Synthesis and application of sialic acid-containing building blocks for glycopeptide libraries. Establishing glycosylation conditions

Article abstract of DOI:10.1039/a908321i

Three different sialic acid-containing building blocks (6-8) were synthesized for use in solid-phase glycopeptide libraries. Investigation of the conditions for glycosylation of threonine (Thr) with various sialic acid donors revealed that the best results were obtained by coupling glycosyl xanthate 2 to the acceptors Fmoc-Thr-OH (5) or the α-azido acid analogue of Thr, 4. Among several catalysts employed, phenylsulfanyl triflate (PST) afforded the best yields. Both the N-Fmoc and α-azido analogues of Thr allowed glycosylation with good stereoselectivity in 80% (→ 8) and 84% (→ 6) yield, respectively. Introduction of a phenylthio group in the 3 position of the sialic acid donor 3, to assist the stereoselective outcome of the glycosylation reaction, gave good results; however difficulties in the removal of the phenylthio auxiliary group made this route less attractive. Both building blocks 6 and 8 were successfully introduced in solid-phase glycopeptide synthesis. Interestingly, alkaline deprotection of the Fmoc group of 8, necessary for subsequent introduction of amino acids, resulted in an immediate attack of the α-amino group on the sialic acid methyl ester to form the lactam 14. This side reaction was also observed during reduction of the azido acid building block 6 under alkaline conditions, but could be suppressed by performing the reduction under acidic conditions. Lactam formation was completely avoided by hydrolysis of the methyl ester prior to reduction of the azide. The Royal Society of Chemistry 2000.

Full text of DOI:10.1039/a908321i

View Article Online / Journal Homepage / Table of Contents for this issue
Synthesis and application of sialic acid-containing building blocks
1
for glycopeptide libraries. Establishing glycosylation conditions†
Koen M. Halkes, Phaedria M. St. Hilaire, Anita M. Jansson, Charlotte H. Gotfredsen and
Morten Meldal*
1
Department of Chemistry, Carlsberg Laboratory, Gamle Carlsberg Vej 10, DK-2500 Valby,
Copenhagen, Denmark
Received (in Cambridge, UK) 18th October 1999, Accepted 2nd May 2000
Published on the Web 9th June 2000
Three different sialic acid-containing building blocks (6–8) were synthesized for use in solid-phase glycopeptide
libraries. Investigation of the conditions for glycosylation of threonine (Thr) with various sialic acid donors revealed
that the best results were obtained by coupling glycosyl xanthate 2 to the acceptors Fmoc-Thr-OH (5) or the α-azido
acid analogue of Thr, 4. Among several catalysts employed, phenylsulfanyl triflate (PST) afforded the best yields.
Both the N-Fmoc and α-azido analogues of Thr allowed glycosylation with good stereoselectivity in 80% (→ 8)
and 84% (→ 6) yield, respectively. Introduction of a phenylthio group in the 3 position of the sialic acid donor 3, to
assist the stereoselective outcome of the glycosylation reaction, gave good results; however difficulties in the removal
of the phenylthio auxiliary group made this route less attractive.
Both building blocks 6 and 8 were successfully introduced in solid-phase glycopeptide synthesis. Interestingly,
alkaline deprotection of the Fmoc group of 8, necessary for subsequent introduction of amino acids, resulted in an
immediate attack of the α-amino group on the sialic acid methyl ester to form the lactam 14. This side reaction was
also observed during reduction of the azido acid building block 6 under alkaline conditions, but could be suppressed
by performing the reduction under acidic conditions. Lactam formation was completely avoided by hydrolysis of the
methyl ester prior to reduction of the azide.
Introduction
the ease of glycopeptide assembly has allowed large and diverse
10
libraries to be synthesized by the split-and-combine method.
In order to synthesize sialic acid-containing glycopeptides for
binding to sialic acid-recognizing receptors like the sialoadhesin
lectin, amino acids glycosylated with sialic acid were required.
Glycosylation of the hydroxy groups of amino acids is con-
sidered to be more difficult than glycosylation of carbohydrates,
presumably due to their poor reactivity arising from the
Sialic acid located at the periphery of glycolipids or glyco-
proteins is involved in a variety of biological phenomena, such
as cell differentiation, inflammation, or tumor progression and
1
metastasis. The role of sialic acid in these interactions can, in
general terms, be considered two-fold. First, it makes a major
contribution to the net negative charge, which through electro-
static repulsion can be important in modulating cell–cell inter-
actions mediated by specific cell-adhesion molecules. Second,
sialic acid can function as a specific recognition determinant at
α
unfavorable H-bonding with the commonly used urethane N -
2
11
protecting groups of amino acids. In addition, the formation
of the glycosidic bond of sialic acid is more difficult compared
3
the cell surface, e.g. leucocyte/E-selectin interaction or the
macrophage sialic acid receptor.
12
with that of other carbohydrates due to three inherent factors.
4
First, the presence of the carboxylic acid function at the ano-
meric center (C-2) electronically disfavors oxonium ion form-
ation. Secondly, the carboxylate group interferes sterically with
the glycoside formation, and finally the lack of a functional
group at C-3 excludes the use of neighboring-group partici-
pation for stereochemical control. Some of the above men-
tioned problems could be circumvented by attaching sialic acid
to an amino acid via a linker molecule. While the use of a linker
molecule to couple sialic acid to amino acids will simplify the
synthesis, it will also reduce the proximity of carbohydrate and
peptide and introduce additional degrees of rotational freedom
that may result in a greater entropic penalty and reduced
binding. Therefore, direct O-sialylation of an amino acid was
preferred in order to retain conformational rigidity in the
The mechanisms of the majority of the above mentioned
interactions are not yet well elucidated due to the difficulty in
obtaining the glycoconjugates in a pure form and sufficient
quantities. The synthesis of the individual, complex oligo-
saccharide structures required to study the ligand–receptor
5
interactions is cumbersome and time consuming. Often, only a
few residues at the non-reducing end of a complex glycan are
6
necessary for tight interaction with the receptor. Therefore, the
use of simplified synthetic molecules that can be rapidly gener-
ated and can mimic the natural ligand can give important
information about the nature and topology of the ligand–
7
receptor interaction. In fairly recent work it has been shown
that complex oligosaccharide structures can be mimicked by
8,9
glycopeptides. The peptide-scaffold may even actively partici-
pate in the ligand–receptor interaction, thus enhancing the
overall affinity of the ligand to the receptor, and furthermore,
glycopeptide through carbohydrate–peptide interactions. While
13–16
many examples of glycosylated amino acids
have been
described in the literature, to our knowledge no sialylated
amino acids O-linked via the glycosidic bond have been previ-
ously reported. In the present paper we describe the synthesis of
three sialylated threonine building blocks, and their application
in solid-phase glycopeptide synthesis.
†
Mass spectra of peptides released as a single bead are available as
supplementary data. For direct electronic access see http://www.rsc.org/
suppdata/p1/a9/a908321i
DOI: 10.1039/a908321i
J. Chem. Soc., Perkin Trans. 1, 2000, 2127–2133
This journal is © The Royal Society of Chemistry 2000
2127

View Article Online
Scheme 1 Reagents: i, DMTST, CH Cl ; ii, PST, CH CN.
2
2
3
Results and discussion
Building block synthesis
Three versatile sialic acid donors, methyl (phenyl 5-acetamido-
4
,7,8,9-tetra-O-acetyl-3,5-dideoxy-2-thio--glycero--galacto-
17
non-2-ulopyranosid)onate 1, O-ethyl S-[methyl(5-acetamido-
4
,7,8,9-tetra-O-acetyl-3,5-dideoxy-2-thio--glycero--galacto-
17
non-2-ulopyranosid)onate] dithiocarbonate (2), and methyl
phenyl 5-acetamido-4,7,8,9-tetra-O-acetyl-5-deoxy-3-S-
phenyl-2,3-dithio--erythro--gluco-non-2-ulopyranosid)-
24
(
has been shown that sialylation reactions progress in higher
yields when less sterically hindered and more nucleophilic
acceptors are used. The use of an acceptor molecule containing
12,18,19
onate
3 were prepared (Scheme 1). Synthesis of the
2
5–27
donors 1 and 2 proceeded as described in the literature with
an α-azido group as the masked α-amine,
and no carboxylic
12
28
comparable yields. Synthesis of the sialic acid donor 3, which
has a C-3 auxiliary phenylthio group for stereocontrol in the
glycosylation reaction, proved to be cumbersome. Yields of
acid protecting group, satisfies both the steric and electronic
requirements. Consequently, the acceptor, 2-(S)-azido-3-(R)-
hydroxybutanoic acid (N -Thr-OH) was synthesized in two
α
3
5
7–77% are reported for the crucial reaction step, in which
steps starting from commercially available H-Thr(tBu)-OH.
20
26,29
benzenesulfenyl chloride is added to the double bond of sialic
acid glycal. However, in our hands, a yield of only 30% was
obtained, partly due to difficulties in the purification of the
complex reaction mixture.
The amine was first converted to the azide
by treatment
30
with freshly prepared triflic azide, followed by subsequent
removal of the tBu group, with 95% trifluoroacetic acid (TFA).
The desired acceptor (4, Scheme 1) was obtained in 93% yield
for the two steps.
Initially Fmoc-Thr-OPfp 9 was selected as the acceptor
molecule, since glycosylation with other carbohydrates has pre-
Glycosylation of 4 with phenylthio donor 1, using N-iodo-
succinimide (NIS)–triflic acid (TfOH) as promoter, was un-
successful and yielded mainly the elimination product of the
donor. Glycosylation using the same donor–acceptor pair
under dimethyl(methylthio)sulfonium triflate (DMTST) acti-
vation yielded the desired compound 6 (Scheme 1) in
59% yield and an α:β ratio of 1:1. Attempts to increase the
α-selectivity during the glycosylation reaction by decreasing the
reaction temperature were unsuccessful and led to formation
of an increasing amount of glycal and a reduced glycosyl-
ation yield.
21,22
viously been successful,
and, furthermore, the Fmoc/OPfp
building block strategy for the synthesis of O- and N-linked
2
3
glycopeptides has been proven to be very versatile. Unfortu-
nately, attempted glycosylation of acceptor 9 with the sialic
acid donors 1 or 2 was unsuccessful, and resulted primarily in
hydrolysis or elimination of the activating group. Typically
yields of approximately 10% were obtained, with no stereo-
selectivity.
31
It was initially thought that the low yields were due to steric
α
hindrance caused by the bulky Fmoc N -protecting group. It
2
128
J. Chem. Soc., Perkin Trans. 1, 2000, 2127–2133

View Article Online
a
Table 1 Glycosylation conditions investigated
Donor
Acceptor
Product
Reaction conditions
Yield (%) (α:β ratio)
b
1
1
1
1
2
2
2
3
2
or 2
9
9
4
4
4
4
4
4
5
DMTST, 0 ЊC
≈10b
NIS–TfOH, Ϫ30 ЊC
NIS–TfOH, Ϫ30 ЊC
DMTST, 0 ЊC
MeSBr, Ϫ30 → 0 ЊC
PST, Ϫ30 ЊC
PST, Ϫ60 ЊC, CH CN–CH Cl (4:1)
PST, Ϫ30 ЊC
PST, Ϫ40 ЊC
≈10
6
6
6
6
7
8
59 (1:1)
19 (1:1)
83 (3:1)
84 (5:1)
93
3
2
2
80 (9:1)
a
b
All reactions were performed in CH CN, unless otherwise indicated, in the presence of powdered 3 Å molecular sieves. Estimated from TLC.
3
Glycosylation of 4 with the xanthate donor 2, using
(MALDI-TOF) mass spectrometry. Peptide 11 (Scheme 2) was
33
methanesulfenyl bromide (MeSBr) as a promoter, afforded the
desired compound 6 in 19% yield, with no stereoselectivity.
synthesized using the syringe technology and Fmoc/OPfp-
derivatized amino acids which were activated with 3,4-dihydro-
3-hydroxy-4-oxo-1,2,3-benzotriazine (Dhbt-OH).
19
However, when phenylsulfenyl triflate (PST) was used as the
catalyst, 6 was formed in 83% yield and an α:β ratio of 3:1
Coupling of 8 with model peptide 11, using O-benzotriazol-1-
yl-N,N,NЈ,NЈ-tetramethyluronium tetrafluoroborate (TBTU)/
N-ethylmorpholine (NEM) activation, gave the desired glyco-
peptide 12 (Scheme 2) as shown by MALDI-TOF mass spectro-
metry. Coupling of 6 with model peptide 11, using TBTU/
NEM activation, afforded glycopeptide 13 (Scheme 2) as shown
by MALDI-TOF mass spectrometry and IR spectroscopy.
Coupling of amino acids subsequent to the incorporation of
sialylated building blocks proved challenging. Removal of the
Fmoc protecting group of 12 under alkaline conditions [20%
piperidine in dimethylformamide (DMF)] resulted in formation
of the sialic acid-Thr lactam and deacetylation gave 14 (Scheme
2). Compound 14 was isolated, upon photolytic cleavage,
by HPLC in 52% overall yield starting from the introduction of
10 to the solid phase. Compound 14 was characterized by
NMR spectroscopy (see Table 2) and ES mass spectroscopy.
Additionally, reduction of the azide of 13 under alkaline
conditions {dithiothreitol (DTT) and 1,8-diazabicyclo[5.4.0]-
undec-7-ene (DBU) in DMF} and removal of O-acetyl groups
1
(
determined by comparison of the 3 -H signals in H NMR).
eq
Changing the solvent system from acetonitrile to an aceto-
nitrile–dichloromethane mixture allowed the lowering of the
reaction temperature to Ϫ60 ЊC, leading to formation of the
product in the same yield (84%), but with an increased stereo-
selectivity (α:β ratio of 5:1). From this mixture the α-anomer
could be isolated in 56% yield.
Since the separation of the α- and β-anomers of 6 by vacuum
liquid chromatography (VLC) was difficult, the glycosylation of
4
with donor 3 having a C-3 auxiliary phenylthio group that
directs the formation of an α-linkage through neighboring-
group participation was investigated. This glycosylation
afforded exclusively the α-anomer of 7 (Scheme 1), in an excel-
lent yield of 93%. Unfortunately, removal of the C-3 auxiliary
group using either tributyltin hydride or triphenyltin hydride
and 2,2Ј-azoisobutyronitrile (AIBN) in refluxing toluene, or
Raney Nickel gave a complex reaction mixture from which no
product could be isolated.
34
To determine whether the improvement in glycosylation yield
and stereoselectivity between acceptor 9 and 4 could be solely
attributed to steric hindrance or to the electron-donating prop-
erties of the azide compared with Fmoc, glycosylation of com-
mercially available Fmoc-Thr-OH (5, Scheme 1) was attempted.
The coupling of donor 2 to acceptor 5, at Ϫ40 ЊC in
acetonitrile using PST as catalyst, surprisingly gave 80% of a
gave also the sialic acid-Thr lactam 14. However, protonation
of the amino function by performing the reduction under
acidic conditions (TMSCl and NaI in acetonitrile) decreased
the formation of lactam, as could be shown by subsequently
incorporation of Fmoc-Gly-OPfp (to give 15, Scheme 2).
Despite the successful coupling of amino acids, the con-
comitant loss of sialic acid from the peptide [masses at 1005.21
ϩ
ϩ
9
:1 mixture of the anomers of 8 (Scheme 1), from which the
([M ϩ Na] ) and 1021.21 ([M ϩ K] ), mass spectra are given
in supplementary material] under the acidic reduction con-
ditions made this route less attractive. An alternative route
would be to remove the protecting groups from sialic acid,
prior to reduction. In this way, no lactam will be formed
during alkaline azide reduction and no loss of sialic acid, due to
acidic azide reduction conditions, will occur. Removal of the
protecting groups, especially the methyl ester, proved to be
difficult on the solid phase. Normal deprotection strategies for
α-anomer could be isolated in a 62% yield. The low yield
obtained with 9 as an acceptor is therefore most likely due to
the combined steric effect of the pentafluorophenyl group and
22
the Fmoc group as previously suggested. Results of the glyco-
sylation reaction are compiled in Table 1.
Glycopeptide synthesis
As a preface to the synthesis of sialic acid-containing glyco-
peptide libraries, the conditions of successful incorporation
of building blocks 6 and 8 on a solid phase were investigated.
35
solution-phase deprotection (LiOH or NaOH in water–
36
alcohol mixtures or LiI in pyridine ) were not successful, as
indicated by the formation of the cyclized side product during
32
The photolabile amide linker 10 was used to facilitate
37,38
azide reduction. However, addition of calcium chloride
to a
1
M solution of LiOH in 70% propan-2-ol in water in conjunc-
tion with sonication of the mixture afforded smoothly the
desired deprotected sialic acid derivative. It is assumed that
2ϩ
the Ca stabilizes electrostatically the transition state during
the hydrolysis process. Azide reduction on the deprotected
glycopeptide (DTT and DBU in DMF) followed by in-
corporation of three more amino acids afforded the desired
compound 16 (Scheme 2). Compound 16 was isolated, upon
photolytic cleavage, by HPLC in 38% overall yield starting from
incorporation of 10 on the solid phase. Compound 16 was
characterized by NMR spectroscopy (see Table 2) and ES mass
spectrometry.
mild cleavage of the glycopeptide from the solid support
[acryloylated bis(2-aminopropyl)poly(ethylene glycol)/acryl-
amide copolymer (PEGA1900)] and to permit expedient analysis
via matrix-assisted laser desorption ionization–time-of-flight
J. Chem. Soc., Perkin Trans. 1, 2000, 2127–2133
2129

View Article Online
a
a
Table 2 Chemical-shift table for compounds 14 and 16
1
4
16
NH
Hα
Hβ
Hγ
Hδ
NH
Hα
Hβ
Hγ
Hδ
2,6-H, 3,5-H, 4-H
7.32, 7.24
Phe
Leu
Gly
Thrb
Ala
7.12, 6.88
4.11
4.42
3.84, 3.57
4.05
4.35
4.45
4.29
4.44
4.53
3.21, 3.02
1.69, 1.53
1.59
1.17
0.87, 0.90
8.40
7.45
8.23
8.32
8.32
3.82
4.52
4.57
4.31
4.45
4.53
4.27
4.11
4.27
1.22
1.21
c1.91
4.17
1.11
b
Prob
2.16, 1.83
1.98, 1.85
1.15
2.13, 1.86
2.02, 1.87
1.20
3.64, 3.48
3.64, 3.52
1.92, 1.85
1.97, 1.79
1.14
2.12, 1.83
2.02, 1.84
1.19
1.69
1.85
3.52, 3.37
3.63, 3.46
Pro
b
Alab
7.82
7.81
Prob
Pro
c1.93
3.63, 3.46
3.64, 3.57
1.95
1.89
3.61, 3.46
3.64, 3.55
4.27
4.11
b
Ala
7.69
7.69
NH
H3ax
H3eq
H4
H5
H6
Ac
1
1
4-OSA
6-OSA
1.35
1.31
2.13
2.63
4.35
4.68
3.52
3.39
3.75
3.54
7.93
1.81
a
b
All spectra were acquired at 25 ЊC in DMSO-d and referenced to DMSO at δ 2.49. The assignments of these resonances are based on connectivi-
6
H
α
α
α
α
δ
δ
β
δ
ties from H (i)–H (i ϩ 1), H (i)–NH(i ϩ 1), H (i)–H (i)–H (i ϩ 1) and Ala H (i)–H (i ϩ 1). If no fully unambiguous sequential assignment could be
made among these the order can be interchanged. These resonances are most likely totally overlapped with the upfield resonance of Hβ.
c
Scheme 2 Reagents: i, TBTU, NEM, DMF; ii, 20% piperidine–DMF; iii, DTT, DBU, DMF; iv, TMSCl, NaI, CH CN; v, Fmoc-Aa-OPfp, Dhbt-
3
i
OH, DMF; vi, LiOH, CaCl , 70% Pr OH–H O; vii, 80% aqueous hydrazine in MeOH.
2
2
In summary, we have synthesized three sialic acid-containing
building blocks, 6, 7 and 8, in high yield and good stereo-
selectivity. Building block 7 proved unsuitable for further use
since the C-3 auxilary phenylthio group could not be removed
from the glycosyl amino acid. Building blocks 6 and 8 were
smoothly introduced during solid-phase glycopeptide synthesis.
The Fmoc building block can be used for synthesizing a library
that gives rise to lactamized products that cannot be further
extended, while the azide building block can be used for the
synthesis of a sialylated glycopeptide library, provided the
methyl ester is hydrolyzed prior to azide reduction.
Experimental
Materials and general methods
All solvents were of p.a. quality, and were distilled from
appropriate drying agents when necessary. DMF was stored
over 3 Å molecular sieves. Reaction mixtures were dried with
Na SO and concentrated under reduced pressure at temper-
2
4
atures less than 30 ЊC (water-bath). VLC was performed on
Merck Silica Gel 60 H, and chromatography under dry con-
ditions was performed on dried silica gel (120 ЊC; 24 h), eluting
with dry solvents. Solid-phase peptide-coupling reactions were
2
130
J. Chem. Soc., Perkin Trans. 1, 2000, 2127–2133

View Article Online
3
9
ꢀ
monitored using the Kaiser test, and solution-phase reactions
were monitored by thin-layer chromatography (TLC) per-
formed on Merck Silica Gel 60 F254 aluminium-backed sheets
with detection by charring with sulfuric acid, or by UV light
when applicable. PEGA1900 resin (300–500 µm) was obtained
from Polymer Laboratories (Amherst, MA). Suitably protected
N -Thr[methyl (5-acetamido-4,7,8,9-tetra-O-acetyl-3,5-
3
dideoxy-D-glycero-ꢀ-D-galacto-non-2-ulopyranosid)onate]-OH 6
via Donor 1. Donor 1 (0.31 g, 0.54 mmol), acceptor 4 (26 mg,
0
.18 mmol), molecular sieves 3 Å, and a magnet were placed in
3
a predried 25 cm flask. The air in the flask was evacuated and
3
replaced by an atmosphere of argon. Dry acetonitrile (4 cm )
was added. The suspension was cooled to 0 ЊC, DMTST was
added (0.137 g, 0.54 mmol) and the suspension was stirred for
2 h at RT. The reaction mixture was filtered over Celite, then
directly applied to a VLC column and eluted with chloroform–
methanol (9:1) to afford 6 (65 mg, 59%) as a 1:1 mixture of
anomers.
α
N -Fmoc amino acids were purchased from NovaBiochem
Switzerland), TBTU and Dhbt-OH from Fluka (Switzerland),
NEM from Merck (Germany), DBU, silver triflate (AgOTf,
recrystallized from toluene) and DTT from Aldrich (USA).
Optical rotations were measured on a Perkin-Elmer 241 polari-
3
1
(
1
Ϫ1
2
Ϫ1
meter and [α] -values are given in units of 10 deg cm g .
D
Preparative HPLC of compounds 14 and 16 was performed
over a 25 × 200 mm semipreparative RP-18 column (Millipore
Delta Pak 15 µ). Eluents A (1% TFA in water) and B (10% of A
with 1% TFA in acetonitrile) were used in a linear gradient,
via Donor 2. Donor 2 (2.23 g, 3.75 mmol), acceptor 4 (0.328
g, 1.87 mmol), molecular sieves 3 Å, and magnet were placed in
a predried 50 cm flask. The air in the flask was evacuated and
replaced by an atmosphere of argon. Acetonitrile (12 cm ) and
dichloromethane (3 cm ) were added and the suspension was
cooled to Ϫ60 ЊC. AgOTf (1.10 g, 4.31 mmol) and benzene-
sulfenyl chloride (0.54 cm , 3.75 mmol) were added and the
reaction mixture stirred for 3 h. The suspension was neutralized
with N,N-diisopropylethylamine (DIPEA) (0.75 cm , 4.31
mmol), filtered over Celite, and concentrated to dryness. The
product was applied to a VLC column and eluted with
chloroform–methanol (9:1) to yield 6 (845 mg, 84%) as a 5:1
mixture of anomers. Separation of the anomers on a VLC
column eluted with toluene–acetone (5:1) afforded the pure
α-anomer of 6 (640 mg, 56%), [α] Ϫ93 (c 1.0, CHCl ) [Found:
3
Ϫ1
starting with 85% A and 15% B, a slope of 0.5% min , and a
3
3
Ϫ1
flow rate of 10 cm min . ES-MS spectra were recorded in
the positive mode on a Fisons VG Quattro Instrument. NMR
spectra were recorded on a Bruker AMX-250 or a Bruker
DRX-600 MHz spectrometer. The H and C resonances were
assigned by H, C, H– H COSY, and HSQC experiments.
NMR spectra were recorded in CDCl , D O or CDCl –CD OD
3
3
1
13
1
13
1
1
3
3
2
3
3
mixtures. Chemical shifts are given in ppm and referenced to
CDCl (δ 7.29 and δ 77.0); J values are given in Hz.
3
H
C
MALDI-TOF mass spectrometry
Beads were irradiated on stainless steel targets with a strong UV
lamp for 30 min. The analyte was extracted on the target from
the beads using 0.5 mm of 70% acetonitrile and then dried at
D
3
(ES-MS positive-ion mode) m/z 619.2. C H N O requires M,
618.5]; δ (600 MHz; CDCl ) 5.44–5.41 (1 H, m, 8-H), 5.38 (1 H,
24
34
4
15
3
H
3
room temperature (RT). The appropriate matrix was added, the
br d, JNH,H5 9.8, NHCOCH ), 5.34 (1 H, br d, J 8.5, J <1,
3 7,8 6,7
β
7-H), 4.97–4.94 (1 H, m, 4-H), 4.59–4.57 (1 H, m, CH ), 4.39
(1 H, dd, J8,9b 2.6, 9a-H), 4.12 (1 H, dd, J8,9a 6.0, J9a,9b 12.4, 9b-
sample dried at 40 ЊC, and the spectrum recorded on a Bruker
Tm
Reflex III MALDI-TOF mass spectrometer. Spectra were
obtained (1–100 pulses) using the lowest power required to
facilitate desorption and ionization. Ions were accelerated
toward the discrete dynode multiplier detector with an acceler-
ation voltage of 20 kV. The matrix α-cyano-4-hydroxycinnamic
H), 4.05 (1 H, dd, J4,5 10.1, 5-H), 3.98 (1 H, br d, J5,6 10.6, 6-H),
α
3.83 (3 H, s, CO CH ), 3.63 (1 H, d, J
3.4, CH ), 2.69
2
3
CHα,CHβ
(1 H, dd, J3eq,4 4.5, J3eq,3ax 12.4, 3eq-H), 2.19, 2.09 and 2.07
(3 H, 3 H, 6 H, 3 s, 4 × O CCH ), 1.95 (3 H, s, NHCOCH ),
2
3
3
γ
3
acid (CHC, 10 mg in 1 cm of 70% acetonitrile) was used to
1.896 (1 H, t, J3ax,4 12.4, 3ax-H), 1.49 (3 H, d, JCHγ,CHβ 6.2, CH );
δ (75 MHz; CDCl ) 171.9, 171.7 (2), 171.2 (2) and 170.8,
analyze both peptides and protected glycopeptides. Unpro-
tected glycopeptides were analyzed using the matrix 2,6-
dihydroxyacetophenone (DHAP) to which was added pyridine
C
3
(4 × O CCH , NHCOCH , and CO CH ), 167.9 (CO H), 100.2
2
3
3
2
3
2
β
(C-2), 73.4 (C ), 73.0 (C-6), 69.5 (C-4), 69.4 (C-8), 67.3
(C-7), 66.7 (C ), 62.9 (C-9), 53.5 (CO CH ), 50.1 (C-5), 38.6
3
3
α
(
DHAP, 10 mg in 1 cm of 70% acetonitrile and 50 mm of
2
3
pyridine). Bradykin (1060.2 mu), renin (1759.0 mu), and
mellitin (2846.5 mu) were used as the standards for internal
calibration of the mass spectra.
(C-3), 23.54 (NHCOCH ), 21.7 and 21.3 (3) (4 × O CCH ),
21.0 (C ).
3
2
3
γ
ꢀ
N -Thr[methyl (5-acetamido-4,7,8,9-tetra-O-acetyl-5-deoxy-3-
3
ꢀ
3
N -Thr-OH 4
S-phenyl-3-thio-D-erythro-ꢁ-L-gluco-non-2-ulopyranosid)onate]-
OH 7
H-Thr(tBu)-OH (1.00 g, 5.71 mmol) and CuSO ؒ5H O were
4
2
3
3
dissolved in a mixture of water (18 cm ) and methanol (36 cm ),
then K CO was added until pH 10 (pH paper) was obtained.
Donor 3 (145 mg, 209 µmol), acceptor 4 (33 mg, 189 µmol),
molecular sieves 3 Å, and magnet were placed in a predried 10
2
3
3
Under vigorous stirring, a solution of triflic azide (12 mmol) in
cm flask. The air in the flask was evacuated and replaced by an
3
3
dichloromethane (30 cm ) was added, and the pH of the solu-
atmosphere of argon. Acetonitrile (5 cm ) was added and the
tion was re-adjusted to 10 by addition of K CO . The reaction
suspension cooled to Ϫ40 ЊC. AgOTf (124 mg, 0.48 mmol)
and benzenesulfenyl chloride (30 mm , 209 µmol) were added
2
3
3
mixture was stirred overnight at RT, then diluted with dichlo-
3
romethane (50 cm ). The layers were separated and the organic
and the reaction mixture was stirred for 1 h. The suspension
3
phase was twice extracted with water. The combined aqueous
was neutralized with DIPEA (82 mm , 0.48 mmol), filtered
layers were acidified using 2 M H SO and extracted with
over Celite, and concentrated to dryness. The product was
applied to a VLC column and eluted with chloroform–
methanol (30:1 → 9:1) to yield 7 (148 mg, 93%), [α]D ϩ99
2
4
3
dichloromethane (3 × 50 cm ). The combined organic extracts
were dried (Na SO ), filtered and concentrated to dryness.
2
4
To the crude mixture was added 95% aq. TFA. The solution
was stirred for 30 min and then diluted with toluene and
concentrated to dryness. The product was applied to a VLC
column and eluted with chloroform–methanol (6:1) to yield
(c 1.0, CHCl ) [Found: (ES-MS positive-ion mode) m/z 727.2.
3
C H N O S requires M, 726.7]; δ (250 MHz; CDCl ) 7.45–
30
38
4
17
H
3
7.22 (5 H, m, SPh), 5.35 (1 H, t, J3,4=4,5 10.9, 4-H), 5.34 (1 H, d,
J 10.7, NHCOCH ), 5.26–5.21 (1 H, m, 8-H), 5.16 (1 H, dd, J
3
6,7
21
β
compound 4 (0.77 g, 93%), [α]D Ϫ50 (c 1.0, MeOH) [Found:
1.56, J7,8 9.25, 7-H), 4.47 (1 H, m, CH ), 4.18 (1 H, dd, J 2.0,
8,9b
(
ES-MS negative-ion mode) m/z 144.8. C H N O requires
J9a,9b 12.2, 9b-H), 4.16 (1 H, t, J5,6 9.92, 5-H), 4.07 (1 H, dd,
4
7
3
3
β
α
M, 145.1]; δ (250 MHz; CDCl ) 4.31 (1 H, dq, CH ), 3.87
6-H), 3.94 (1 H, d, JCHα,CHβ 3.4, CH ), 3.90 (1 H, dd, J
H
3
8,9a
α
γ
(
1 H, d, JCHα,CHβ 3.2, CH ), 1.42 (3 H, d, J
5.9, CH );
6.4, 9a-H), 3.82 (3 H, s, CO CH ), 3.14 (1 H, d, J
10.9, 3-H),
CHγ,CHβ
2
3
3eq,4
β
α
δ (75 MHz; CDCl ) 173.1 (COOH), 76.1 (C ), 69.0 (C ),
2
2.04, 2.03, 1.97 and 1.88 (each 3 H, 4 s, 4 × O CCH ), 1.75 (3 H,
2 3
γ
C
3
γ
0.2 (C ).
s, NHCOCH ), 1.20 (3 H, d, J
6.3, CH ).
3
CHγ,CHβ
J. Chem. Soc., Perkin Trans. 1, 2000, 2127–2133
2131

View Article Online
ꢀ
N -Fmoc-Thr[methyl (5-acetamido-4,7,8,9-tetra-O-acetyl-3,5-
Synthesis of 13
dideoxy-D-glycero-ꢀ-D-galacto-non-2-ulopyranosid)onate]-
OH 8
To resin bound peptide 11 (42 mg resin, 8.09 µmol NH func-
2
tions) was added a solution of 7 (15 mg, 24.3 µmol), TBTU (6.9
3
3
Donor 2 (1.10 g, 1.85 mmol), acceptor 5 (0.315 g, 0.92 mmol),
molecular sieves 3 Å, and magnet were placed in a predried 50
mg, 23.5 µmol), and NEM (6 mm ) in 400 mm of dry DMF.
After 3 h, the reagents were removed by suction and the resin
was washed with DMF (8×). Photolytic release of the com-
pound, followed by MALDI-TOF mass spectrometry, estab-
lished the identity of the product. Mass calc.: 1241.24 [M ϩ
3
cm flask. The air in the flask was evacuated and replaced by an
3
atmosphere of argon. Acetonitrile (8 cm ) was added and the
suspension cooled to Ϫ35 ЊC. AgOTf (0.95 g, 3.69 mmol) and
3
ϩ
ϩ
ϩ
benzenesulfenyl chloride (133 mm , 0.92 mmol) were added and
Na] , 1257.15 [M ϩ K] ; mass found: 1241.28 [M ϩ Na] ,
ϩ
the reaction mixture was stirred for 3 h. The suspension was
1257.24 [M ϩ K] .
3
neutralized with DIPEA (0.64 cm , 3.69 mmol), filtered over
Celite, and concentrated to dryness. The product was applied to
a VLC column and eluted with chloroform–methanol (9:1) to
yield 8 (599 mg, 80%) as a 9:1 mixture of anomers. Separation
of the anomers on a VLC column eluted with toluene–acetone
Synthesis of 14
via Compound 12. Compound 12 (10 mg resin) was treated
with 20% piperidine in DMF solution. After 20 min, suction
was applied and the resin was washed with DMF (8×). The
Kaiser test showed that no amine functions were present. The
Fmoc protecting group was removed, then the acetyl protecting
groups were removed by overnight treatment of the resin with
(
(
4:1) afforded the pure α-anomer of 8 (464 mg, 62%), [α]D ϩ8
c 1.0, CHCl ) [Found: (ES-MS positive-ion mode) m/z 814.6.
3
C H N O requires M, 814.8]; δ (600 MHz; CDCl ) 7.98–
39
46
2
17
H
3
7
.22 (8 H, m, Fmoc ArH), 5.51 (1 H, br d, JNH,H5 9.3,
3
3
8
0% aq. hydrazine (56 mm ) in methanol (1 cm ). The product
NHCOCH ), 5.37–5.31 (1 H, m, 8-H), 5.21 (1 H, br d, J 8.7,
3
7,8
was photolytically released of the resin, and purified by HPLC
to yield 14 (1.9 mg, 52% overall yield starting from the
introduction of 10 on resin). The identity of the product was
established by NMR spectroscopy (see Table 2) and ES-MS.
β
J6,7 <1, 7-H), 4.85–4.77 (1 H, m, 4-H), 4.49–4.47 (1 H, m, CH ),
.35–4.29 (3 H, m, Fmoc CH2 and 9a-H), 4.29 (1 H, d,
JCHα,CHβ 2.3, CH ), 4.17 [1 H, t, J(CH, Fmoc CH ) 7.0, Fmoc
4
α
2
CH], 4.01 (1 H, dd, J8,9a 5.9, J9a,9b 12.4, 9b-H), 3.93–3.89 (2 H,
ϩ
ϩ
Mass calc.: 953.24 [M ϩ H] , 975.25 [M ϩ Na] ; mass found:
m, 5- and 6-H), 3.71 (3 H, s, CO CH ), 2.54 (1 H, dd, J
4.4,
2
3
3eq,4
ϩ
ϩ
9
53.5 [M ϩ H] , 975.6 [M ϩ Na]
J3eq,3ax 12.4, 3eq-H), 2.08, 2.07, 1.97 and 1.96 (each 3 H, 4 s,
4
× O CCH ), 1.90 (1 H, t, J
12.4, 3ax-H), 1.84 (3 H, s,
2
3
3ax,4
via Compound 13. Compound 13 (15 mg resin) was treated
γ
NHCOCH ), 1.31 (3 H, d, J
CDCl ) 172.8, 171.8, 171.6, 171.3, 171.1 and 170.8 (4 ×
O CCH , NHCOCH , and CO CH ), 168.2 (CO H), 157.3
6.2, CH ); δ (75 MHz;
3
CHγ,CHβ
C
3
with 300 mm of a 0.1 M DTT solution in DMF to which was
added DBU (1 equiv. compared with 13). After 1 h, suction was
applied and the resin was washed with DMF (8×). The Kaiser
test showed that no amine functions were present. The acetyl
protecting groups were removed by overnight treatment of the
3
2
3
3
2
3
2
(
Fmoc CO), 144.4 (2), 141.9 (2), 128.3 (2), 127.7 (2), 125.7 (2)
and 120.6 (2) (12 × Fmoc ArC), 100.1 (C-2), 73.1 (C-6), 72.2
β
α
(
C ), 69.6 (C-4), 69.3 (C-8), 67.9 (C ), 67.8 (C-7), 63.1 (C-9),
3
resin with 80% aq. hydrazine (56 mm ) solution in methanol
5
3
9.4 (Fmoc CH ), 53.6 (CO CH ), 50.0 (C-5), 47.7 (Fmoc CH),
2 2 3
3
(
1 cm ).
8.4 (C-3), 23.7 (NHCOCH ), 21.7 and 21.4 (3) (4 ×
3
γ
O COCH ), 20.7 (C ).
2
3
Synthesis of 15
A solution of sodium iodide (3.6 mg, 24 µmol) in dry aceto-
nitrile (100 mm ) was added to resin-bound compound 13 (11.6
Synthesis of model peptide 11
3
PEGA1900 resin was washed with dichloromethane (6×) in a
syringe fitted with a Teflon filter, then dried under vacuum
mg resin, 2.4 µmol) and, after 5 min, chlorotrimethylsilane (3.6
3
mm , 38 µmol) was injected. After 30 min, suction was applied
(
(
lyophilizer) for at least 24 h before use. Fmoc-Lys(Fmoc)-OH
3 equiv.) was coupled to the resin, using TBTU (2.9 equiv.)–
and the resin was washed successively with 10% aq. Na S O
2
2
3
(
3×), water (2×), and acetonitrile (6×). To the resin was added a
NEM (5 equiv.) activation, in order to double the capacity of
the resin (final loading 0.24 mmol g ). The Fmoc protecting
groups were removed by treatment with 20% piperidine in
DMF and the photolabile linker 10 (1.5 equiv.) was attached
using TBTU (1.4 equiv.)–NEM (3 equiv.) activation.
solution of Fmoc-Gly-OPfp (11.1 mg, 24 µmol) and Dhbt-OH
Ϫ1
3
(
2.5 mg, 8 µmol) in 200 mm of dry DMF. After 6 h, the
reagents were removed by suction and the resin was washed
with DMF (8×). Photolytic release of the compound, followed
by MALDI-TOF mass spectrometry, established the identity of
All remaining peptide couplings were performed with the
Fmoc amino acid OPfp ester (3 equiv.), which were activated
with Dhbt-OH (1 equiv.). The progress of each coupling was
followed by the Kaiser test. The Fmoc group was removed
using 20% piperidine in DMF solution. After each coupling or
deprotection step the resin was washed with DMF (8×). Photo-
lytic release of the compound, followed by MALDI-TOF mass
spectrometry, established the identity of the product. Mass
ϩ
the products. Mass calc.: 1494.59 [M ϩ Na] , 1510.53 [M ϩ
ϩ
ϩ
ϩ
ϩ
K] ; mass found: 1494.30 [M ϩ Na] (35%), 1510.26 [M ϩ K]
ϩ
(
(
30%), 1199.32 [MLact ϩ K] (7.5%), 1183.27 [M ϩ Na]
Lact
ϩ
2.5%), 1021.21 [M Ϫ SA ϩ K] (12.5%), 1005.23 [M Ϫ SA ϩ
Na] (12.5%).
ϩ
Synthesis of 16
3
3
ϩ
ϩ
ϩ
A solution of 50 mm of 1 M aq. LiOH in 1.45 cm of 0.8 M
i
calc.: 619.74 [M ϩ H] , 641.73 [M ϩ Na] , 657.84 [M ϩ K] ;
ϩ
ϩ
CaCl in 70% Pr OH–H O was added to resin-bound compound
2 2
mass found: 619.31 [M ϩ H] , 641.29 [M ϩ Na] , 657.25
ϩ
13 (13 mg resin). The mixture was sonicated for 3 h at 0 ЊC, then
the reagents were removed by suction and the resin was washed
successively with water (3×), 95% acetic acid (2×), water (3×),
[
M ϩ K] .
Synthesis of 12
5
% DIPEA in DMF (3×), and DMF (6×).
The resin was treated with 300 mm of a 0.1 M DTT solution
3
To resin-bound peptide 11 (38 mg resin, 7.9 µmol) was added a
solution of 8 (19 mg, 23.8 µmol), TBTU (6.9 mg, 21.5 µmol),
in DMF to which was added DBU (1 equiv. with respect to 13).
After 1 h, suction was applied and the resin was washed with
DMF (8×). The Kaiser test established the presence of amine
functions. All remaining peptide couplings were performed
with the Fmoc amino acid OPfp esters (3 equiv.), which were
activated with Dhbt-OH (1 equiv.). The progress of each coup-
ling was followed by the Kaiser test. The Fmoc group was
3
3
and NEM (6 mm ) in 400 mm dry DMF. After 3 h, the
reagents were removed by suction and the resin was washed
with DMF (8×). Photolytic release of the compound, followed
by MALDI-TOF mass spectrometry, established the identity of
ϩ
the product. Mass calc.: 1437.54 [M ϩ Na] , 1453.53 [M ϩ
ϩ
ϩ
ϩ
K] ; mass found: 1437.39 [M ϩ Na] , 1453.39 [M ϩ K] .
2
132 J. Chem. Soc., Perkin Trans. 1, 2000, 2127–2133

View Article Online
removed using 20% piperidine in DMF. After each coupling or
deprotection step the resin was washed with DMF (8×).
11 R. Polt, L. Szabo, J. Treiberg, Y. Li and V. J. Hruby, J. Am. Chem.
Soc., 1992, 114, 10249.
1
1
2 T. Ercegovic and G. Magnusson, J. Org. Chem., 1995, 60, 3378.
3 M. Meldal, Glycopeptide Synthesis, in Neoglycoconjugates:
Preparation and Application, eds. Y. C. Lee and R. T. Lee, Academic
Press, San Diego, 1994, ch. 4, pp. 145–198.
The product was photolytically released of the resin, and
purified by HPLC to yield 16 (1.6 mg, 38% overall yield starting
from the introduction of 10 on resin). The identity of the prod-
uct was established by NMR spectroscopy (see Table 2) and ES-
14 H. Paulsen, S. Peters and T. Bielfeldt, Chemical Synthesis of
Glycopeptides, in Glycoproteins, Elsevier, Amsterdam, 1st edn.,
1995, pp. 87–121.
ϩ
ϩ
MS. Mass calc.: 1328.54 [M ϩ H] , 1350.60 [M ϩ Na] ; mass
ϩ
ϩ
found: 1328.7 [M ϩ H] , 1350.7 [M ϩ Na] , 1372.7 [M Ϫ H ϩ
ϩ
15 C. M. Taylor, Tetrahedron, 1998, 54, 11317.
2
Na]
1
6 G. Arsequell and G. Valencia, Tetrahedron: Asymmetry, 1997, 8,
2
839.
1
1
7 A. Marra and P. Sina, Carbohydr. Res., 1989, 187, 35.
Acknowledgements
8 Y. Ito and T. Ogawa, Tetrahedron, 1990, 46, 89.
The present work has been carried out at the SPOCC center
and supported by the Danish National Research Foundation.
Dr Les Miranda and Pia Breddam are acknowledged for meas-
uring ES-MS spectra. Bent Ole Petersen is acknowledged for
running the 600 MHz NMR spectra.
19 V. Martichonok and G. M. Whitesides, Carbohydr. Res., 1997, 302,
123.
2
2
0 S. Thea and G. Cevasco, Tetrahedron Lett., 1988, 29, 2865.
1 A. M. Jansson, M. Meldal and K. Bock, Tetrahedron Lett., 1990, 31,
6
991.
2
2
2 M. Meldal and K. J. Jensen, J. Chem. Soc., Chem. Commun., 1990,
83.
4
3 E. Meinjohanns, M. Meldal, H. Paulsen and K. Bock, J. Chem. Soc.,
Perkin Trans. 1, 1995, 405.
References
1
2
R. Schauer, Trends Glycosci. Glycotechnol., 1997, 9, 315.
R. Schauer, A. K. Shukla, C. Schröder and E. Müller, Pure Appl.
Chem., 1984, 56, 907.
M. Vinson, S. Mucklow, A. P. May, E. Y. Jones, S. Kelm and
P. R. Crocker, Trends Glucosci. Glycotechnol., 1997, 9, 283.
D. Nath, P. A. van der Merwe, S. Kelm, P. Bradfield and
P. R. Crocker, J. Biol. Chem., 1995, 270, 26184.
H. Ishida, Y. Ohta, Y. Tsukada, M. Kiso and A. Hasegawa,
Carbohydr. Res., 1993, 246, 75.
Y. C. Lee, Binding Modes of Mammalian Hepatic Gal/GalNAc
Receptors, in Carbohydrate Recognition in Cellular Function, eds.
G. Bock and S. Harnett, John Wiley & Sons, Chichester, England,
24 K. Okamoto and T. Goto, Tetrahedron, 1990, 46, 5835.
25 R. U. Lemieux and R. M. Ratcliffe, Can. J. Chem., 1979, 57, 1244.
26 J. Zaloom and D. C. Roberta, J. Org. Chem., 1981, 46, 5173.
27 P. Zimmermann and R. R. Schmidt, Liebigs Ann. Chem., 1988,
663.
28 M. Elofsson, S. Roy, L. A. Salvador and J. Kihlberg, Tetrahedron
Lett., 1996, 37, 7645.
29 P. B. Alper, S.-C. Hung and C.-H. Wong, Tetrahedron Lett., 1996,
37, 6029.
30 C. J. Cavender and V. J. Shiner, J. Org. Chem., 1972, 37, 3567.
31 P. Fugedi and P. J. Garegg, Carbohydr. Res., 1986, 149, C9.
32 C. P. Holmes and D. G. Jones, J. Org. Chem., 1995, 60, 2318.
33 M. K. Christensen, M. Meldal and K. Bock, J. Chem. Soc., Perkin
Trans. 1, 1993, 1453.
34 E. Meinjohanns, M. Meldal, T. Jensen, O. Werdelin, L. Galli-
Stampino, S. Mouritsen and K. Bock, J. Chem. Soc., Perkin Trans. 1,
1997, 871.
35 T. W. Greene and P. G. M. Wuts, in Protective Groups in Organic
Synthesis, John Wiley & Sons, New York, 2nd edn., 1991.
36 F. Elsinger, J. Schreiber and A. Eschenmoser, Helv. Chim. Acta,
1960, 43, 113.
37 R. Pascal and R. Sola, Tetrahedron Lett., 1997, 38, 4549.
38 R. Pascal and R. Sola, Tetrahedron Lett., 1998, 39, 5031.
39 E. Kaiser, R. L. Colescott, C. D. Bossinger and P. I. Cook, Anal.
Biochem., 1970, 34, 595.
3
4
5
6
1
st edn., 1989, pp. 80–95.
7
M. Meldal, I. Christiansen-Brams, M. K. Christensen, S. Mouritsen
and K. Bock, Synthesis and Biological Application of Glycosylated
Peptide Templates, in Complex Carbohydrates in Drug Research.
Structural and Functional Aspects, eds. K. Bock and H. Clausen,
Munksgaard, Copenhagen, 1st edn., 1994, pp. 153–165.
M. K. Christensen, M. Meldal, K. Bock, H. Cordes, S. Mouritsen
and H. Elsner, J. Chem. Soc., Perkin Trans. 1, 1994, 1299.
8
9
S.-H. Wu, M. Shimazaki, C.-C. Lin, L. Qiao, W. J. Moree, G. Weitz-
Schmidt and C.-H. Wong, Angew. Chem., Int. Ed. Engl., 1996, 35,
8
8.
1
0 P. M. St. Hilaire, T. L. Lowary, M. Meldal and K. Bock, J. Am.
Chem. Soc., 1998, 120, 13312.
J. Chem. Soc., Perkin Trans. 1, 2000, 2127–2133
2133