Synthesis, conformation and T-helper cell stimulation of an O-linked glycopeptide epitope containing extended carbohydrate side-chains

Article abstract of DOI:10.1016/S0968-0896(02)00388-7

To answer the question whether or not T cells to immunodominant protein fragments recognize glycosylated antigens, we synthesized a series of glycopeptides corresponding to peptide 31D, a major T-helper cell epitope of the rabies virus nucleoprotein. Thr4 of the epitope is known to allow mono- or disaccharide side-chain substitutions in either α- or β-anomeric configuration without interfering with MHC-binding. To model naturally occurring glycoprotein fragments that carry extended sugar chains, we prepared Fmoc-Ser/Thr-OPfp building blocks containing α- and β-linked linear tri- and heptasaccharides. Peptide 31D was synthesized with the complex carbohydrates attached to Thr4, and the T-helper cell activity of the glycopeptides was determined. Addition of α-linked carbohydrates, that mimic most of the natural O-linked glycoproteins, resulted in a major drop in the T-cell stimulatory ability in a sugar length-dependent manner. In contrast, the cytosolic glycoprotein mimicking β-linked glycopeptides retained their T-cell stimulatory activity, with the trisaccharide-containing analogue being almost as potent as the unglycosylated peptide. When the peptides were preincubated with diluted human serum, all peptides lost their ability to stimulate the 9C5.D8-H hybridoma. These findings indicated that (i) in contrast to cytosolic glycosylation, incorporation of long O-linked carbohydrates into T-helper cell epitopes abrogates the antigenicity of these protein fragments, and (ii) glycosylation is not a viable alternative to improve the immunogenic properties of subunit peptide vaccines. Glycosylation with all four carbohydrate moieties similarly destroyed the inducible α-helical structure of peptide 31D as detected by CD, indicating that the differences in the T-cell activity were not due to different peptide conformations.

Full text of DOI:10.1016/S0968-0896(02)00388-7

Bioorganic & Medicinal Chemistry 10 (2002) 3859–3870
Synthesis, Conformation and T-Helper Cell Stimulation of an
O-Linked Glycopeptide Epitope Containing Extended
Carbohydrate Side-Chains
Mare Cudic, Hildegund C.J. Ertl and Laszlo Otvos, Jr.*
The Wistar Institute, 3601 Spruce Street, Philadelphia, PA 19104, USA
Received 10 September 2001; accepted 7 December 2001
Abstract—To answer the question whether or not T cells to immunodominant protein fragments recognize glycosylated antigens,
we synthesized a series of glycopeptides corresponding to peptide 31D, a major T-helper cell epitope of the rabies virus nucleo-
protein. Thr4 of the epitope is known to allow mono- or disaccharide side-chain substitutions in either a- or b-anomeric configur-
ation without interfering with MHC-binding. To model naturally occurring glycoprotein fragments that carry extended sugar
chains, we prepared Fmoc-Ser/Thr-OPfp building blocks containing a- and b-linked linear tri- and heptasaccharides. Peptide 31D
was synthesized with the complex carbohydrates attached to Thr4, and the T-helper cell activity of the glycopeptides was deter-
mined. Addition of a-linked carbohydrates, that mimic most of the natural O-linked glycoproteins, resulted in a major drop in the
T-cell stimulatory ability in a sugar length-dependent manner. In contrast, the cytosolic glycoprotein mimicking b-linked glyco-
peptides retained their T-cell stimulatory activity, with the trisaccharide-containing analogue being almost as potent as the
unglycosylated peptide. When the peptides were preincubated with diluted human serum, all peptides lost their ability to stimulate
the 9C5.D8-H hybridoma. These findings indicated that (i) in contrast to cytosolic glycosylation, incorporation of long O-linked
carbohydrates into T-helper cell epitopes abrogates the antigenicity of these protein fragments, and (ii) glycosylation is not a viable
alternative to improve the immunogenic properties of subunit peptide vaccines. Glycosylation with all four carbohydrate moieties
similarly destroyed the inducible a-helical structure of peptide 31D as detected by CD, indicating that the differences in the T-cell
activity were not due to different peptide conformations.
# 2002 Elsevier Science Ltd. All rights reserved.
Introduction
is towards side-chain substituted peptide antigens. Pro-
teins expressed in a diversity of intracellular sites,
including the cell surface, endoplasmic reticulum and
cytosol can gain access to the MHC molecule, albeit
with different efficiencies.4
The way the immune system presents antigens for T-cell
stimulation is well understood by now. The major histo-
compatibility complex (MHC) class I and class II anti-
gen-processing pathways play an essential role in the
activation of pathogen-specific T lymphocytes by pre-
senting peptide fragments derived from foreign pro-
teins.1 The basic unit of recognition is a trimolecular
complex consisting of the T-cell receptor (TcR), the
MHC molecule, and the MHC-bound peptide ligand.2
In the first step, the processed peptide antigens bind to
the MHC determinant followed by recognition by the
TcR coming from the opposite direction.3 Whereas no
T-cell stimulation occurs without snug fit to the MHC
proteins, it is not clear how tolerant the T-cell receptor
Glycosylation is a very common co-translational modi-
fication of proteins, and a basic immunological question
is whether native glycopeptides can be included in the
natural T-cell epitope repertoire.5 For peptides to have
antagonistic activity for T-cells (targeting, for example,
autoimmune diseases), binding to the MHC complex is
a prerequisite that is not necessarily fulfilled by syn-
thetic glycopeptides.6 While N-terminal glycosylation of
an MHC class II-restricted hen egg lysozyme-originated
epitope results in increased MHC-binding of synthetic
peptides, internal glycosylation abolishes the ability of
the same peptide to bind MHC.7 Among the internally
glycosylated analogues the study did not find any dif-
ference in binding when the sugar was positioned on an
*Corresponding author. Tel.: +1-215-898-3772; fax: +1-215-898-
5821; e-mail: otvos@wistar.upenn.edu
0968-0896/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(02)00388-7

3860
M. Cudic et al. / Bioorg. Med. Chem. 10 (2002) 3859–3870
asparagine or on an adjacent serine. Haurum and co-
workers demonstrated that a MHC class I binding pep-
tide can be modified by both a- and b-linked O-glyco-
sylation as well as by b-linked N-glycosylation without
affecting its binding to MHC class I, provided the
modifications do not involve amino acids located at
MHC anchor positions.8 Some of the glycopeptide-
MHC complexes retain their ability to stimulate cyto-
toxic T-cell lymphocyte (CTL) lines, but another subset
of the glycosylated analogues are not able to bind the
TcR anymore.8 In another experiment, recognition by T
cells, directed to a non-glycosylated decapeptide, is lost
when a single galactosyl moiety is attached N-terminally
to the antigens.9 This decapeptide is likely the shortest
core sequence that binds to the corresponding MHC
class II, and any further modification, such as glyco-
sylation is not tolerated. T-cells directed to the N-term-
inally glycosylated version of this peptide recognize
glycopeptides having significant variation in the sac-
charide structure. The authors proposed that these T
cells recognize a peptide structure specific to glycopep-
tide-MHC complexes, and that the recognition does not
involve specific interaction between the carbohydrate
moiety and the T-cell receptor. Later their view was
modified when they discovered a new, glycopeptide-
specific set of T-cells. They demonstrated that the new
T-cells are specific for the immunizing galabiose, and do
not recognize the same peptide coupled to cellobiose.10
Likewise, when neo-glycopeptides are used to generate
CTL, the polyclonal cell population preferentially kills
target cells treated with glycopeptide compared to those
treated with the core peptide.11 The resulting glycopeptide-
specific CTL and unrestricted carbohydrate-specific
CTL populations exhibit different TcR expression
panels.
reduce the candidate protein fragments in the process of
mapping the antigenic regions. Prediction methods for
identifying epitopic peptides are ever-so-popular,16ꢀ18
because a successful algorithm would reduce the num-
ber of peptides required to be synthesized and assayed,
and thereby greatly facilitate the identification of T-cell
epitopes of proteins.19
Earlier we found that mid-chain incorporation of
b-linked mono- and disaccharides into asparagines
greatly diminishes the T-cell stimulatory activity due to
the reduced ability of the glycopeptides to bind to MHC
class II.20,21 This finding seems to support the idea
about the lack of natural N-glycosylated epitopes.
b-O-glycosylation with short sugars similarly decreases
the T-cell stimulation, but the extent of the decrease is
dependent upon the peptide sequence and the host
amino acid.21,22 Our data on the short sugar-containing
glycopeptides indicate that in contrast to glycosylated
asparagine, the T-cell epitope repertoire may contain
both a- and b-linked O-glycopeptides. Since many nat-
ural O-glycosylated protein fragments (e.g., mucins,
blood group antigens, etc.) contain longer carbohydrate
side-chains,23 the next question we addressed in the
current report was how the carbohydrate length affected
the peptides’ ability to stimulate T-helper cells. Our
model is peptide 31D, an immunodominant epitope
delineated from the rabies virus nucleoprotein.24 The
sequence of peptide 31D contains a threonine residue in
position 4, which was earlier shown not to be involved
in binding to the I-Ek MHC II molecule, and could be
decorated with O-linked short carbohydrates without
interfering with MHC binding.25 We also studied the
conformation-modifying effect of incorporation of long
sugars into T-cell epitopic peptides. To complete the
immunological and structural studies first the synthetic
procedures for the preparation of O-glycopeptides con-
taining a- and b-linked long carbohydrates had to be
worked out. Here we report the synthesis of not only the
glycosylated threonine building blocks used for the
immunological and structural studies, but of the analo-
gous serine derivatives as well. These glycopeptide pre-
cursors will open new avenues in studying an array of
biological, biochemical and pharmacological properties
of model glycopeptides and fragments of native glyco-
proteins.
As outlined above, experiments using model T-cell epi-
topes have demonstrated that carbohydrate can mod-
ulate T-cell responses in a variety of ways. Significantly,
short sugar-containing glycopeptide-specific T-cell
responses have been detected to native glycoproteins.12
While investigations using viral glycoproteins suggest
that the large N-linked glycans influence the T cell
response to glycoproteins but that this does not involve
specific interactions between the carbohydrate and the
TcR, the smaller O-linked carbohydrates found on col-
lagen or mucins are able to elicit a carbohydrate specific
T cell responses.13 The ability of carbohydrate to influ-
ence T-cell recognition of antigen has important con-
sequences for a broad range of immune responses.
Results and Discussion
Synthesis of O-glycosylated amino acid building blocks
Peptides presented by human class I MHC molecules in
vivo encompass up to 0.1% glycopeptides, containing
the cytosolic O-b-GlcNAc substitution, and synthetic
peptides carrying this substitution are efficiently trans-
ported into the endoplasmic reticulum.14 Studies on the
immunogenicity of glycosylated type II collagen indi-
cate that antigen presenting cells can indeed process
glycoproteins to glycopeptides, which are presented to
MHC class II proteins and the complex elicits T cell
responses.15 In addition to a pure scientific inquiry, if
naturally glycosylated epitopes were excluded from the
antigenic pool, this knowledge would significantly
For modeling the extended carbohydrate chains
attached to the fragments of natural O-glycoproteins we
returned to oligomers of maltose. These were used suc-
cessfully earlier for the immunological and confor-
mational analysis of N-linked glycopeptides.26,27 The
starting compounds were maltotriose, maltohexaose
and maltoheptaose. These oligosaccharides were
acetylated with acetic acid anhydride/pyridine.28 Mix-
tures of a- and b-anomers of peracetylated sugars (for
maltotriose the a:b ratio was 1:1, and for maltohexaose
and maltoheptose 3:1) were separated by silica gel

M. Cudic et al. / Bioorg. Med. Chem. 10 (2002) 3859–3870
3861
chromatography. The anomeric configuration was con-
verted in order to obtain b-anomers of the peracetylated
sugars, needed for the ensuing synthetic strategy. As a
first step, the a-anomers were converted into acetylated
glycosyl bromides with titanium tetrabromide.29 This
unusual protocol had to be employed, because the more
conventional treatment with hydrogen bromide in acetic
acid28 failed to provide the target products due to
intense decomposition, especially in case of malto-
hexaose and maltoheptaose. After bromination, the
halogen was replaced by an acetate group using mercu-
ric acetate in acetic acid, concomitant with simple
inversion of the anomeric bonds (Scheme 1).
longer sugars because apparently an increase in the size
of the carbohydrate is accompanied by an increase in
the yield of the glycosylated Fmoc-Ser/Thr-OH deriva-
tives.31 Trisaccharide 1 and heptasaccharide 3 were
coupled to Fmoc-Ser/Thr-OH using a Lewis acid
ꢁ
(BF3 Et2O) as a promoter. After the crude products
were purified by flash chromatography, the carboxyl
group of the amino acids was activated with penta-
fluorophenol to afford Fmoc-protected and b-glycosy-
lated Thr 4 and 6 as well as the corresponding Ser 5 and
7 derivatives in relatively high yields (60, 48, 55, and
44%, respectively) (Scheme 2). Clearly, our yields using
long saccharides were comparable or better than those
reported for the same derivatives carrying mono-
saccharide side-chains. The purity of the final O-glyco-
sylated Ser/Thr building blocks was confirmed by
HPLC chromatography and MALDI-MS (Table 1).
The next step was to prepare Fmoc-protected glycosyl
serine and glycosyl threonine building blocks for the
solid-phase synthesis of O-glycopeptides. The protocol
introduced by Salvador et al.30 gives an easy access to
the target b-O-glycosylated building blocks, which carry
the protective groups of choice (O-acetyl and N-a-Fmoc)
for automated glycopeptide synthesis. We did not
expect any major difficulty in applying this strategy for
The formation of the a-glycosidic linkage can only be
achieved if a nonparticipating substituent at C-2 of the
glycosyl donor is used in a glycosylation reaction. The
azido group was found to be efficient for such stereo-
Scheme 1. Reagents: (i) Ac2O/pyridine; (ii) TiBr4; (iii) Hg(CH3COO)2.

3862
M. Cudic et al. / Bioorg. Med. Chem. 10 (2002) 3859–3870
Scheme 2. Reagents: (i) BF3–OEt2; (ii) Pfp–OH, DIC.
Table 1. Synthetic O-glycosylated amino acid building blocks and their characterization
Compound
RP-HPLC retention
time (min)
MALDI mass spectrometry
Calculated
(Da)
Observed
(M+Na)+ (m/z)
Fmoc-Thr(maltotriose b1!O)-OPfp 4
33.3
31.7
37.7
36.5
33.1
31.8
38.0
36.9
1413
1399
2566
2552
1396
1382
2549
2535
1437
1423
2589
2575
1419
1406
2572
2558
Fmoc-Ser(maltotriose b1!O)-OPfp 5
Fmoc-Thr(maltoheptose b1!O)-OPfp 6
Fmoc-Ser(maltoheptose b1!O)-OPfp 7
Fmoc-Thr(maltose-GalN3 a1!O)-OPfp 18
Fmoc-Ser(maltose-GalN3 a1!O)-OPfp 19
Fmoc-Thr(maltohexaose-GalN3 a1!O)-OPfp 20
Fmoc-Ser(maltohexaose-GalN3 a1!O)-OPfp 21
selective a-glycosylation and for the subsequent conver-
sion into an acetamido function, which is predominantly
present at C-2 in natural O-glycoproteins.29 The forma-
tion of a-linked glycosides of serine and threonine can be
further promoted by the presence of fused cyclic acetal at
the C-4 and C-6 hydroxyls of N-acetylgalactosamine.32
Accordingly, methyl 2-azido-4,6-di-O-benzylidene-2-
deoxy-b-d-galactopyranoside,29,33ꢀ35 was used for the
preparation of the a-O-glycosylated building blocks.
Our strategy towards the synthesis of a-anomers 18–
21 with extended carbohydrate chains involved the
preparation of fully acetylated trisaccharide/hepta-
saccharide bromides 16, 17 as glycosyl donors which
were then coupled to Fmoc-Ser/Thr pentafluorophenyl
esters to yield Fmoc - Ser/Thr(Ac7 - maltose - Ac2 - Gal-
N3/Ac19-maltohexaose-Ac2-GalN3-a1!O)-OPfp 18–21
(Scheme 3). Peracetylated b-sugars, b-d-maltose octa-
acetate and the peracetylated maltohexaose 2, were
treated with thiophenol in the presence of boron tri-
fluoride etherate to give the corresponding phenyl-
b-thioglycosides 8 and 9.36 The trisaccharide/hepta-
saccharide glycosyl donors were then prepared from
peracetylated maltose/maltohexaose phenyl b-1-thioglyco-
sides 8 and 9. These carbohydrates underwent a con-
densation reaction with methyl 2-azido-4,6-di-O-
benzylidene-2-deoxy-b-d-galactopyranoside promoted by
N-iodosuccinimide-triflic acid to exclusively afford the
b-anomers 10 and 11 in good yields. The thioglycoside
donors were used on the basis of their superiority com-
pared with trichloroacetimidate donors because of the
ease of their preparation and the high yields they are
reported to provide.35 The benzylidine protective group
was removed in aqueous acetic acid at 80 ꢂC resulting in
the formation of partially acetylated glycosyl azides 12
and 13. Acetylation of the hydroxyl groups and ace-
tolysis of the anomeric methyl glycoside were carried
out in a single step using a mixture of acetic acid and
sulfuric acid (50:1, v:v), affording anomeric mixtures of
the peracetylated trisaccharides or heptasaccharides 14
or 15. These peracetylated carbohydrates were converted
to the corresponding glycosyl bromides 16 and 17, now
in a-anomeric configuration, using titanium tetra-
bromide. Finally, a modified Koenigs-Knorr glyco-
sylation reaction was employed for the sugar
incorporation into Fmoc-protected pentafluorophenyl
esters of serine and threonine. This reaction is based
on an observation that when the glycosylation is car-
ried out with AgClO4 at ꢀ40 ꢂC, the a-isomer is pre-
ferably formed.33 Accordingly, we coupled the Fmoc-
Ser/Thr-OPfp derivatives with glycosyl bromide donors
16 and 17 in the presence of silver perchlorate at ꢀ40 ꢂC
to obtain the stereochemically almost clean a-anomers
of Thr 18 and 20 as well as the Ser derivatives 19 and 21
in excellent yields (72, 63, 65, and 60%, respectively).
The synthesis of the a-linked glycoamino acids is out-
lined in Scheme 3. All synthesized glycoamino acid
derivatives were purified by silica gel chromatography.
The purity of the final O-glycosylated Ser/Thr building
blocks was confirmed by HPLC chromatography and
MALDI-MS (Table 1). MHC class II epitopic peptide
31D as well as its a- and b-glycosylated analogues were
then prepared by stepwise incorporation of unglyco-
sylated threonine or the corresponding glycoamino acids
4, 6, 18, and 20 during conventional solid-phase peptide
synthesis. Similar to the glycoamino acid building
blocks, the final glycopeptides were characterized by

M. Cudic et al. / Bioorg. Med. Chem. 10 (2002) 3859–3870
3863
Scheme 3. Reagents: (i) PhSH, BF3–OEt2; (ii) PhCH(OMe)2, TsOH; (iii) NIS, TfOH, ꢀ45 ꢂC; (iv) 80% AcOH, 80 ꢂC; (v) Ac2O–H2SO4, ꢀ25 ꢂC; (vi)
TiBr4. The final coupling to Fmoc-Thr/Ser-OPfp utilized AgClO4 at ꢀ45 ꢂC.
RP-HPLC and MALDI-MS (Table 2). Figure 1 shows
the RP-HPLC profile and the mass spectrum of the
most complex glycopeptide, carrying an a-linked hepta-
saccharide side-chain attached to Thr4. One dimen-
sional NMR spectra of the 31D glycopeptides
supported the expected structures. The same glycoamino
acid derivatives were used to synthesize a series of glyco-
peptides corresponding to the human epithelial mucin 1
protein. These mucin glycopeptides were characterized
by homonuclear two-dimensional NMR experiments.
Identity of the products and their anomeric configur-
ation was confirmed by comparison of the protein shifts
with those of model compounds in the sugar database
and by analysis of the TOCSY and NOESY spectra.

3864
M. Cudic et al. / Bioorg. Med. Chem. 10 (2002) 3859–3870
Table 2. Characterization of the synthetic peptide (31D) and its glycosylated analogues
Peptide
Sequence
Side-chain modification (X)
RP-HPLC
(min)
(M+H)+
(m/z)
(Calculated,
Da)
31D
AVYTRIMMNGGRLKR
AVYT(X)RIMMNGGRLKR
AVYT(X)RIMMNGGRLKR
AVYT(X)RIMMNGGRLKR
AVYT(X)RIMMNGGRLKR
—
29.9
24.5
26.7
23.8
26.2
1766
2252
2293
2901
2941
1765
2251
2292
2900
2940
31D-Tb-(Glc)3
31D-Ta-(Glc)2GalNAc
31D-Tb-(Glc)7
31D-Ta-(Glc)6GalNAc
Glc(a1–4)Glc(a1–4)Glc(b1-O)
Glc(a1–4)-Glc(b1–4)-GalNAc(a1-O)
[Glc(a1–4)-Glc]3-Glc(b1-O)
[Glc(a1–4)-Glc]3-GalNAc(a1-O)
Figure 2. Stimulation of the 9C5.D8.H T-cell hybridoma with syn-
thetic peptides. The solid line (crosses) represents the unmodified
native peptide epitope 31D. The other four analogues carried carbo-
hydrate side chains. The solid symbols represent trisaccharide sub-
stitutions, and the open symbols correspond to heptasaccharide
substitutions. The two curves with circles represent a-linked sugars,
and the triangles correspond to the b-derivatives. The glycopeptides are:
31D-Tb-(Glc)3 (dots), 31D-Tb(Glc)7 (dashes); 31D-Ta-(Glc)2GalNAc
(double dots and dashes), and 31D-Ta-(Glc)6GalNAc (dots and dashes).
The positive control was a rat concanavalin A solution (RCAS), reg-
ularly used to propagate the lymphokine-dependent HT-2 cell line.
Figure 1. Reversed-phase high performance liquid chromatogram and
matrix-assisted laser ionization/desorption mass spectrum (inset) of
peptide 31D-Ta-(Glc)6GalNAc. The glycopeptide eluted at 26.2 min
as a single peak and exhibited the expected molecular ion of 2941 Da.
The peak at 1472 m/z unit is the double charged mass spectrometry
product.
High resolution conformational analysis of the pro-
ducts will be published elsewhere.
MHC-binding. Either the TcR was unable to accom-
modate the bulky carbohydrate, or, less likely, the
sugar interfered with the formation of the trimolecular
complex needed to elicit CD4+ T-cell responses. The
same glycopeptide remained somewhat stimulatory,
when the heptasaccharide was attached in the non-
natural b-anomeric configuration. This further sup-
ported the idea, that the MHC groove could accom-
modate the long carbohydrate side-chain attached to
Thr4.
T-cell stimulation
To answer the question how O-linked extended car-
bohydrates modulate T cell responses, the 31D glyco-
peptides were tested for their ability to stimulate T
helper cells. Peptide 31D, in context with the I-Ek
MHC determinant, vigorously stimulates cytokine
secretion by the 9C5.D8-H T-cell hybridoma. From
our earlier studies we know that Thr4 can carry mono
or disaccharide side-chains in either a- or b-anomeric
configuration without interfering with MHC-binding.
In the current study, addition of a-linked carbo-
hydrates, that mimic the sugar structures of natural
O-linked glycoproteins, resulted in a major drop in the
T-cell stimulatory ability in a sugar length-dependent
manner (Fig. 2). In contrast, the b-linked glycopeptides
retained most of their T-cell stimulatory activity, at
least at higher antigen concentrations, with the tri-
saccharide-containing analogue being as potent as the
non-glycosylated peptide (Fig. 2). These experiments
demonstrated that the stimulation of T-helper cells
was diminished if an extended a-linked sugar moiety
was attached to a Thr residue, not involved in
When the glycopeptides were preincubated in 50%
human serum for 3 h, all peptides lost their T-cell sti-
mulatory ability, probably due to cleavage by serum
proteases. Although some peptides were probably
cleaved by the 10% fetal bovine serum (FBS) used as a
medium for the T-cell stimulation assay, this proteoly-
tic strength has been shown to be inadequate to cleave
significant amounts of peptide 31D or its analogues.37
These findings indicated that (i) long and short sugars
coupled to hydroxyamino acids in native T cell epi-
topes are non-equivalent in terms of antigenicity, and
(ii) glycosylation of peptide epitopes is not an efficient
alternative to improve the half-live, and consequently
the immunogenic properties of subunit peptide vac-
cines.

M. Cudic et al. / Bioorg. Med. Chem. 10 (2002) 3859–3870
3865
These results are in accord with earlier studies on the
antigenicity and immunogenicity of MHC class II
restricted O-linked glycoproteins or synthetic glyco-
peptides that carry shorter sugar side-chains. When
T-cells were generated to various O-glycosylated frag-
ments of type II collagen, most of the resulting hybri-
domas specifically recognized the collagen peptide
glycosylated with a monosaccharide.38 Likewise, while
glycopeptides carrying mono- and disaccharides
attached to the mouse hemoglobin-derived T helper cell
epitope Hb(67–76) elicit measurable immune responses,
the same peptide with a natural mucin core 2 tri-
saccharide structure remains without any immunogenic
properties.39 In the third example the disaccharide pre-
sent in the T antigen is extended with sialylation to yield
a trisaccharide-coupled peptide epitope. Regardless of
whether the carbohydrate is attached to a serine or a
threonine residue, the peptide carrying trisaccharide is
always less immunogenic than the analogous peptide
coupled to the disaccharide moiety.40 Although the
authors explain these findings with the shielding effects
of sialic acid in cellular recognition processes, a simple
sugar-length dependent inhibition of interaction with
the T-cell receptor can not be excluded either. To the
best of our knowledge, O-glycopeptides containing
longer than trisaccharides have not been studied for
their T-cell stimulatory properties.
effect of long carbohydrates attached to the peptide
backbone is very limited. Our data provided some
insight into the natural conformation of glycoproteins
and their fragments.
In 50% aqueous trifluoroethanol, unglycosylated pep-
tide 31D exhibited a loosened a-helical structure (Fig.
3). While the general shape of the CD curve resembled
those of classical a-helices,42 the blueshift of the bands
representing the pp* transition, especially that of the
negative band to 205 nm, clearly documented a desta-
bilized helical conformation. The calculated a-helix
content based on the ellipticity value at 208 nm43
reached 28%. Glycosylation with all four carbohydrate
moieties similarly destroyed the inducible a-helical
structure of peptide 31D (Fig. 3). This conformation-
modifying effect could be seen from the blueshift of the
pp* bands which was accompanied by the significant
decrease of the intensity of the positive band, as well as
from the considerable drop of the intensity ratio of the
222:201–205 nm bands. Actually all four glycopeptides
exhibited very similar CD spectra, characterized by a
positive band at 186 nm, a negative band at 201 nm and
a negative shoulder at 222 nm. These spectral features
resembled those of type I (III) b-turns or mixtures of
type I and type II turns.44 These data indicated that the
differences in the T-cell activity, as regulated by the
anomeric configuration of the carbohydrate side-chains,
were not due to different peptide conformations. Sig-
nificantly, the local secondary structure around the
glycosylation sites of O-glycopeptides appeared to be a
reverse-turn, regardless of the length or the anomeric
configuration of the carbohydrate moiety.
It is worth speculating on the T cell antigenicity of
native glycopeptides as regulated by the differences in the
anomeric configuration and the length of the carbo-
hydrate side-chains. While the less frequent cytosolic
glycosylation involves a single GlcNAc residue coupled
to serines or threonines via a b-anomeric linkage,41 most
other glycoproteins contain a-linked extended carbo-
hydrate side-chains. In accord with earlier studies on
MHC class I epitopes, our b-linked trisaccharide-con-
taining 31D variant retained most of the antigenic
properties of the native peptide epitope. Apparently this
type of glycosylation, native or artificial, does not
interfere with binding to the MHC determinants or the
recognition by the TcR. These and our earlier results
also indicate that our particular 9C5.D8-H T cell
hybridoma does not directly bind to the hydroxyl group
of Thr4. For all other a-linked glycoprotein-mimicking
glycopeptides, if our single example can be generalized,
the interaction with the TcR is abrogated upon sugar
incorporation. Whether this reflects the lack of native
O-glycosylated epitopes for the a-linked glycoproteins
can be later investigated by using the various glyco-
sylated 31D fragments as immunogens.
Conformational analysis
Peptide 31D is a prototype of protein fragments
assuming an inducible a-helical structure.20 The con-
formational analysis of the 31D glycopeptides carrying
the extended sugar chains served two purposes. First,
we wanted to see whether the observed differences in the
T-cell stimulatory ability of the a- and b-linked carbo-
hydrates could be rationalized by different secondary
structures of the glycopeptides. Second, the current state
of knowledge concerning the conformation-modifying
Figure 3. Circular dichroism spectra of peptide 31D and its glycopep-
tide analogues. The spectra were acquired in 50% aqueous tri-
fluoroethanol at a peptide concentration of 0.2 mg/mL. The solid line
represents unmodified peptide 31D, the rest are the glycopeptides. The
substitutions are as follows: a-linked trisaccharide, peptide 31D-Ta-
(Glc)2GalNAc—short dashes; b-linked trisaccharide, peptide 31D-Tb-
(Glc)3—dots; a-linked heptasaccharide, peptide 31D-Ta-(Glc)6Gal-
NAc—long dashes; and b-linked heptasaccharide, peptide 31D-
Tb(Glc)7—dots and dashes.

3866
M. Cudic et al. / Bioorg. Med. Chem. 10 (2002) 3859–3870
Incorporation of O-linked carbohydrates into synthetic
peptides are reported to result in small, but measurable
conformational changes on the peptide backbone. In
accordance with our data, the 204 nm negative CD
band is blueshifted to 200 nm upon multiple glyco-
sylation of a glycophorin fragment, when alternating
serines and threonines are decorated with a GalNAc
monosaccharide moiety.45 Another stabilization of a
b-turn structure was documented by addition of two
GalNAc units at the two ends of the V3 loop of the
human immunodeficiency virus glycoprotein.46 O-Glyco-
sylation also appears to modulate the conformation of
synthetic mucin peptides,47,48 although the similarity of
the CD curves representing polyproline II helices and
unordered conformations as well as the lack of intra-
molecular hydrogen bonding, sources of NOE cross-
peaks in two-dimensional NMR, renders the polyproline
II conformation of free peptides in solution indis-
tinguishable from an irregular backbone structure.49
The literature generally lacks conformational studies on
O-linked glycopeptides with extended sugar chains. In
one of the rare reports, only marginal changes in the
backbone conformation was detected when a hexa-
peptide was decorated with two linear tetrasaccharide
moieties.50
Organizer) equipped with a Dynamax UV–vis-detector
Model UV-C (Rainin Technologies) and an CR501-
integrator (Shimadzu Corporation) was used. A Pheno-
menex 5 mm C18 300 A (4ꢃ250 mm) Jupiter column was
used at a flow rate 1 mL minꢀ1 for analytical runs, and
a Phenomenex 10 mm C18 300 A (21.2ꢃ250 mm) Jupiter
column was used at a flow rate 10 mL minꢀ1 for pre-
parative separations. Solvent A was 0.1% aqueous tri-
fluoroacetic acid and solvent B was 0.1% trifluoroacetic
acid in acetonitrile. The gradient is given as percentage
of solvent B in the solvent A/solvent B mixture.
Matrix-assisted laser desorption/ionization mass
spectrometry (MALDI-MS) was carried out at the
Wistar Institute Protein Microchemistry Laboratory on
a Voyager Biospectrometry Workstation by standard
methods. Fmoc-Thr(OtBu)-OPfp, Fmoc-Ser(OtBu)-
OPfp, Fmoc-Thr-OH and Fmoc-Ser-OH were pur-
chased from Bachem. Maltotriose, maltoheptaose, and
b-d-maltose octaacetate were purchased from Sigma,
methyl-2-azido-2-deoxy-b-d-galactopyranoside
from
Senn Chemicals, and silver perchlorate from Aldrich.
Thioacetic acid was purchased from Fluka.
Synthesis of O-glycosylated amino acid building blocks
In conclusion, we prepared glycoamino acid building
blocks, ready for incorporation into synthetic glyco-
peptides, carrying a- and b-linked trisaccharide and
heptasaccharide side-chains attached to serines and
threonines. These glycoamino acid derivatives will be
invaluable tools to study the biological and biochemical
effects the presence of carbohydrates make in natural
and artificial glycoprotein fragments. In our particular
example, an analogue of a native T-helper cell epitopic
peptide, carrying an a-linked extended sugar chain lost
its ability to stimulate the corresponding T-cell hybri-
doma. In contrast, the cytosolic-glycosylation mimick-
ing b-linked short trisaccharide did not interfere with
recognition by the TcR. Threonine-bound long sugars
appear to break helices and stabilize reverse-turns.
b-Anomers of peracetylated sugars (maltotriose, malto-
hexaose, and maltoheptaose) (1–3) were prepared
according to standard methods by acetylation of
unprotected d-carbohydrates, and an a!b anomeric
conversion of the fully acetylated sugars.28
Synthesis of Fmoc-Ser/Thr(Ac10-maltotriose/Ac22-malto-
heptaose ꢀ1!O)-OPfp
ꢁ
The Lewis acid, BF3 Et2O (39 mL, 0.3 mmol) was added
to the b-peracetylated tri-/heptasaccharide (0.1 mmol)
and the N-a-Fmoc-Ser/Thr-OH (0.12 mmol) in dry
CH2Cl2 (10 mL) under argon atmosphere at room tem-
perature. The progress of the glycosylation reaction was
monitored by TLC and analytical HPLC (isocratic elu-
tion of 50% solvent B for 5 min followed by a linear
gradient from 50 to 100% solvent B for 45 min). When
the reaction was found not to progress further, the
reaction mixture was diluted with CH2Cl2 (15 mL),
washed with 1 M aqueous HCl (5 mL) and water (15
mL), dried over sodium sulfate (Na2SO4) and con-
centrated in vacuum. The residue was purified by flash
chromatography (EtOAc/toluene/HOAc=10:5:1, v/v).
Pentafluorophenol (37 mg, 0.2 mmol) and 1,3-diiso-
propylcarbodiimide (31 mL, 0.2 mmol) were added to the
Experimental
General procedures
All air- and moisture-sensitive reactions were performed
under argon atmosphere. The solvents, purchased from
Fluka, were devoid of any trace of water and kept over
molecular sieves. Acid-washed molecular sieves (AW-
300), activated powdered molecular sieves, and Celite
(Celite 521) were purchased from Aldrich. The progress
of the reactions was monitored by thin-layer chroma-
tography (TLC) on Merck Silica Gel 60 F254 plates,
visualized by charring with 10% aqueous sulfuric acid
and/or by UV light when applicable. Flash-column
chromatography was performed on Silica Gel columns
(Merck, 230–400 mesh).
purified
Fmoc-Ser/Thr(Ac10-maltotriose/Ac22-malto-
heptaose b1!O)-OH (0.1 mmol), dissolved in dry
CH2Cl2 (10 mL), at room temperature. After the active
ester formation was completed (approximately 2–3 h),
the mixture was washed with water (10 mL), dried and
concentrated by rotary evaporation. The crude products
were purified by RP-HPLC (preparative runs using an
isocratic elution of 50% solvent B for 5 min followed by
a linear gradient from 50 to 100% solvent B for 45 min).
The purity of the final b1!O-glycosylated Ser/Thr
building blocks was confirmed by RP-HPLC chro-
matography and MALDI-MS.
For the reversed-phase HPLC a Beckman HPLC system
(two 110B solvent delivery modules, a System Gold
analogue interface module 406, and a Beckman System

M. Cudic et al. / Bioorg. Med. Chem. 10 (2002) 3859–3870
3867
Nꢁ-(Fluoren-9-ylmethoxycarbonyl)-3-O-[2,3,4,6-tetra-O-
acetyl-ꢁ-D-glucopyranosyl-(1!4)-2,3,6-tri-O-acetyl-ꢁ-D
-glucopyranosyl-(1!4)-2,3,6-tri-O-acetyl-ꢀ-D -gluco-
pyranosyl]-L-threonine pentafluorophenyl ester (4). The
protected and activated glycoamino acid derivative 4 was
prepared by the condensation of peracetylated maltotriose
1 (100 mg, 0.1 mmol) with N-a-Fmoc-Thr-OH (42 mg, 0.12
mmol) according to the general procedure outlined in the
previous paragraph. Purification of the pentafluorophenyl
ester by flash chromatography followed by RP-HPLC
afforded the final trisaccharide-containing activated amino
acid derivative 4 (85 mg, 60%). MALDI-MS; calcd for
C63H68F5N1O30 1413.37, found (M+Na)+ 1436.50.
Phenyl 2,3,4,6-tetra-O-acetyl-ꢁ-D-glucopyranosyl-(1!4)-
2,3,6-tri-O-acetyl-1-thio-ꢀ-D-glucopyranoside (8). The
Lewis acid, BF3 Et2O (280 mL, 2.1 mmol) was added to
ꢁ
b-d-maltose octaacetate (500 mg, 0.7 mmol) and thio-
phenol (111 mL, 1.05 mmol) in dry CH2Cl2 (20 mL)
under argon atmosphere. After 12 h at room tempera-
ture, the reaction mixture was washed two times with
saturated aqueous sodium bicarbonate (NaHCO3), and
then with water. The separated organic layer was dried
over sodium sulfate (Na2SO4) and concentrated under
vacuum. The residue was purified by flash chroma-
tography (EtOAc:toluene=1:1, v/v) to obtain the thio-
glycoside 8 (320 mg, 58%). MALDI-MS; calcd for
C32H40O17S1 728.20, found (M+Na)+ 751.20.
Nꢁ-(Fluoren-9-ylmethoxycarbonyl)-3-O-[2,3,4,6-tetra-O-
acetyl-ꢁ-D-glucopyranosyl-(1!4)-2,3,6-tri-O-acetyl-ꢁ-
D-glucopyranosyl-(1!4)-2,3,6-tri-O-acetyl-ꢀ-D-gluco-
pyranosyl]-L-serine pentafluorophenyl ester (5). Compound
5 was prepared by reaction of 1 (100 mg, 0.1 mmol) with
N-a-Fmoc-Ser-OH (39 mg, 0.12 mmol). Purification of
the crude product by flash chromatography, and
RP-HPLC of the corresponding pentafluorophenyl ester
afforded the protected activated ester of serine 5 (77 mg,
55%). MALDI-MS; calcd for C62H66F5N1O30 1399.36,
found (M+Na)+ 1422.59.
Phenyl 2,3,4,6-tetra-O-acetyl-ꢁ-D-glucopyranosyl-(1!4)-
2,3,6-tri-O-acetyl-ꢁ-D-glucopyranosyl-(1!4)-2,3,6-tri-
O-acetyl-ꢁ-D-glucopyranosyl-(1!4)-2,3,6-tri-O-acetyl-
ꢁ-D-glucopyranosyl-(1!4)-2,3,6-tri-O-acetyl-ꢁ-D-gluco-
pyranosyl-(1!4)-2,3,6-tri-O-acetyl-1-thio-ꢀ-D-gluco-
pyranoside (9). In a similar manner, compound 9 was
prepared starting from peracetylated b-d-maltohexaose
2 (1500 mg, 0.82 mmol) and thiophenol (130 mL, 1.23
ꢁ
mmol) in dry CH2Cl2 (20 mL) under BF3 Et2O (328 mL,
2.46 mmol) promotion. Purification of the crude pro-
duct by flash chromatography (EtOAc:toluene=3:1,
v/v) afforded the hexasaccharide thioglycoside 9 (1150
mg, 75%). MALDI-MS; calcd for C80H104O49S1
1880.54, found (M+Na)+ 1903.78.
Nꢁ-(Fluoren-9-ylmethoxycarbonyl)-3-O-[2,3,4,6-tetra-O-
acetyl-ꢁ-D-glucopyranosyl-(1!4)-2,3,6-tri-O-acetyl-ꢁ-
D-glucopyranosyl-(1!4)-2,3,6-tri-O-acetyl-ꢁ-D-gluco-
pyranosyl-(1!4)-2,3,6-tri-O-acetyl-ꢁ-D-glucopyranosyl-
(1!4)-2,3,6-tri-O-acetyl-ꢁ-D-glucopyranosyl-(1!4)-
2,3,6-tri-O-acetyl-ꢁ-D-glucopyranosyl-(1!4)-2,3,6-tri-O-
acetyl-ꢀ-D-glucopyranosyl]-L-threonine pentafluorophenyl
ester (6). The protected and activated glycoamino acid
derivative 6 was prepared by the condensation of the
peracetylated heptasaccharide 3 (213 mg, 0.1 mmol)
with N-a-Fmoc-Thr-OH (42 mg, 0.12 mmol) according
to the general procedure described above. Purification
of the crude product by flash chromatography fol-
lowed by RP-HPLC afforded the final threonine deri-
vative 6 (125 mg, 48%). MALDI-MS; calcd for
C111H132F5N1O62 2565.71, found (M+Na)+ 2588.69.
Methyl 2,3,4,6-tetra-O-acetyl-ꢁ-D-glucopyranosyl-(1!4)-
2,3,6-tri-O-acetyl-ꢀ-D-glucopyranosyl-(1!3)-2-azido-
4,6-O-benzylidene-2-deoxy-ꢀ-D-galactopyranoside (10).
Phenyl b-thioglycoside 8 (720 mg, 0.96 mmol) and
methyl 2-azido-4,6-O-benzylidene-2-deoxy-b-d-galacto-
pyranoside (255 mg, 0.81 mmol) were dissolved in dry
CH2Cl2 (20 mL). Acid-washed molecular sieves (AW-
300, 1.8 g) were added and the reaction mixture was
stirred at room temperature for 1 h under argon atmo-
sphere. The flask was cooled down to ꢀ45 ꢂC and
N-iodosuccinimide (234 mg, 1.02 mmol) was added.
After a period of 10 min, trifluoromethane sulfonic acid
(84 mL, 0.96 mmol) in acetonitrile (1 mL) was added
slowly and the reaction mixture was vigorously stirred
for another 2 h. Triethylamine (0.1 mL) was added at
the same temperature to neutralize the acid and the
mixture was filtered through Celite. The organic filtrate
was then washed with saturated aqueous sodium thio-
sulfate (Na2S2O3) and sodium bicarbonate (NaHCO3).
The separated organic layer was dried over Na2SO4 and
was concentrated. The residue was purified by flash
chromatography (EtOAc/toluene=2:1, v/v) to obtain
the pure methyl glycoside 10 (495 mg, 66%). MALDI-
MS; calcd for C40H51N3O22 925.30, found (M+Na)+
948.31.
Nꢁ-(Fluoren-9-ylmethoxycarbonyl)-3-O-[2,3,4,6-tetra-O-
acetyl-ꢁ-D-glucopyranosyl-(1!4)-2,3,6-tri-O-acetyl-ꢁ-
D-glucopyranosyl-(1!4)-2,3,6-tri-O-acetyl-ꢁ-D-gluco-
pyranosyl-(1!4)-2,3,6-tri-O-acetyl-ꢁ-D-glucopyranosyl-
(1!4)-2,3,6-tri-O-acetyl-ꢁ-D-glucopyranosyl-(1!4)-
2,3,6-tri-O-acetyl-ꢁ-D-glucopyranosyl-(1!4)-2,3,6-tri-O-
acetyl-ꢀ-D-glucopyranosyl]-L-serine pentafluorophenyl ester
(7). Compound 7 was prepared by the reaction of the
heptasaccharide 3 (213 mg, 0.1 mmol) with N-a-Fmoc-
Ser-OH (39 mg, 0.12 mmol). Purification of the crude
product by flash chromatography, and RP-HPLC of the
corresponding pentafluorophenyl ester afforded the
protected activated ester of serine 7 (115 mg, 44%).
MALDI-MS; calcd for C110H130F5N1O62 2551.70,
found (M+Na)+ 2575.30.
Methyl 2,3,4,6-tetra-O-acetyl-ꢁ-D-glucopyranosyl-(1!4)-
2,3,6-tri-O-acetyl-ꢁ-D-glucopyranosyl-(1!4)-2,3,6-tri-O
-acetyl-ꢁ-D-glucopyranosyl-(1!4)-2,3,6-tri-O-acetyl-ꢁ-
D-glucopyranosyl-(1!4)-2,3,6-tri-O-acetyl-ꢁ-D-gluco-
pyranosyl-(1!4)-2,3,6-tri-O-acetyl-ꢁ-D-glucopyranosyl-
(1!3)-2-azido-4,6-O-benzylidene-2-deoxy-ꢀ-D-galacto-
pyranoside (11). In a similar manner, the methyl hexa-
Synthesis of Fmoc-Ser/Thr(Ac7-maltose-Ac2-GalN3/Ac19--
maltohexaose-Ac2-GalN3 ꢁ1!O)-OPfp. Methyl 2-azido-
4,6-O-benzylidene-2-deoxy-b-d-galactopyranoside was
prepared according to Paulsen et al.17

3868
M. Cudic et al. / Bioorg. Med. Chem. 10 (2002) 3859–3870
saccharide glycoside 11 was prepared from phenyl
b-thioglycoside 9 (1431 mg, 0.76 mmol) and methyl
2-azido-4,6-O-benzylidene-2-deoxy-b-d-galactopyranoside
(200 mg, 0.64 mmol). Purification of the crude product
by flash chromatography (EtOAc/toluene=3:1,v/v)
afforded product 11 (900 mg, 68%). MALDI-MS; calcd
for C88H115N3O54 2077.64, found (M+Na)+ 2100.97.
quantitative yields. MALDI-MS; calcd for C86H115
N3O57 2101.62, found (M+Na)+ 2124.95.
2,3,4,6-Tetra-O-acetyl-ꢁ-D-glucopyranosyl-(1!4)-2,3,6-
tri-O-acetyl-ꢀ-D-glucopyranosyl-(1!3)-2-azido-4,6-di-O-
acetyl-2-deoxy-ꢁ-D-galactopyranosyl bromide (16). The
acetyl trisaccharide glycoside 14 (250 mg, 0.26 mmol)
and titanium tetrabromide (195 mg, 0.53 mmol) were
stirred in dry CH2Cl2 and EtOAc mixture (10:1, v/v, 11
mL) for 12 h at room temperature. The solution was
diluted with dry toluene (50 mL), and anhydrous
sodium acetate was added until the mixture became
colorless. The reaction mixture was filtered through
Celite, and the filtrate was dried over Na2SO4 and con-
centrated in vacuum. The crude product was purified by
flash chromatography (EtOAc/toluene=2:1, v/v) to
obtain the pure glycosyl bromide 16 (100 mg, 40%).
MALDI-MS; calculated for C36H48Br1N3O23 969.19,
found (M+Na)+ 992.08.
Methyl 2,3,4,6-tetra-O-acetyl-ꢁ-D-glucopyranosyl-(1!4)-
2,3,6-tri-O-acetyl-ꢀ-D-glucopyranosyl-(1!3)-2-azido-2-
deoxy-ꢀ-D-galactopyranoside (12). To remove the
benzylidene protection, compound 10 (450 mg, 0.49
mmol) was heated with 80% aqueous acetic acid (5 mL)
for 2–3 h at 80 ꢂC. The resulting mixture was diluted
with toluene and evaporated to dryness. The crude
product was purified by flash chromatography (EtOAc/
toluene/HOAc=10:5:1, v/v/v) to obtain the pure
deprotected methyl trisaccharide 12 (250 mg, 61%).
MALDI-MS; calcd for C33H47N3O22 837.27, found
(M+Na)+ 859.88.
2,3,4,6-Tetra-O-acetyl-ꢁ-D-glucopyranosyl-(1!4)-2,3,6-
tri-O-acetyl-ꢁ-D-glucopyranosyl-(1!4)-2,3,6-tri-O-acetyl-
ꢁ-D-glucopyranosyl-(1!4)-2,3,6-tri-O-acetyl-ꢁ-D-gluco-
pyranosyl-(1!4)-2,3,6-tri-O-acetyl-ꢁ-D-glucopyranosyl-
(1!4)-2,3,6-tri-O-acetyl-ꢁ-D-glucopyranosyl-(1!3)-2-
azido-4,6-di-O-acetyl-2-deoxy-ꢁ-D -galactopyranosyl
bromide (17). In a similar manner, the acetyl hepta-
saccharide glycoside 15 (600 mg, 0.29 mmol) was used
to prepare the corresponding bromide 17. The reaction
was run in the presence of titanium tetrabromide (210
mg, 0.57 mmol). Purification of the crude hepta-
saccharide bromide by flash chromatography (EtOAc/
toluene=3:1, v/v) afforded the pure product 17 (300
mg, 50%). MALDI-MS; calcd for C84H112Br1N3O55
2121.53, found (M+Na)+ 2146.72.
Methyl 2,3,4,6-tetra-O-acetyl-ꢁ-D-glucopyranosyl-(1!4)-
2,3,6-tri-O-acetyl-ꢁ-D-glucopyranosyl-(1!4)-2,3,6-tri-O-
acetyl-ꢀ-D-glucopyranosyl-(1!4)-2,3,6-tri-O-acetyl-ꢁ-D-
glucopyranosyl-(1!4)-2,3,6-tri-O-acetyl-ꢁ -D -gluco-
pyranosyl-(1!4)-2,3,6-tri-O-acetyl-ꢁ-D-glucopyranosyl-
(1!3)-2-azido-2-deoxy-ꢀ-D-galactopyranoside (13). In a
similar manner, compound 13 was prepared by acidic
hydrolysis of the benzylidene derivative 11 (700 mg, 0.34
mmol). After flash chromatography (EtOAc/toluene/
HOAc=10:5:1, v/v/v) the methyl heptasaccharide 13
was obtained as a white solid (550 mg, 81%). MALDI-
MS; calcd for C81H111N3O54 1989.61, found (M+Na)+
2012.86.
Acetyl 2,3,4,6-tetra-O-acetyl-ꢁ-D-glucopyranosyl-(1!4)-
2,3,6-tri-O-acetyl-ꢀ-D-glucopyranosyl-(1!3)-2-azido-
4,6-di-O-acetyl-2-deoxy-1-O-D-galactopyranoside (14).
The solution of the methyl glycoside 12 (250 mg, 0.3
mmol) in acetic anhydride (3 mL), was stirred at
ꢀ20 ꢂC, and a precooled (0 ꢂC) mixture of acetic anhy-
dride and sulfuric acid [Ac2O/H2SO4 (50:1, v/v, 2 mL)]
was added to it during a time period of 5 min. The
acetylation reaction was allowed to proceed for 5 days
at-20 ꢂC. The mixture was diluted with cold CH2Cl2 (50
mL) and was carefully washed with aqueous sodium
bicarbonate solution (NaHCO3) and water. The organic
layer was dried over Na2SO4 and was evaporated to
complete dryness after adding and removing toluene.
Product 14 was obtained in practically quantitative
yields. MALDI-MS; calcd for C38H51N3O25 949.28,
found (M+Na)+ 972.32.
Nꢁ-(Fluoren-9-ylmethoxycarbonyl)-3-O-[2,3,4,6-tetra-O-
acetyl-ꢁ-D-glucopyranosyl-(1!4)-2,3,6-tri-O-acetyl-ꢀ-D-
glucopyranosyl-(1!3)-2-azido-4,6-di-O-acetyl-2-deoxy-
ꢁ-D-galactopyranosyl]-L-threonine pentafluorophenyl ester
(18). Bromide 16 (100 mg, 0.1 mmol) and Fmoc-Thr-
OPfp (102 mg, 0.2 mmol) were dissolved in a mixture of
dry CH2Cl2 and toluene (1:1, v/v, 5 mL). Activated pow-
dered molecular sieves (Aldrich, 150 mg) were added and
the reaction mixture was stirred at room temperature for 1
h under argon atmosphere. Then the reaction flask was
cooled down to ꢀ40ꢂC and silver perchlorate (AgClO4)
(21 mg, 0.1 mmol) was added. The progress of the glyco-
sylation reaction was monitored by TLC (EtOAc/tolu-
ene=2:1, v/v). After 2 h, a complete disappearance of the
starting glycosyl bromide 16 was observed. The reaction
mixture was allowed to warm up to room temperature,
was diluted with chloroform (10 mL) and was filtered
through Celite. The organic filtrate was washed with
water, separated, dried over Na2SO4 and concentrated.
The residue was purified by flash chromatography
(EtOAc/toluene=2:1, v/v) to obtain the pure activated
glycoamino acid 18 (102 mg, 72%). MALDI-MS; calcd
for C61H65F5N4O28 1396.37, found (M+Na)+ 1419.26.
Acetyl 2,3,4,6-tetra-O-acetyl-ꢁ-D-glucopyranosyl-(1!4)-
2,3,6-tri-O-acetyl-ꢁ-D-glucopyranosyl-(1!4)-2,3,6-tri-O
-acetyl-ꢁ-D-glucopyranosyl-(1!4)-2,3,6-tri-O-acetyl-ꢁ-
D-glucopyranosyl-(1!4)-2,3,6-tri-O-acetyl-ꢁ-D-gluco-
pyranosyl-(1!4)-2,3,6-tri-O-acetyl-ꢁ-D-glucopyranosyl-
(1!3)-2-azido-4,6-di-O-acetyl-2-deoxy-1-O-D-galacto-
pyranoside (15). In a similar manner, the acetyl hepta-
saccharide glycoside 15 was prepared starting from the
methyl derivative 13 (600 mg, 0.3 mmol). The reaction
proceeded smoothly and the product was isolated in
Nꢁ-(Fluoren-9-ylmethoxycarbonyl)-3-O-[2,3,4,6-tetra-O-
acetyl-ꢁ-D-glucopyranosyl-(1!4)-2,3,6-tri-O-acetyl-ꢀ-D-
glucopyranosyl-(1!3)-2-azido-4,6-di-O-acetyl-2-deoxy-

M. Cudic et al. / Bioorg. Med. Chem. 10 (2002) 3859–3870
3869
ꢁ-D-galactopyranosyl]-L-serine pentafluorophenyl ester
(19). In a similar manner, the serine derivative 19 was
prepared from glycosyl bromide 16 (100 mg, 0.1 mmol)
and Fmoc-Ser-OPfp (99 mg, 0.2 mmol) in the presence
of silver perchlorate (AgClO4) (21 mg, 0.1 mmol). The
crude product was purified by flash chromatography
(EtOAc/toluene=2:1, v/v) to obtain the pure activated
glycoamino acid 19 (90 mg, 65%). MALDI-MS; calcd
for C60H63F5N4O28 1382.35, found (M+Na)+ 1405.67.
retained. Still polymer-bound, the azido group on the
sugar ring was transformed into an acetamido group
using thioacetic acid for 4 days with 4 consecutive
changes of the reagent. After reducing the azido group,
the Fmoc protection on the N-terminal amino acid was
removed using 20% piperidine in dimethyl formamide
(DMF). The resulting glycopeptides as well unmodified
peptide 31D were cleaved from the solid support with
trifluoroacetic acid (TFA) in the presence of thioanisole
(5%), and water (5%) as scavengers for 2 h. Deacetyla-
tion of the sugar hydroxyl groups was accomplished by
a treatment with 0.1 M NaOH.20 A reaction time of 10
min was sufficient for the peptides containing tri-
saccharides, and 20 min was needed for those carrying
heptasaccharide side-chains. After cleavage, the pep-
tides were purified by RP-HPLC, (preparative runs:
isocratic elution of 5% solvent B for 5 min followed by
a linear gradient from 5 to 65% solvent B for 120 min).
The final products were characterized by RP-HPLC
(during the analytical runs the same gradient was
developed over a 45 min period) and MALDI-MS.
Nꢁ-(Fluoren-9-ylmethoxycarbonyl)-3-O-[2,3,4,6-tetra-O-
acetyl-ꢁ-D-glucopyranosyl-(1!4)-2,3,6-tri-O-acetyl-ꢁ-
D-glucopyranosyl-2,3,6-tri-O-acetyl-ꢁ-D-glucopyranosyl-
2,3,6-tri-O-acetyl-ꢁ-D-glucopyranosyl-2,3,6-tri-O-acetyl-
ꢁ-D-glucopyranosyl-2,3,6-tri-O-acetyl-ꢁ-D-glucopyranosyl-
(1!3)-2-azido-4,6-di-O-acetyl-2-deoxy-ꢁ-D-galacto-
pyranosyl]-L-threonine pentafluorophenyl ester (20). In a
similar manner, the activated threonine derivative car-
rying a long carbohydrate side-chain 20 was prepared
from the corresponding glycosyl bromide 17 (214 mg,
0.1 mmol) and Fmoc-Thr-OPfp (102 mg, 0.2 mmol) in
the presence of silver perchlorate (AgClO4) (21 mg, 0.1
mmol). The crude product was purified by flash chroma-
tography (EtOAc/toluene=3:1, v/v) to obtain the pure
glycoamino acid 20 (160 mg, 63%). MALDI-MS; calcd
for C109H129F5N4O60 2548.71, found (M+Na)+
2571.90.
Conformational analysis
Circular dichroism (CD) spectra were performed on a
Jasco J720 instrument at room temperature in a 0.2 mm
pathlength cell. Double-distilled water and spectroscopy
grade trifluoroethanol were used as solvents. Peptide
concentration was 0.2 mg/mL, as determined by RP-
HPLC. Curves were smoothed by algorithm provided
by Jasco. Mean residue ellipticity [ꢀ]MR is expressed in
deg cm2 dmolꢀ1 using mean residue weights calculated
from the molecular weight of the glycopeptides divided
by the number of amide bonds in the sequences.
Nꢁ-(Fluoren-9-ylmethoxycarbonyl)-3-O-[2,3,4,6-tetra-O-
acetyl-ꢁ-D-glucopyranosyl-(1!4)-2,3,6-tri-O-acetyl-ꢁ-
D-glucopyranosyl-2,3,6-tri-O-acetyl-ꢁ-D-glucopyranosyl-
2,3,6-tri-O-acetyl-ꢁ-D-glucopyranosyl-2,3,6-tri-O-acetyl-
ꢁ-D-glucopyranosyl-2,3,6-tri-O-acetyl-ꢁ-D-glucopyranosyl
(1!3)-2-azido-4,6-di-O-acetyl-2-deoxy-ꢁ-D-galacto-
pyranosyl] - L - serine pentafluorophenyl ester (21). In a
similar manner, the serine derivative 21 was prepared
from the glycosyl bromide 17 (214 mg, 0.1 mmol) and
Fmoc-Ser-OPfp (99 mg, 0.2 mmol) in the presence of
silver perchlorate (AgClO4) (21 mg, 0.1 mmol). The
crude glycoamino acid was purified by flash chroma-
tography (EtOAc/toluene=3:1, v/v) to obtain the pure
serine building block 21 (153 mg, 60%). MALDI-MS;
calcd for C108H127F5N4O60 2534.70, found (M+Na)+
2558.27.
Cytokine release assay
The 9C5.D8.H cell line, a fusion product between a
thymoma cell line and splenocytes from C3H/He mice
immunized with inactivated rabies virus of the Evelyn–
Rokitnicki–Abelseth (ERA) strain, was used for the
studies. These T-cell hybridomas secrete IL-2 in
response to peptide 31D presented by the I-Ek MHC
determinant as it was detailed earlier.24 The 9C5.D8.H
cells were grown in Dulbecco’s Modified Eagles med-
ium (DMEM) supplemented with 10% FBS, glutamine
and antibiotics. Low passage numbers were used for
the experiments. The IL-2 and IL-4 sensitive HT-2 cell
line was grown in DMEM supplemented with 10%
FBS, 10% rat concanavalin A supernatant and 1 mM
2-mercaptoethanol.
Solid-phase (glyco)peptide synthesis
Unmodified peptide 31D and the corresponding glyco-
sylated analogues were synthesized by conventional
solid-phase methods. The peptide chain assembly pro-
ceeded on a Rainin PS3 automated synthesizer and
using TentaGel S-Ram-Fmoc resin with an initial load
of 0.3 mmol/g (Advanced ChemTech, Louisville, KY).
Standard Fmoc-chemistry was used throughout51 with a
4-molar excess of the acylating amino acids and HATU
(1-hydroxy-7-azabenzotriazole uronium salt) activation,
recommended for the synthesis of complex peptides.52
The glycoamino acids were incorporated in the same
manner as the unmodified amino acids, except that they
were coupled in a 1.5 molar excess (to reduce glyco-
amino acid usage) in the presence of di-isopropyl-ethyl-
amine (DIPEA). On completion of the assembly of the
peptide chain, the N-terminal Fmoc-protection was
Splenocytes from young naive C3H/He mice were pre-
pared and irradiated with 1800 rads. Cells were plated
at a density of 3ꢃ106 cell/wells onto 48-well Costar
plates in 300 mL of DMEM supplemented with 10%
FBS and 1 mM 2-mercaptoethanol. Peptides were added
followed by the addition of 1ꢃ105 9C5.D8.H to each
well and the medium was adjusted to a final volume of 1
mL. Cell-free supernatants were harvested 24 h later; 75
mL of concentrated supernatant was added in triplicate
to round bottom microtiter plate wells that contained
2ꢃ103 HT-2 cells. The final volume was 150 mL. Plates

3870
M. Cudic et al. / Bioorg. Med. Chem. 10 (2002) 3859–3870
3
were pulsed 3 days later for 6 h with 0.5 mCi of [H]-
thymidine. Cells were then harvested onto filtermates
and incorporation of radioactivity was determined in a
b-counter. The specificity of this assay for this particular
peptide epitope and T-helper cell line was documented
earlier when under identical assay conditions peptide
31D failed to stimulate any other rabies virus-related T
cell clone.53
23. Lennarz, W. J. The biochemistry of glycoproteins and pro-
teoglycans; Plenum Press: NY, 1981.
24. Ertl, H. C. J.; Dietzschold, B.; Gore, M.; Otvos, L., Jr.;
Larson, J. K.; Wunner, W. H.; Koprowski, H. J. Virol. 1989,
63, 2885.
25. Prammer, K. V.; Ertl, H. C. J.; Otvos, L., Jr. Biomed.
Pept. Proteins Nucl. Acids 1995, 1, 163.
26. Urge, L.; Jackson, D. C.; Gorbics, L.; Wroblewski, K.;
Graczyk, G.; Otvos, L., Jr. Tetrahedron 1994, 50, 103.
27. Jackson, D. C.; Drummer, H. E.; Urge, L.; Otvos, L., Jr.;
Brown, L. E. Virology 1994, 199, 422.
Acknowledgements
28. Wolfrom, M. L.; Thompson, A. Methods Carbohydr.
Chem. 1963, 2, 211.
29. Paulsen, H.; Bielfeldt, T.; Peters, S.; Meldal, M.; Bock, K.
Liebigs Ann. Chem. 1994, 369.
30. Salvador, L. A.; Elofsson, M.; Kihlberg, J. Tetrahedron
1995, 51, 5643.
The authors wish to thank Dr. Walter Gerhard for cri-
tical reading of the manuscript. This work was sup-
ported by NIH grant GM45011.
31. Elofsson, M.; Roy, S.; Walse, B.; Kihlberg, J. Carbohydr.
Res. 1993, 246, 89.
References and Notes
32. Qiu, D.; Gandhi, S. S.; Koganty, R. R. Tetrahedron Lett.
1996, 37, 595.
33. Paulsen, H.; Peters, S.; Bielfeldt, T.; Meldal, M.; Bock, K.
Carbohydr. Res. 1995, 268, 17.
34. Meinjohanns, E.; Meldal, M.; Schleyer, A.; Paulsen, H.;
Bock, K. J. Chem. Soc., Perkin Trans. 1 1996, 985.
35. Satyanarayana, J.; Gururaja, T. L.; Naganagowda, G. A.;
Ramasubbu, N.; Levine, M. J. J. Peptide Res. 1998, 52, 165.
36. Ferrier, R. J.; Furneaux, H. Methods Carbohydr. Chem.
1980, 8, 251.
37. Otvos, L., Jr.; Pease, A. M.; Bokonyi, K.; Giles-Davis,
W.; Rogers, M. E.; Hintz, P. A.; Hoffmann, R.; Ertl, H. C. J.
J. Immunol. Meth. 2000, 233, 95.
1. Pamer, E. G. Clin. Infect. Dis. 1999, 28, 714.
2. Maffei, A.; Harris, P. E. Peptides 1998, 19, 179.
3. Chicz, R. M.; Urban, R. G. Immunol. Today 1994, 15, 155.
4. Sant, A. J. Immunol. Res. 1991, 13, 253.
5. Bacik, I.; Link-Snyder, H.; Anton, L. A.; Russ, G.; Chen,
W.; Bennink, J. R.; Urge, L.; Otvos, L., Jr.; Dudlowska, B.;
Eisenlohr, L.; Yewdell, J. W. J. Exp. Med. 1997, 186, 479.
6. Ishioka, G. Y.; Lamont, A. G.; Thompson, D.; Bulbow,
N.; Gaeta, F. C. A.; Sette, A.; Grey, H. M. J. Immunol. 1992,
148, 2446.
7. Mouritsen, S.; Meldal, M.; Christiansen-Brams, I.; Elsner,
H.; Werdelin, O. Eur. J. Immunol. 1994, 24, 1066.
8. Haurum, J. S.; Arsequell, G.; Lellouch, A. C.; Wong,
S. Y. C.; Dwek, R. A.; McMichael, A. J.; Elliott, T. J. Exp.
Med. 1994, 180, 739.
38. Corthay, A.; Backlund, J.; Broddefalk, J.; Michaelsson,
E.; Goldschmidt, T. J.; Kihlberg, J.; Holmdahl, R. Eur. J.
Immunol. 1998, 28, 2580.
9. Harding, C. V.; Kihlberg, M.; Elofsson, M.; Magnusson,
G.; Unanue, E. R. J. Immunol. 1993, 151, 2419.
10. Deck, B.; Eloffson, M.; Kihlberg, J.; Unanue, E. R.
J. Immunol. 1995, 155, 1074.
11. Abdel-Motal, U. M.; Berg, L.; Rosen, A.; Bengtsson, M.;
Thorpe, C. J.; Kihlberg, J.; Dahmen, J.; Magnusson, G.;
Karlsson, K. A.; Jondal, M. Eur. J. Immunol. 1996, 26, 544.
12. Carbone, F. R.; Gleeson, P. A. Glycobiology 1997, 7, 725.
13. Kihlberg, J.; Elofsson, M. Curr. Med. Chem. 1997, 4, 85.
14. Haurum, J. S.; Hoier, I. B.; Arsequell, G.; Neisig, A.;
Valencia, G.; Zeuthen, J.; Neefjes, J.; Elliott, T. J. Exp. Med.
1999, 190, 145.
39. Galli-Stampino, L.; Meinjohans, E.; Frische, K.; Meldal,
M.; Jensen, T.; Werdelin, O.; Mouritsen, S. Cancer Res. 1997,
57, 3214.
40. Komba, S.; Werdelin, O.; Jensen, T.; Meldal, M. J. Pept.
Sci. 2000, 6, 585.
41. Hayes, B. D.; Hart, G. W. Curr. Opin. Biol. 1994, 4, 692.
42. Woody, R. W. In The Peptides: Analysis, Synthesis and
Biology, vol. 7, Hruby, V. J., Ed.; Academic Press: San Diego,
1985; pp 15–114.
43. Greenfield, N.; Fasman, G. D. Biochemistry 1969, 8, 4108.
44. Perczel, A.; Hollosi, M.; Sandor, P.; Fasman, G. D. Int. J.
Peptide Protein Res. 1993, 41, 223.
15. Broddefalk, J.; Backlund, J.; Almquist, F.; Johansson, M.;
Holmdahl, R.; Kihlberg, J. J. Am. Chem. Soc. 1998, 120, 7676.
16. de Lalla, C.; Sturniolo, T.; Abbruzzese, L.; Hammer, J.;
Sidoli, A.; Sinigaglia, F.; Panina-Bordignon, P. J. Immunol.
1999, 163, 1725.
45. Schuster, O.; Klich, G.; Sinnwell, V.; Kranz, H.; Paulsen,
H.; Meyer, B. J. Biomol. NMR 1999, 14, 33.
46. Huang, X.; Barchi, J. J.; Lung, F.-D. T.; Roller, P. P.;
Nara, P. L.; Muschik, J.; Garrity, R. R. Biochemistry 1997, 36,
10846.
17. Udaka, K.; Wiesmuller, K. H.; Kienle, S.; Jung, G.;
Tamamura, H.; Yamagishi, H.; Okumura, K.; Walden, P.;
Suto, T.; Kawasaki, T. Immunogenetics 2000, 51, 816.
18. Lu, J.; Celis, E. Cancer Res. 2000, 60, 5223.
19. Brusic, V.; Rudy, G.; Honeyman, G.; Hammer, J.; Harri-
son, L. Bioinformatics 1998, 14, 121.
47. Spencer, D. I.; Missailidis, S.; Denton, G.; Murray, A.;
Brady, K.; Matteis, C. I.; Searle, M. S.; Tendlerm, S.; Price,
M. R. Biospectroscopy 1999, 5, 79.
48. Kirnarsky, L.; Prakash, O.; Vogen, S. M.; Nomoto, M.;
Hollingsworth, M. A.; Sherman, S. Biochemistry 2000, 39,
12076.
20. Otvos, L., Jr.; Urge, L.; Xiang, Z. Q.; Krivulka, G. R.;
Nagy, L.; Szendrei, G. I.; Ertl, H. C. J. Biochim. Biophys. Acta
1994, 1224, 68.
49. Siligardi, G.; Drake, A. F. Biopolymers 1995, 37, 281.
50. Agrawal, P. K.; Jacquinet, J.-C.; Krishna, N. R. Glyco-
biology 1999, 9, 669.
21. Otvos, L., Jr.; Krivulka, G. R.; Urge, L.; Szendrei, G. I.;
Nagy, L.; Xiang, Z. Q.; Ertl, H. C. J. Biochim. Biophys. Acta
1995, 1267, 55.
22. Otvos, L., Jr.; Cappelletto, B.; Varga, I.; Wade, J. D.;
Xiang, Z. Q.; Kaiser, K.; Stephens, L. J.; Ertl, H. C. J. Bio-
chim. Biophys. Acta 1996, 1313, 11.
51. Fields, G. B.; Noble, R. L. Int. J. Pept. Protein Res. 1990,
35, 161.
52. Angell, Y. M.; Garcia-Echeverria, C.; Rich, D. H. Tetra-
hedron Lett. 1994, 35, 5891.
53. Ertl, H. C. J.; Dietzschold, B.; Otvos, L., Jr. Eur. J.
Immunol. 1991, 21, 1.