibitors.
Scheme 2 Reagents and conditions: enzyme reactions; (i) α-2,6-sialyl-
transferase, CMP-NeuAc; (ii) α-1,3-fucosyltransferase V, GDP-fucose
recombinant human α-1,3-fucosyltransferase V and alkaline
phosphatase to ensure a low concentration of GDP. In this case,
the rate of transfer was significantly slower, however, after a
reaction time of 5 days, MALDI-TOF mass spectrometry indi-
cated completion of the reaction. The crude reaction mixture
was purified by ion exchange column chromatography followed
by biogel P2 size exclusion column chromatography to give 6 in
an excellent yield of 99%.
The presence of the linker in compounds 5 and 6 was con-
firmed by the shifts of H8RЈ and H8SЈ and a large coupling
constant J8RЈ,8SЈ that were also observed in compound 4 (4.56,
3.79 ppm, J8RЈ,8SЈ 16.5 Hz), for compound 5 (4.52, 3.79 ppm,
J8RЈ,8SЈ 16.6 Hz) and for compound 6 (4.47, 3.74 ppm, J8RЈ,8SЈ
17.6 Hz). The presence the α-sialyl linkage of compound 5 was
confirmed by a typical shift of H4Љ (δ 3.57 ppm) and the differ-
ence in chemical shift of H9a and H9b (H9Љa – H9Љb = 0.23
ppm).30 The typical shift and coupling constant of the anomeric
proton of the α-fucosyl linkage (H1Љ δ 5.04 ppm, J1Љ,2Љ 5.1 Hz)
confirmed its presence in compound 6.
Based on the observation that compound 4 is an appropriate
substrate for glycosyl transferases, apparent kinetic parameters
for α-(2,6)-sialyltransferase catalyzed transfer of CMP-
(14C)Neu5Ac to LacNAc and conformationally constraint
compounds 3 and 4 were determined using a reported assay20
(Table 2). The Km for 1 was in close agreement with previous
data and its Vmax was set at 1. As can be seen in Table 2, very
similar Km and relative Vmax values were obtained for each com-
pound. Compound 4 is contained in the B-conformer on the
energy surface of LacNAc whereas the previously studied 3 can
adopt the A- and the B-conformer. When a transferase prefers
the A-conformer, it is to be expected that only 3 gives favorable
kinetic parameters. On the other hand, when the B-conformer is
recognized both compounds should display good transform-
ations. Thus, the results indicate that α-(2,6)-sialyltransferase
recognizes LacNAc in the B-conformation.
Experimental
General remarks
Chemicals were purchased from Aldrich and Fluka and used
without further purification. Molecular sieves were activated at
350 ЊC for 3 h in vacuo. Dichloromethane was distilled from
CaH2 and stored over 4Å molecular sieves. All the reactions
were performed under anhydrous conditions and monitored by
TLC on Kieselgel 60 F254 (Merck). Detection was by examin-
ation under UV light (254 nm) and by charring with 10% sul-
furic acid in methanol. Flash chromatography was performed
on silica gel (Merck, mesh 70–230). Extracts were concentrated
1
under reduced pressure at <40 ЊC (bath). H NMR (1D, 2D)
and 13C NMR spectra were recorded on a Varian Merc300 spec-
trometer and Varian 500, 600 and 800 MHz spectrometers
equipped with Sun workstations. For 1H and 13C NMR spectra
recorded in CDCl3, chemical shifts (δ) are given in ppm relative
to solvent peaks (1H, δ 7.26; 13C, δ 77.3) as internal standard for
In the case of fucosyltransferase V, only relative rates could
be measured due to high Michaelis–Menten constants. It is,
however, apparent, that this transferase responds differently to
the imposed constraints and in this case the ethylene bridged
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 8 9 1 – 3 8 9 9
3895

View Article Online
protected compounds and using NAc proton (2.025 ppm) and
OMe carbon (59.82 ppm) as internal standard for deprotected
molecules. Negative ion matrix assisted laser desorption ionis-
ation time of flight (MALDI-TOF) mass spectra were recorded
using an HP-MALDI instrument using gentisic acid matrix.
Optical rotations were measured on a Jasco P-1020 polarimeter,
and [α]D values are given in units of 10Ϫ1 deg cm3 gϪ1 at 26 ЊC,
50 mm cell.
72.78, 72.42, 71.60, 68.29, 56.71, 55.70, 51.44, 23.75. MALDI-
TOF: m/z 806.3 [M ϩ Na]ϩ. Anal. Calc. (%) for C43H50N4O10:
C, 65.97; H, 6.44; N, 7.16; O, 20.44. Found: C, 66.04; H, 6.55;
N, 7.26.
Methyl 2-acetamido-6-amino-3-O-benzyl-2,6-dideoxy-4-O-
(3,4,6-tri-O-benzyl-ꢀ-D-galactopyranosyl)-ꢀ-D-glucopyranoside
(10)
Recombinant human α-1,3-fucosyltransferase V, rat liver
α-2,6-sialyltransferase, CTP, CMP-Neu5Ac and calf alkaline
phosphatase were purchased from Calbiochem. ACS liquid
scintillation cocktail was obtained from Fisher Scientific.
1,3-Propane dithiol (151 µL, 1.5 mmol) and triethylamine
(209 µL, 1.5 mmol) were added to a stirred solution of 9
(120 mg, 0.15 mmol) in a mixture of pyridine and water (4 : 1,
v/v, 20 mL). The mixture was stirred at room temperature for
6 h TLC (chloroform–methanol, 9 : 1, v/v) indicated comple-
tion of the reaction. The mixture was co-evaporated with tolu-
ene, concentrated in vacuo and then chromatographed over
flash silica gel (gradient DCM–methanol, 1 : 0 to 9 : 1, v/v) to
afford compound 10 (90.8 mg, 80%). [α]D= Ϫ14.9 (c 1.39,
Methyl 2-acetamido-6-azido-3-O-benzyl-2,6-dideoxy-4-O-(2-O-
acetyl-3,4,6-tri-O-benzyl-ꢀ-D-galactopyranosyl)-ꢀ-D-gluco-
pyranoside (8)
Sodium azide (130 mg, 2.0 mmol) was added to a stirred solu-
tion of 7 (438.7 mg, 0.5 mmol) in dry N,N-dimethylformamide
(20 mL). After stirring for 2 h at 90ЊC, TLC (chloroform–meth-
anol, 9 : 1, v/v) indicated completion of the reaction. The reac-
tion mixture was diluted with ethyl acetate (100 mL) and suc-
cessively washed with water (2 × 20 mL) and brine (20 mL),
followed by drying over MgSO4. After evaporation of the sol-
vent, the residue was purified by flash silica gel chromatography
(hexane–acetone, gradient 3 : 1 to 1 : 1, v/v) to afford compound
8 (403.9 mg, 98%) as a white foam. [α]D = 3.8 (c 5.03, CHCl3).
1H NMR (CDCl3, 300 MHz): δ 7.40–7.10 (m, 20H, arom), 6.04
(d, 1H, JNH,2 8.5 Hz, NHCOCH3), 5.33 (dd, 1H, J2Ј,3Ј 10.2 Hz,
H-2Ј), 4.93, 4.58 (AB q, 2H, JAB 11.3 Hz, OCH2Ph), 4.70, 4.64
(AB q, 2H, JAB 11.8 Hz, OCH2Ph), 4.60, 4.52 (AB q, 2H,
JAB 12.1 Hz OCH2Ph), 4.56 (d, 1H, J1,2 9.1 Hz, H-1), 4.40, 4.33
(AB q, 2H, JAB 12.9 Hz, OCH2Ph), 4.36 (d, 1H, J1Ј,2Ј 8.0 Hz,
H-1Ј), 3.97 (d, 1H, J3Ј,4Ј 2.8 Hz, H-4Ј), 3.92 (m, 1H, H-2), 3.85
(m, 1H, H-3), 3.74 (t, 1H, J3,4 = J4,5 5.5 Hz, H-4), 3.70–3.40 (m,
4H, H-6a, H-6b, H-6aЈ, H-6bЈ), 3.67 (m, 1H, H-5), 3.52 (m, 1H,
H-5Ј), 3.46 (m, 1H, H-3Ј), 3.44 (s, 3H, OCH3), 1.99 (s, 3H,
CH3C(O)NH), 1.85 (s, 3H, CH3C(O)O). 13C NMR (CDCl3, 125
MHz): δ 170.03, 169.78, 138.18, 137.62, 137.57, 128.29, 128.08,
128.07, 128.00, 127.74, 127.72, 127.67, 127.53, 127.47, 127.28,
101.20, 100.48, 79.86, 76.58, 75.47, 74.58, 74.08, 73.67, 73.49,
72.92, 72.47, 72.02, 71.74, 68.02, 56.44, 51.76, 51.71, 23.33.
21.06. MALDI-TOF: m/z 847.6 [M ϩ Na]ϩ. Anal. Calc. (%) for
C45H52N4O11: C, 65.52; H, 6.35; N, 6.79; O, 21.33. Found: C,
65.22; H, 6.40; N, 6.84.
1
CHCl3). H NMR (CDCl3, 300 MHz): δ 7.40–7.10 (m, 20H,
arom), 6.10 (br s, 1H, NHCOCH3), 5.02, 4.74 (AB q, 2H,
JAB 11.5 Hz, OCH2Ph), 4.90, 4.64 (AB q, 2H, JAB 11.3 Hz,
OCH2Ph), 4.78, 4.54 (AB q, 2H, JAB 11.8 Hz, OCH2Ph), 4.69
(d, 1H, J1,2 8.0 Hz, H-1), 4.65 (d, 1H, J1Ј,2Ј 7.7 Hz, H-1Ј), 4.32,
4.22 (AB q, 2H, JAB 11.8 Hz, OCH2Ph), 4.10 (t, 1H, J3,4 9.3 Hz,
H-3), 3.94 (t, 1H, J4,5 9.3 Hz, H-4), 3.92 (m, 1H, H-2), 3.85 (d,
1H, J3Ј,4Ј 2.2 Hz, H-4Ј), 3.67 (m, 1H, H-5), 3.48 (m, 1H, H-2Ј),
3.58–3.40 (m, 5H, H-3Ј, H-2Ј, H-6aЈ, H-6bЈ), 3.44 (s, 3H,
OCH3), 3.33 (dd, 1H, J6b,5 4.4 Hz, J6b,6a 11.8 Hz, H-6b), 3.12
(dd, 1H, J6a,5 5.0 Hz, H-6a), 1.90 (s, 3H, CH3C(O)NH). 13C
NMR (CDCl3, 125 MHz): δ 170.68, 139.14, 138.88, 138.63,
137.88, 128.44, 128.19, 127.92, 127.78, 127.62, 127.48, 127.34,
104.85, 101.74, 82.33, 80.84, 78.68, 74.74, 73.93, 73.69, 73.56,
73.00, 72.56, 68.69, 57.18, 56.97, 41.92, 23.75. MALDI-TOF:
m/z 779.5 [M ϩ Na]ϩ. Anal. Calc. (%) for C43H52N2O10: C,
68.24; H, 6.92; N, 3.70; O, 21.14. Found: C, 68.35; H, 7.04; N,
3.81.
Methyl 2-acetamido-3-O-benzyl-6-chloroacetamido-2,6-
dideoxy-4-O-(2-O-acetyl-3,4,6-tri-O-benzyl-ꢀ-D-galactopyrano-
syl)-ꢀ-D-glucopyranoside (11)
Chloroacetic anhydride (363 mg, 2.12 mmol) was added to a
solution of 10 (100 mg, 0.13 mmol) in a mixture of dichloro-
methane and methanol (1 : 1, mL). The mixture was stirred at
room temperature for 16 h. TLC analysis (chloroform–
methanol, 9 : 1, v/v) indicated completion of the reaction. The
reaction mixture was diluted with dichloromethane (20 mL),
washed successively with a saturated solution of NaHCO3
(10 mL), water (2 × 10 mL) and brine (10 mL), followed by
drying over MgSO4. After evaporation of the solvent, the resi-
due was purified by flash silica gel chromatography (hexane–
acetone, gradient 3 : 1 to 1 : 1, v/v) followed by size exclusion
column chromatography (LH20, MeOH–DCM, 1 : 1, v/v) to
Methyl 2-acetamido-6-azido-3-O-benzyl-2-dideoxy-4-O-(3,4,6-
tri-O-benzyl-ꢀ-D-galactopyranosyl)-ꢀ-D-glucopyranoside (9)
Sodium methoxide (10.8 mg, 0.2 mmol) was added to the
stirred solution of 8 (256.2 mg, 0.31 mmol) in methanol
(20 mL). The mixture was left stirring at room temperature for
48 h. TLC (hexane–acetone, 2 : 1, v/v) indicated completion of
the reaction. The mixture was neutralized with Dowex 50Hϩ
WX4–200 ion exchange resin until pH = 7, filtered and concen-
trated in vacuo. The residue was purified by flash silica gel
chromatography (gradient hexane–acetone, 3 : 1 to 1 : 1, v/v) to
afford compound 9 (237.7 mg, 98%) as a white foam. [α]D = 10.5
(c 5.04, CHCl3). 1H NMR (CDCl3, 300 MHz): δ 7.40–7.10 (m,
20H, arom), 5.65 (d, 1H, JNH,2 7.9 Hz, NHCOCH3), 4.80, 4.62
(AB q, 2H, JAB 11.9 Hz, OCH2Ph), 4.79, 4.53 (AB q, 2H, JAB
11.4 Hz, OCH2Ph), 4.65 (d, 1H, J1,2 7.5 Hz, H-1), 4.50, 4.48
(AB q, 2H, JAB 11.4 Hz, OCH2Ph), 4.35 (d, 1H, J1Ј,2Ј 7.9 Hz,
H-1Ј), 4.29, 4.21 (AB q, 2H, JAB 11.9, OCH2Ph), 3.97 (t, 1H, J2,3
8.4, J3,4 8.8, H-3), 3.87 (d, 1H, J3Ј,4Ј 2.7 Hz, H-4Ј), 3.83 (m, 1H,
H-2Ј), 3.71 (dd, 1H, J3,4 = J4,5 7.9 Hz, H-4), 3.66–3.24 (m, 6H,
H-5, H-5Ј H-6a, H-6b, H-6aЈ, H-06bЈ), 3.44–3.24 (m, 1H,
H-3Ј), 3.40 (s, 3H, OCH3), 1.78 (s, 3H, CH3C(O)NH). 13C
NMR (CDCl3, 125 MHz): δ 170.50, 138.79, 138.67, 137.89,
137.82, 128.62, 128.49, 128.29, 127.95, 127.75, 127.59, 127.48,
103.52, 101.18, 82.30, 78.62, 78.01, 74.77, 73.71, 73.64, 73.55,
afford compound 11 (97.4 mg, 90%) as a white foam. [α]D
=
Ϫ3.97 (c 1.05, CHCl3). 1H NMR (CDCl3, 300 MHz): δ 7.40–7.2
(m, 20H, arom), 7.17 (m, 1H, NHCO), 5.56 (d, 1H, JNH,2
7.3 Hz, NHCOCH3), 4.89, 4.59 (AB q, 2H, JAB 11.9 Hz,
OCH2Ph), 4.72, 4.64 (AB q, 2H, JAB 11.9 Hz, OCH2Ph), 4.69
(d, 1H, J1,2 7.9 Hz, H-1), 4.63, 4.55 (AB q, 2H, JAB 11.9 Hz,
OCH2Ph), 4.53 (d, 1H, J1Ј,2Ј 7.0 Hz, H-1Ј), 4.34, 4.28 (AB q, 2H,
JAB 11.9 Hz, OCH2Ph), 4.02 (s, 2H, CH2CO), 4.03 (m, 1H,
H-3), 3.97 (dd, 1H, J2Ј,3Ј 8.8 Hz, H-2Ј), 3.91 (d, 1H, J3Ј,4Ј 2.2 Hz,
H-4Ј), 3.69 (t, 1H, J4,5 8.4 Hz, H-4), 3.52 (m, 1H, H-2), 3.47 (s,
3H, OCH3), 3.41 (dd, 1H, J3Ј,4Ј 11.8 Hz, H-3Ј), 3.88–3.58 (m,
4H, H-5, H-5Ј, H-6aЈ, H-6bЈ), 3.58 (m, 1H, H-6b), 3.36 (dd, 1H,
J6a,5 4.4 Hz, J6a,6b 12.3 Hz, H-6a), 1.84 (s, 3H, CH3C(O)NH).
13C NMR (CDCl3, 125 MHz): δ 170.36, 166.39, 138.82, 138.74,
138.18, 137.90, 128.86, 128.63, 128.52, 128.42, 128.30, 128.03,
127.99, 127.92, 127.76, 127.67, 127.61, 127.56, 103.76, 101.27,
82.39, 79.37, 78.80, 74.79, 74.17, 74.05, 73.82, 73.72, 73.68,
73.03, 72.73, 72.04, 68.51, 57.01, 56.16, 42.94, 40.90, 23.81.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 8 9 1 – 3 8 9 9
3896

View Article Online
MALDI-TOF: m/z 855.33 [M ϩ Na]ϩ. Anal. Calc. (%) for
C45H53N2O11: C, 64.86; H, 6.41; N, 3.36; O, 21.12. Found: C,
64.97; H, 6.47; N, 3.47.
Methyl 2-acetamido-3-O-benzyl-2,6-dideoxy-2Ј,6-(2Ј-O-methyl-
enelactam)-4-O-(3,4,6-tri-O-benzyl-ꢀ-D-galactopyranosyl)-ꢀ-D-
glucopyranoside (12)
Sodium hydride (60% in oil suspension, 40 mg, 1.1 mmol) was
added to a solution of 11 (90 mg, 0.11 mmol) in dry N,N-di-
methylformamide (20 mL). After stirring at room temperature
16 h, TLC (chloroform–methanol, 9 : 1, v/v) indicated comple-
tion of the reaction. The reaction mixture was quenched with
methanol, diluted with ethyl acetate (50 mL) and successively
washed with water (2 × 10 mL) and brine (10 mL), followed by
drying over MgSO4. After evaporation of the solvents, the
residue was then purified by flash silica gel chromatography
(hexane–acetone, gradient 3 : 1 to 1 : 1, v/v) to afford compound
12 (49 mg, 56%) as a white foam. [α]D = Ϫ106 (c 0.40, CHCl3).
1H NMR (DMSO, 300 MHz): δ 7.20–7.05 (m, 20H, arom), 7.94
(d, 1H, JNH,2 8.8 Hz, NHCOCH3), 7.74 (t, 1H, J 6.5 Hz,
NHCO), 4.78 (d, 1H, J1,2 6.4 Hz, H-1), 4.76, 4.44 (AB q, 2H,
JAB 11.2 Hz, OCH2Ph), 4.75, 4.51 (AB q, 2H, JAB 11.7 Hz,
OCH2Ph), 4.73, 4.68 (AB q, 2H, JAB 12.2 Hz OCH2Ph), 4.72 (d,
1H, OCH–HCO), 4.57 (d, 1H, J 11.72 Hz, OCH–HCO), 4.27
(d, 1H, J1Ј,2Ј 8.3 Hz, H-1Ј), 4.11, 3.98 (AB q, 2H, JAB 12.2 Hz,
OCH2Ph), 3.88 (d, 1H, J3Ј,4Ј <2 Hz, H-4Ј), 3.74 (m, 1H, H-3),
3.69 (dd, 1H, J 9.3 Hz, H-2), 3.67 (m, 1H, H-6bЈ), 3.64 (m,
2,3
1H, H-6aЈ), 3.57 (m, 1H, H-5), 3.53 (m, 1H, H-4), 3.51 (m, 1H,
H-3Ј), 3.34 (m, 1H, H-5Ј), 3.31 (m, 1H, H-6b), 3.30 (s, 3H,
OCH3), 3.29 (m, 1H, H-2Ј), 2.93 (dd, 1H, J6a,5 4.9 Hz, J6a,6b 8.3
Hz, H-6a), 1.89 (s, 3H, CH3C(O)NH). 13C NMR (DMSO,
125 MHz): δ 177.51, 171.98, 129.5–126.13 (arom.), 103.98,
101.79, 83.79, 82.71, 81.30, 74.58, 74.07, 73.89, 73.20, 72.90,
72.75, 72.32, 71.75, 68.31, 68.00, 56.35, 54.67, 40.30, 23.41.
MALDI-TOF: m/z 820.2 [M ϩ Na]ϩ. Anal. Calc. (%) for
C45H52N2O11: C, 67.82; H, 6.58; N, 3.52; O, 22.08. Found: C,
67.93; H, 6.62; N, 3.59.
Methyl 2-acetamido-2,6-dideoxy-(2Ј-O-,6-N-methylenelactam)-
4-O-(ꢀ-D-galactopyranosyl)-ꢀ-D-glucopyranoside (4)
10% Palladium on charcoal (32 mg) was added to a solution of
12 (30 mg, 0.04 mmol) in ethanol (3 mL). The mixture was
vigorously stirred under an atmosphere of hydrogen for 16 h.
TLC analysis (chloroform–methanol, 9 : 1, v/v) indicated com-
pletion of the reaction. After filtration using Celite and concen-
tration of the filtrate under reduced pressure, the crude product
was purified by column chromatography (Iatrobeads, chloro-
form–methanol–water, 74 : 24 : 2, v/v/v) to afford compound 4
(16.9 mg, 97%) as a white amorphous solid. [α]D = Ϫ61.9
(c 0.24, D2O). 13C NMR (D2O, 125 MHz): δ 106.1 (C-1Ј), 103.8
(C-1), 90.1 (C-4), 87.1 (C-2Ј), 78.4 (C-5Ј), 76.1 (OCH2CO), 75.7
(C-3), 74.6 (C-3Ј), 70.8 (C-4Ј), 68.5 (C-5), 63.2 (C-6Ј), 59.8
(OCH3), 56.8 (C-2), 44.6 (C-6), 25.3 (COCH3). MALDI-TOF:
m/z 459.3 [M ϩ Na]ϩ. For 1H NMR data see Table 3.
Methyl 5-acetamido-3,5-dideoxy-D-glycero-ꢁ-D-galacto-2-nonu-
lopyranosyl-(2–6)-ꢀ-D-galactopyranosyl-(1–4)-2-acetamido-2,6-
dideoxy-(2Ј-O-,6-N-methylenelactam)-4-O-ꢀ-D-glucopyranoside
(5)
Incubations were carried out in 150 µL sodium cacodylate buf-
fer (25 mM, pH 7.2) containing 0.5% Triton X100 and bovine
serum albumin (1 mg mlϪ1). Acceptor 4 (2.00 mg, 4.6 µmol),
CMP-Neu5Ac (2.0 mg), alkaline phosphatase (3 U), and sialyl-
tranferase (6.9 mU) were incubated at 37 ЊC for 36 h. Every
12 h, phosphatase (3 U), enzyme (6.9 mU) and CMP-Neu5Ac
(2.0 mg) were added to the incubation mixture to compensate
for any hydrolysis of the donor. The reaction was monitored by
TLC on silica gel plates using CHCl3–MeOH–H2O (60 : 35 : 6,
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 8 9 1 – 3 8 9 9
3897

View Article Online
v/v/v) as the eluent. The acceptor 1 has an Rf of 0.32 and com-
pound 2 had an Rf of 0.06. At the completion of the reaction,
the sample was purified on ion-exchange chromatography using
Dowex 1 × 8–200 (PO42Ϫ, 100–200 mesh) followed by biogel P2
to afford compound 5 (3.20 mg, 99%). [α]D = Ϫ2.54 (c 2.8,
D2O). 13C NMR (D2O, 125 MHz): δ 181.34(COCH3), 181.25
(COCH3), 180.15 (CO2H), 109.92 (C-1Ј), 107.66 (C-1), 94.18,
90.50, 80.00, 79.83(OCH2CO), 79.25, 79.08, 77.93, 74.75, 74.21
(C-9Љ), 72.28, 69.22, 69.18, 68.87 (C-6Ј), 59.82, 57.92, 47.82
(C-6), 46.13 (C-3Љ), 28.32(COCH3), 28.28(COCH3), 23.03.
MALDI-TOF: m/z = 727.3 [Mϩ ϩ Na]. For 1H NMR data see
Table 3.
ance for ring closure was 0.3 Å for bond length and 20Њ for
bond angles. The systematic search was performed with steps of
5Њ on all torsion angles except for the amine group (angle µ2) for
which only values of 0Њ and 180Њ were tested. The van der Waals
cut off (steric penetration allowed) was fixed to 0.6. The
hydroxylic hydrogen atoms were not taken into account during
this rigid search. The energy was evaluated for each conform-
ation but the electrostatic contribution was not included. Con-
formations were accepted in an energy window of 50 kcal
molϪ1. The resulting conformations were then analyzed by a
home-made clustering program to determine conformational
families.
Methyl ꢁ-D-fucopyranosyl-(1–3)-[ꢀ-D-galactopyranosyl-(1–4)]-
2-acetamido-2,6-dideoxy-(2Ј-O-6-N-methylenelactam)-D-gluco-
pyranoside (6)
Geometry optimization. Compound 4 was fully optimized
with the TRIPOS force-field. Partial charges were assigned
according to the PIM energy parameters for carbohydrates34
and the dielectric constant was set to a value of 4.0. Energy
minimisation was performed with inclusion of one distance
constraint (H1 ؒ ؒ ؒ H3 set to 3.0 Å).
Incubations were carried out in 150 µL sodium cacodylate
buffer (25 mM, pH 6.5) containing MnCl2 (8 mM), ATP
(1.6 mM) and NaN3 (1.6 mM). Acceptor 4 (1.10 mg, 2.5 µmol),
GDP-Fucose (1.0 mg), alkaline phosphatase (4 U) and fucosyl-
transferase V (7.5 mU) are incubated at 37 ЊC for 4 days. Every
24 h, GDP-Fucose (1.0 mg), alkaline phosphatases (4 U) and
FucT V (7.5 mU) were added to the incubation mixture to
compensate for any hydrolysis of the donor. The reaction was
monitored by TLC on silica gel plates using CHCl3–MeOH–
H2O (60 : 35 : 6, v/v) as the eluent. The acceptor 4 has an Rf of
0.51 and compound 3 has an Rf of 0.33. After the completion
of the reaction, the sample was purified on ion-exchange chro-
matography on a Dowex 1 × 8–200 (ClϪ, 100–200 mesh) fol-
lowed by biogel P2 to afford compound 6 (1.44 mg, 99%). [α]D =
Ϫ26.86 (c 0.1, D2O). 13C NMR (D2O, 125 MHz): δ 180.85
(COCH3), 106.14, 105.04, (C-1, C-1Ј), 101.72 (C-1Љ), 91.82,
86.84, 86.53, 84.04, 78.25, 77.89, 77.46, 75.42, 74.60, 71.69,
71.86, 70.83, 70.22, 69.22, 68.40, 68.10, 63.66, 59.82, 57.95.
MALDI-TOF: m/z = 605.23 [Mϩ ϩ Na]. For 1H NMR data see
Table 3.
Heteronuclear 3JC–H coupling constants across the glycosidic
linkages were calculated using the Karplus-type equation
proposed by Tvaroska.25
3JC–H = 5.7cos2Φ Ϫ 0.6cosΦ ϩ 0.5
NMR Spectroscopy
Data were collected on Varian 500, 600 and 800 MHz spectro-
meters. The methyl amide disaccharide (5 mg) was lyophilized
from D2O and re-dissolved in 0.45 ml D2O. The data for signal
assignment was acquired on this sample at 25 ЊC with standard
Varian gradient COSY and HSQC pulse sequences. Data was
processed with nmrPipe software.35 For the measurement of
heteronuclear coupling constants and the measurement of
NOEs, 0.12 ml of pyridine-d5 was added to the sample, and the
temperature was raised to 50 ЊC to change the chemical shifts of
the overlapped H3 and H4 signals and remove second-order
effects. Proton–proton coupling constants were extracted from
the one-dimensional proton spectra in D2O and in 21% pyrid-
ine–D2O. For the observation of NOEs to the amide proton,
the sample was recovered, and dissolved in 10% DMSO-d6–
pyridine-d5. A two-dimensional ROESY experiment was used
to obtain qualitative NOE data.
Guanosine-5Ј-diphosphate-ꢀ-L-fucose, disodium salt (14)
GDP-Fucose was synthesized following a reported procedure.31
Bis(cyclohexylammonium) β--fucopyranosyl phosphate (354
mg, 0.80 mmol) was dissolved in H2O (10 mL), applied to a Bio-
Rad AGW-X2 cation-exchange column (Et3Nϩ) and eluted
with H2O (100 mL). The solution was evaporated, coevapo-
rated with MeOH (2 × 10 mL), and dried under vacuum to give
triethylammonium) β--fucopyranosyl phosphate 13 (290 mg).
A mixture of 14 (290 mg, 0.80 mmol) and 4-morpholine-N,NЈ-
dicyclohexylcarboxamidinium guanosine 5Ј-monophospho-
morpholidate (1 g, 1.27 mmol) was co-evaporated with dry
pyridine (2 × 2 mL). 1H-Tetrazole (165 mg, 2.4 mmol) and dry
pyridine (1.5 mL) were added and the solution was stirred at
room temperature. After a while, product started to precipitate.
The reaction was monitored by TLC (i-PrOH–NH4OAc (1 M)–
H2O (3 : 2 : 1, v/v/v)). The residue was passed through Dowex
resin 50 Naϩ with water, followed by purification using C-18
reverse phase chromatography in H2O to afford compound 13
(211 mg, 45%) as a white amorphous solid. 1H NMR data was
in agreement with those reported.31
A 10 mg sample of LacNac was dissolved in D2O and pyrid-
ine-d5 added to give a final solution of 30% pyridine–D2O.
Chemical shifts of this sample at 40 ЊC were correlated to the
data obtained in pure D2O at 25 ЊC, using standard HSQC and
TOCSY.
For LacNAc,
a
variation of the one-dimensional
HSQMBC24 was used in analogy to work done on lactose.37 In
the experiment, the initial proton pulse of the one-dimensional
HSQMBC sequence was replaced by
a selective one-
dimensional TOCSY sequence, and the H1 of the glucosamine
residue was selectively irradiated. Thus, only protons of the
glucosamine spin system continue through the remainder of the
pulse sequence. The result is a multiplet which arises from coup-
ling between glucosamine H4 and galactosyl C1Ј, thus avoiding
any contribution to the multiplet from 3JC1–H3 or 3JC1–H3Ј
.
Heteronuclear coupling constants. For compound 4, the three-
bond proton–carbon coupling constants (3JCH) were obtained
from a two-dimensional HSQMBC experiment, according to
Williamson et al.24 The data set was 1325 × 64 complex points,
Molecular modeling
Nomenclature. Torsion angles at the glycosidic linkage have
been defined with reference to the hydrogen atom: ΦH = H1Ј–
C1Ј–O1Ј–C4 and ΨH = C1Ј–O1Ј–C4–H4.
1
covering a H range of 2.3 ppm and 13C range of 78.3 ppm at
800 MHz. Two-dimensional data sets were zero-filled in t2 to
16384 points, and linear predicted and zero-filled in t1 to 512
points, giving 0.11 Hz/point resolution in t2 and 30 Hz/points
in t1. Coupling constants were measured by fitting a one-
dimensional reference spectrum to a one-dimensional trace
through the center of the cross-peak of interest, which contains
Conformational analysis of 4. All energy calculations were
performed using the TRIPOS force field32 complemented by
carbohydrate energy parameters.33 The possible conformations
of compound 4 were determined using the SEARCH procedure
of SYBYL. The ring closure bond was chosen to be ν2. Toler-
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 8 9 1 – 3 8 9 9
3898

View Article Online
the anti-phase heteronuclear coupling.36 The reference spectra
were constructed by adding a one-dimensional trace through
the center of the appropriate auto-peak (e.g. H1Ј–C1Ј) with a
copy of that trace, inverted 180Њ in phase and offset by a vari-
able number of points. The coupling constant was estimated
from the number of points offset for the reference spectrum that
best matched the cross-peak pattern, based on minimization of
their difference.
12 M. J. Milton and D. R. Bundle, J. Am. Chem. Soc., 1998, 120,
10547–10548.
13 R. Alibes and D. R. Bundle, J. Org. Chem., 1998, 63, 6288–6301.
14 S. A. Wacowich-Sgarbi and D. R. Bundle, J. Org. Chem., 1999, 64,
9080–9089.
15 J. L. Koviach, M. D. Chappell and R. L. Halcomb, J. Org. Chem.,
2001, 66, 2318–2326.
16 G. Thoma, J. L. Magnani, G. T. Patton, B. Ernst and W. Jahnke,
Angew. Chem., Int. Ed., 2001, 40, 1941–1945.
17 A. Imberty, E. Mikros, J. Koca, R. Mollicone, R. Oriol and S. Pérez,
Glycoconjugate J., 1995, 12, 331–349.
NOE experimental procedures. For each compound, selective
one-dimensional NOESY experiments27 were used to obtain
buildup curves for the anomeric and other resolved protons.
Linearity of the NOE buildup curves was observed at mixing
times beyond 300 ms. The isolated spin-pair approximation was
used to estimate the distances, by using the intra-ring H1Ј–H2Ј
NOE as an internal reference (3.09 Å).34 Qualitative NOEs
were also observed in the DMSO–pyridine solution with a two-
dimensional ROESY experiment.
18 L. G. Baum and J. C. Paulson, Virology, 1991, 180, 10–15.
19 S. I. Hakomori, Adv. Cancer Res., 1989, 52, 257.
20 K. B. Wlasichuk, M. Kashem, P. V. Nikrad, P. Bird, C. Jiang and
A. P. Venot, J. Biol. Chem., 1993, 268, 13971–13977.
21 T. de Vries, C. A. Srnca, M. M. Palcic, S. J. Sweidler, D. H. van den
Eijnden and B. A. Macher, J. Biol. Chem., 1995, 270, 8712.
22 W. Dullenkopf, G. Ritter, S. Fortunato, L. J. Old and R. Schmidt,
Chem. Eur. J., 1999, 5, 2432–2438.
23 E. Hounsell, Prog. Nucl. Magn. Reson. Spectrosc., 1995, 27, 445–
474.
24 R. Williamson, B. Marquez, W. Gerwick and K. Kover, Magn.
Reson. Chem., 2000, 38, 265–273.
Acknowledgements
25 I. Tvaroska, M. Hricovini and E. Petrakova, Carbohydr. Res., 1989,
189, 359–362.
26 S. A. Wacowich-Sgarbi, C. C. Ling, A. Otter and D. R. Bundle,
J. Am. Chem. Soc., 2001, 123, 4362–4363.
This research was supported by the Office of the Vice President
of Research of the University of Georgia.
27 K. Stott, J. Stonehous, J. Keeler, T.-L Hwang and A. J. Shaka, J. Am.
Chem. Soc., 1995, 117, 4199–4200.
28 M. Martín Pastor, J. Felix-Espinosa, J. L. Asensio and J. Jiménez-
Barbero, Carbohydr. Res., 1997, 298, 15–49.
29 Y. Kajihara, H. Kodama, T. Wakabayashi, K. Sato and
H. Hashimoto, Carbohydr. Res., 1993, 247, 179–193.
30 G. J. Boons and A. V. Demchenko, Chem. Rev., 2000, 100, 4539.
31 V. Wittmann and C. H. Wong, J. Org. Chem., 1997, 62, 2144–
2147.
32 M. Clark, R. D. I. Cramer and N. J. van den Opdenbosch, Comput.
Chem., 1989, 10, 982–1012.
33 A. Imberty, E. Bettler, M. Karababa, K. Mazeau, P. Petrova and
S. Pérez, in Perspectives in Structural Biology; Indian Academy of
Sciences and Universities Press, Hyderabad, 1999, pp 392–409.
34 S. Perez, K. Mazeau, A. Imberty, M. Karababa and E. Bettler, http://
webenligne.cermav.cnrs.fr/databank/monosaccharides/index.html.
35 F. Delaglio, S. Grzesiek, G. Vuister, G. Zhu, J. Pfeifer and A. Bax,
J. Biol. NMR, 1995, 6, 277–293.
References
1 A. Varki, Glycobiology, 1993, 3, 97–130.
2 R. A. Dwek, Chem. Rev., 1996, 96, 683–720.
3 S. W. Homans, Glycobiology, 1993, 3, 551–555.
4 A. Imberty and S. Perez, Chem. Rev., 2000, 100, 4567–4588.
5 C. F. Brewer and T. K. Dam, Chem. Rev., 2002, 102, 387.
6 J. P. Carver and D. A. Cumming, Pure Appl. Chem., 1987, 59, 1465–
1476.
7 A. J. Petrescu, S. M. Petrescu, R. A. Dwek and M. R. Wormald,
Glycobiology, 1999, 9, 343–352.
8 M. C. Galan, A. P. Venot, A. P. J. Glushka, A. Imberty and G. J.
Boons, J. Am Chem. Soc., 2002, 124, 5964–5973.
9 D. R. Bundle, R. Alibes, S. Nilar, A. Otter, M. Warwas and
P. Zhang, J. Am. Chem. Soc., 1998, 120, 5317–5318.
10 M. Wilstermann, J. Balogh and G. Magnusson, J. Org. Chem., 1997,
62, 3659–3665.
11 N. A. Navarre, A. van Oijen, A. Imberty, A. Poveda, J. Jimenez-
Barbero, A. Cooper, M. A. Nutley and G. J. Boons, Chem. Eur. J.,
1999, 5, 2281–2294.
36 J. Keeler, D. Neuhaus and J. J. Titman, Chem. Phys. Lett., 1988, 146,
545–548.
37 L. Poppe and H. van Halbeek, J. Magn. Reson., 1991, 93, 214–217.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 8 9 1 – 3 8 9 9
3899