Synthesis and structural model of an α(2,6)-sialyl-T glycosylated MUC1 eicosapeptide under physiological conditions

Article abstract of DOI:10.1002/chem.200600144

To study the effect of O-glycosylation on the conformational propensities of a peptide backbone, a 20residue peptide (GSTAPPAHGVT-SAPDTRPAP) representing the full length tandem repeat sequence of the human mucin MUC1 and its analogue glycosylated with the

Full text of DOI:10.1002/chem.200600144

FULL PAPER
DOI: 10.1002/chem.200600144
Synthesis and Structural Model of an aACTHRE(UNG 2,6)-Sialyl-T Glycosylated MUC1
Eicosapeptide under Physiological Conditions
Sebastian Dziadek,[a, c] Christian Griesinger,*[b] Horst Kunz,*[a] and
Uwe M. Reinscheid[b, c]
Abstract: To study the effect of O-gly-
cosylation on the conformational pro-
pensities of a peptide backbone, a 20-
residue peptide (GSTAPPAHGVT-
SAPDTRPAP) representing the full
length tandem repeat sequence of the
human mucin MUC1 and its analogue
glycosylated with the (2,6)-sialyl-T anti-
gen on Thr11, were prepared and in-
vestigated by NMR and molecular
modeling. The peptides contain both
the GVTSAP sequence, which is an ef-
fective substrate for GalNAc transfer-
anti-MUC1 monoclonal antibodies. It
has been shown that glycosylation of
threonine in the GVTSAP sequence is
a prerequisite for subsequent glycosyla-
tion of the serine at GVTSAP. Further-
more, carbohydrates serve as additional
epitopes for MUC1 antibodies. Investi-
gation of the solution structure of the
sialyl-T glycoeicosapeptide in a H2O/
D2O mixture (9:1) under physiological
conditions (258C and pH 6.5) revealed
that the attachment of the saccharide
side-chain affects the conformational
equilibrium of the peptide backbone
near the glycosylated Thr11 residue.
For the GVTSAregion, an extended,
rod-like secondary structure was found
by restrained molecular dynamics sim-
ulation. The APDTR region formed a
turn structure which is more flexibly
organized. Taken together, the joined
sequence GVTSAPDTR represents the
largest structural model of MUC1 de-
rived glycopeptides analyzed so far.
Keywords: antigens · conformation
analysis
·
glycopeptides
·
ases, and the PDTRP fragment,
a
solid-phase synthesis
known epitope recognized by several
Introduction
out affecting healthy tissue. It is therefore necessary to iden-
tify suitable structural elements that clearly distinguish a
tumor cell from a normal cell. Such important cancer-selec-
tive structural information is observed in the tumor-associat-
ed mucin MUC1 which is a heavily O-glycosylated mem-
brane glycoprotein present at the interface between many
epithelia and their extracellular environments.[4–6] The ex-
tracellular domain of MUC1 consists of tandem repeats
comprising 20 amino acids of the sequence GSTAP-
PAHGVTSAPDTRPAP containing five O-glycosylation
sites. In epithelial tumor cells the expression of MUC1 is
drastically increased. This MUC1 over-expression is accom-
panied by the downregulation of a glucosaminyltransferase
(C-2GnT-1) and the concomitant over-expression of differ-
ent sialyltransferases resulting in the formation of short, pre-
maturely sialylated glycan side-chains such as the sialyl-TN,
Tumor immunotherapy utilizing the remarkable specifity of
the human immune system for a selective attack on malig-
nant cells would be a highly attractive approach for the
treatment of cancer.[1–3] An essential requirement for the de-
velopment of a functional antitumor vaccine is to focus the
highly specific immune reactions on tumor cells ideally with-
[a] Dr. S. Dziadek, Prof. Dr. H. Kunz
Institut für Organische Chemie der Universität Mainz
Duesbergweg 10–14, 55099 Mainz (Germany)
Fax : (+49)6131-39-24786
E-mail: hokunz@uni-mainz.de
[b] Prof. Dr. C. Griesinger, Dr. U. M. Reinscheid
Max-Planck Institute for Biophysical Chemistry
Am Fassberg 11, 37077 Gçttingen (Germany)
Fax : (+49)551-201-2202
(2,6)-sialyl-T saccharide antigens.[7]
aACTHER(UGN 2,3)-sialyl-T, and aACHTREUNG
E-mail: cigr@nmr.mpibpc.mpg.de
Moreover, the incomplete glycosylation in tumor cells is
supposed to lead to a changed conformation of the protein
backbone[8] and to the exposure of peptide epitopes, which
are masked in normal cells. Avariety of monoclonal anti-
bodies recognize these epitopes and specifically bind to ma-
lignant but not normal epithelial cells. Most antibodies are
[c] Dr. S. Dziadek, Dr. U. M. Reinscheid
S. Dziadek and U. M. Reinscheid contributed equally to this work.
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author: Complete list of
ROE values used for the structure calculation of the sialyl-T glycoei-
cosapeptide 13.
Chem. Eur. J. 2006, 12, 4981 – 4993
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4981

directed to the immunodominant PDTRPAP motif on the
MUC1 tandem repeat.[9] These tumor-associated structure
alterations render glycopeptide partial structures from
MUC1 valuable target molecules for the generation of im-
munostimulating antigens.
Immunizations of mice with a vaccine construct consisting
of a glycopeptide sequence from the MUC1 tandem repeat
carrying a sialyl-TN side-chain conjugated via a flexible
spacer with a specific TH-cell epitope from ovalbumin lead
to the induction of a strong, highly specific humoral immune
response against the tumor-associated MUC1 glycopep-
tide.[10] The isolated antibodies from the mouse sera exclu-
sively recognized a combination of saccharide as well as
peptide structural elements as determined by a neutraliza-
tion experiment.[10] In contrast, neither the unglycosylated
MUC1 peptide sequence alone nor the sialyl-TN saccharide
antigen attached to a different peptide chain from MUC4
were capable of binding to and hence neutralizing the anti-
body. These significant results prompted our interest in in-
vestigating the structural propensities of tumor-associated
MUC1 glycopeptides, in particular the influence of the O-
glycans on the conformation of the peptide chain, under
nearly physiological conditions in aqueous solution. We
herein propose a valuable structural model of the immuno-
genically relevant parts of the MUC1 peptide core.
Results and Discussion
Syntheses of MUC1-derived glycopeptides: To allow de-
tailed NMR based structural elucidation, eicosapeptides and
glycopeptides representing the full length tandem repeat se-
quence of the mucin MUC1 were synthesized according to
an efficient convergent strategy. In order to be able to study
the effect of O-glycosylation on the conformational propen-
sities of the underlying peptide backbone, in addition to the
glycoeicosapeptide carrying the complex tumor-associated
Scheme 1. a) MeCN, a,a-dimethoxytoluene, cat. PTSA, RT, 15 h, 75%.
b) Hg(CN)2, CH3NO2/CH2Cl2 3:2, 4 Š MS, 18 h, 93%. c) 80% AcOH,
808C, 1 h, 82%. d) MeSBr, AgOTf, 3 Š MS, CH3CN/CH2Cl2 2:1, 4 h,
ꢀ658C, 61% a-anomer, 12% b-anomer. e) TFA, anisole, 85%. DTBP=
di-tert-butylpyridine; PTSA=p-toluenesulfonic acid; TFA=trifluoroace-
tic acid.
aACHTREUNG(2,6)-sialyl-T antigen the unglycosylated MUC1 eicosapep-
tide was assembled on a solid support. The glycopeptide
structure was accessible by incorporating a pre-formed O-
glycosyl amino acid into the sequential glycopeptide synthe-
sis.
Biomimetic synthesis of the O-glycosyl amino acid building
block: The preparation of the aACTHRE(UNG 2,6)-sialyl-T threonine
removal of the benzylidene acetal with aqueous acetic acid
at 808C furnished a suitable sialyl acceptor 5. For the regio-
and stereoselective sialylation of the 6-OH group in 5, the
xanthate[16,17] 6 of the N-acetyl neuraminic acid benzyl ester
building blocks for solid-phase glycopeptide synthesis ac-
cording to a linear synthetic approach[11] mimicking the
glycan biosynthesis in tumor cells is outlined in Scheme 1.
The N-Fmoc and tert-butyl ester protected TN-antigen
threonine derivative 1[12,13] served as synthon for the assem-
bly of the (2,6)-sialyl-T-antigen. It was converted to the 4,6-
benzylidene acetal 2 using a,a-dimethoxytoluene in the
presence of catalytic p-toluenesulfonic acid in acetonitrile.
The subsequent stereoselective b-galactosylation to form the
blocked disaccharide 4[11] was accomplished employing the
6-O-benzyl protected galactosyl bromide 3[14] activated with
mercury(ii) cyanide under Helferich[15] conditions. Selective
activated with methylsulfenyl triflate[18] as
a promoter
proved to be an efficient donor. Using a mixture of acetoni-
trile[19] and dichloromethane for the glycosylation reaction,
the desired aACTHRE(UNG 2,6)-sialyl-T-threonine derivative 7 was ob-
tained in a yield of 61% after preparative RP-HPLC. Subse-
quent acidolysis of the tert-butyl ester with trifluoroacetic
acid and anisole (10:1) yielded the N-Fmoc protected (2,6)-
sialyl-T-threonine building block 8 which was incorporated
into the sequential solid-phase synthesis without O-acetyla-
tion of the sterically hindered 4-OH group.
4982
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 4981 – 4993

Glycosylated Peptides
FULL PAPER
Scheme 2. Solid-phase peptide synthesis of MUC1 eicosapeptide 10.
Solid-phase glycopeptide synthesis: Prior to the synthesis of
different glycoeicosapeptides, the unglycosylated full length
tandem repeat sequence of MUC1 was assembled by con-
densing N-Fmoc[20] and side-chain protected amino acids
(10 equiv) on Tentagel resin[21] 9 functionalized with Fmoc-
proline via the trityl linker (Scheme 2).
Couplings were achieved by activating the amino acid
building blocks with O-(1H-benzotriazol-1-yl)-N,N,N’,N’-tet-
ramethyluronium hexafluorophosphate (HBTU)[22]/N-hy-
droxybenzotriazole (HOBt)[23] and DIPEA. Following each
coupling step, unreacted amino components were capped
with acetic anhydride/HOBt and DIPEA. After completion
of the MUC1 consensus sequence and acetylation of the
amino terminus, the peptide was liberated from the resin by
acidolysis of the trityl linker under simultaneous removal of
all acid-labile side-chain protecting groups. Purification by
preparative HPLC and subsequent lyophilization furnished
the target structure 10 in a yield of 66%.
zole-1-yl)-N,N,N’,N’-tetramethyluronium
hexafluorophos-
phate
(HATU)
and
N-hydroxy-7-azabenzotriazole
(HOAt).[24] Following the coupling reaction and capping of
unreacted amino groups, the MUC1 sequence was complet-
ed by standard condensations of Fmoc-amino acids and final
acetylation of the amino terminus. Simultaneous detachment
of the glycopeptide from the polymeric support and cleav-
age of the acid-labile amino acid side-chain protecting
groups was achieved by treatment with a mixture of tri-
fluoroacetic acid, triisopropylsilane and water. The resulting
partially deblocked MUC1 glycopeptide 12 was purified by
preparative RP-HPLC and isolated in a yield of 44% based
on the proline loaded resin 9a. Final deprotection of the
glycan portion by hydrogenolysis of the benzyl groups and
O-deacetylation under ZemplØn[25] conditions furnished the
target structure 13 which was obtained in 55% yield after
purification by RP-HPLC.
To generate suitable model structures for conformational
analyses, the threonine residue at position-11, that is, outside
the immunodominant PDTRP domain, was chosen for the
attachment of the tumor-associated saccharide side chains.
For this purpose, the protected resin-bound nonapeptide 11
representing the C-terminal segment of the MUC1 tandem
repeat was prepared according to the Fmoc protocol[20]
(Scheme 3).
For the synthesis of the (2,6)-sialyl-T glycoeicosapeptide
one part of the functionalized resin 11 was employed, which
was liberated from the Fmoc group by treatment with piper-
idine (20%) in NMP. Subsequently, the glycosylated threo-
nine derivative 8 (Scheme 1) was coupled manually to the
resin-bound peptide fragment employing an excess of only
1.7 equivalents of the precious building block activated with
a combination of the coupling reagents O-(7-azabenzotria-
NMR analysis of the eicosapeptide from MUC1 and its gly-
cosylated analogue: Anumber of NMR studies have been
undertaken to elucidate the structural effects of glycosyla-
tion on peptides in general and on MUC1-derived peptides
in particular.[8,26] The drawback of “insufficient structure”
under physiological conditions was circumvented by lower-
ing the temperature, and/or addition of solvents and adjust-
ment of pH to low values.[27–31] As an example, Kirnarsky
et al.[32,33] studied glycosylated 15-mers of MUC1 at low tem-
peratures (5 and 108C) in water, and 9-mers in DMSO.
We were able to study structural effects of complex carbo-
hydrates on MUC1 derived peptides consisting of the full
length tandem repeat in a phosphate buffer of pH 6.5 at
258C. Using NMR restrained molecular dynamics structured
areas could be derived of a full length repeat substituted by
complex carbohydrates under physiological conditions. Con-
Chem. Eur. J. 2006, 12, 4981 – 4993
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4983

H. Kunz, C. Griesinger et al.
Scheme 3. a) Solid-phase peptide synthesis (SPPS): Fmoc removal (20% piperidine/NMP); coupling (steps 1–8: 1 mmol Fmoc-AA-OH, HBTU/HOBt/
DIPEA, DMF; capping: Ac2O, DIPEA, HOBt 4:1:0.12. Synthesis of (2,6)-sialyl-T glycoeicosapeptide 13 from MUC1: b) solid-phase glycopeptide syn-
thesis (SPGS): Fmoc removal (20% piperidine/NMP); coupling step 9: 1.7 equiv 8, HATU/HOAt/NMM, DMF, 4 h; steps 10–19: 1 mmol Fmoc-AA-OH,
HBTU/HOBt/DIPEA, DMF; capping: Ac2O, DIPEA, HOBt 4:1:0.12; c) TFA/TIS/H2O 15:0.9:0.9, 2 h, 44% based on the pre-loaded resin 9a); d) i) H2,
5% Pd/C, MeOH, 20 h, ii) NaOMe/methanol, pH 9.5, 55% over two steps. HATU=O-(7-aza-benzotriazole-1-yl)-N,N,N’,N’-tetramethyluronium hexa-
fluorophosphate,
HBTU=O-(1H-benzotriazol-1-yl)-N,N,N’,N’-tetramethyluronium
hexafluorophosphate,
HOAt=N-hydroxy-7-azabenzotriazole,
HOBt=N-hydroxybenzotriazole; NMM=N-methylmorpholine; TIS=triisopropylsilane; Pmc=2,2,5,7,8-pentamethylchroman-6-sulfonyl.
sequently, direct comparisons with earlier results[8,26–34] have
to be treated cautiously.
One-dimensional proton NMR spectra showed one pre-
dominant resonance for each amide NH suggesting either
one dominant isomer or fast conformational averaging on
the NMR time scale in phosphate buffer. The coexistence of
slowly interconverting conformers could be ruled out by the
absence of exchange cross peaks in the ROESY spectra and
Proton and 13C NMR resonances were assigned for two
MUC1 derived peptides: the unglycosylated full length
tandem repeat sequence 10 and the glycosylated peptide 13
(Tables 1–3).
4984
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 4981 – 4993

Glycosylated Peptides
FULL PAPER
Table 1. Chemical shifts (in ppm) of the MUC1 peptide 10 in H2O/D2O (9:1, pH 6.5) at 298 K.[a]
NH
Ha
Ca
Hb
Cb
Hg
Cg
Hd
Cd
G1
S2
T3
A4
P5
P6
A7
H8
8.212
8.276
8.143
8.148
–
3.893
4.457
4.270
4.51
4.61
4.291
4.15
4.615
42.45
55.50
58.83
47.82
58.72
nd
49.7
nd
nd
59.57
58.97
55.32
47.85
nd
50.33
58.99
51.19
nd
–
–
–
–
–
–
–
–
–
–
–
–
–
–
47.6
47.5
–
a: 3.78, b: 3.818
4.155
61.12
67.07
15.36
27.98
29.28
16.4
1.114
–
1.948
1.931
18.74
–
24.7
24.6
1.249
a: 2.263, b: 1.794
a: 2.18, b: 1.762
1.241
a: 3.207, b: 3.11
–
a: 3.742, b: 3.528
a: 3.724, b: 3.578
–
8.316
8.384
8.347
8.056
8.245
8.219
8.268
–
8.505
7.965
8.174
–
–
–
–
26.34
–
H4: 7.228
C4: 117.425
H2: 8.516
C2: 133.67
G9
a: 3.871, b: 3.915
4.140
–
–
–
–
–
–
–
–
–
–
–
nd
–
V10
T11
S12
A13
P14
D15
T16
R17
P18
A19
P20
2.02
4.137
30.18
67.07
61.25
15.38
29.29
35.19
67.07
27.46
29.28
15.16
28.88
a: 0.862, b: 0.852
a: 18.39, b: 17.61
4.327
4.371
4.521
4.323
4.64
4.236
4.558
4.325
1.116
–
–
1.945
–
1.095
1.58
1.93
–
18.74
–
–
24.5
–
18.83
23.98
24.5
–
a: 3.75, b: 3.891
1.281
a: 2.19, b: 1.829
a: 2.874, b: 2.792
4.132
a: 1.76, b: 1.66
a: 2.19, b: 1.796
1.284
–
a: 3.716, b: 3.525
–
–
3.127
a: 3.728, b: 3.525
–
–
40.611
nd
–
8.326
–
4.481
4.305
47.65
nd
a: 2.22, b: 1.92
1.945
24.5
a: 3.693, b: 3.580
nd
[a] nd: not determined.
Table 2. Chemical shifts (in ppm) of the (2,6)-sialyl-T glycoeicosapeptide 13 in H2O/D2O (9:1, pH 6.5) at 298 K.
NH
15N
Ha
Ca
Hb
Cb
Hg
Cg
Hd
Cd
G1
S2
T3
A4
P5
P6
A7
H8
8.216
8.280
8.149
8.152
–
114.16
115.37
115.73
127.99
–
3.890
4.454
4.267
4.505
4.607
4.290
4.142
4.588
42.52
55.49
58.87
47.68
58.57
60.05
48.20
52.49
43.05
59.28
57.11
54.96
47.75
60.20
51.40
58.88
51.38
60.30
47.46
61.89
–
–
–
–
–
–
–
–
–
–
–
–
a: 3.816, b: 3.777
4.144
1.248
a: 2.260, b: 1.795
a: 2.178, b: 1.754
1.242
a: 3.187, b: 3.100
–
61.04
66.95
15.21
27.82
29.11
16.28
26.59
–
30.13
77.44
61.50
14.96
29.19
38.08
66.95
29.11
29.22
15.13
29.19
1.110
–
1.944
1.920
18.65
–
24.50
24.50
a: 3.732, b: 3.527
a: 3.724, b: 3.532
47.8
47.6
–
134.15
–
–
–
–
–
–
8.316
8.328
8.328
8.016
8.660
8.450
8.432
124.35
117.14
120.09
120.03
117.02
116.23
126.18
–
110.10
114.16
124.69
–
–
–
–
H4: 7.186
117.63
–
a: 18.42, b: 17.67
H2: 8.38
G9
a: 3.838, b: 3.912
4.244
–
–
–
–
–
V10
T11*
S12
A13
P14
D15
T16
R17
P18
A19
P20
2.006
4.216
a: 0.890, b: 0.870
4.562
4.387
4.367
4.310
4.517
4.230
4.524
4.314
1.216
–
–
1.958
–
1.095
1.594
1.908
–
18.27
–
–
24.50
–
18.70
23.81
24.49
–
a: 3.760, b: 3.685
1.300
a: 2.215, b: 1.835
a: 2.64, b: 2.567
4.143
a: 1.751, b: 1.676
a: 2.186, b: 1.808
1.284
–
–
47.7
–
a: 3.722, b: 3.556
8.383
7.948
8.184
–
8.240
–
–
–
3.125
a: 3.730, b: 3.528
–
–
40.56
47.6
–
126.12
–
4.483
4.136
a: 2.124, b: 1.810
1.890
24.39
a: 3.653, b: 3.543
47.4
Table 3. Chemical shifts (in ppm) of the (2,6)-sialyl-T glycoeicosapeptide 13 in H2O/D2O (9:1, pH 6.5) at 298 K.
GalNAc
Gal
NeuNAc
H1:
H2:
H3:
H4:
4.860
4.125
3.900
4.095
4.007
3.830
3.478
7.445
1.917
C1:
C2:
C3:
C4:
C5:
C6:
99.31
48.26
77.16
68.99
69.63
63.85
H1’:
H2’:
H3’:
H4’:
H5’:
H6a’:
H6b’:
4.325
3.408
3.495
3.800
3.530
3.675
3.627
C1’:
C2’:
C3’:
C4’:
C5’:
C6’:
104.70
70.65
72.63
68.64
74.91
61.05
H3eq’’:
H3ax’’:
H4’’:
H5’’:
H6’’:
H7’’:
H8’’:
H9a’’:
H9b’’:
NH:
2.580
1.543
3.564
3.727
3.600
3.790
3.484
3.777
3.550
7.954
1.935
C3’’:
40.15
C4’’:
C5’’:
C6’’:
C7’’:
C8’’:
C9’’
68.26
51.85
72.49
71.77
68.26
62.65
H5:
H6a:
H6b:
NH:
AcNH:
15N:
CAc
121.68
22.03
:
15N:
CAc:
123.04
21.90
AcNH:
Chem. Eur. J. 2006, 12, 4981 – 4993
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4985

H. Kunz, C. Griesinger et al.
the correct number of resonances in the 1D proton spec-
trum. The presence of strong Ha(i)–NH(i+1) ROE signals
and the absence of Ha(i)–HaA(i+1) cross peaks confirmed
AHCTREUNG
CTHREUNG
that all amide bonds were in the trans configuration. A trans
configuration with respect to the X-Pro peptide bonds was
indicated by characteristic ROE cross peaks from the Ha
(X) to the Hd (Pro), the absence of cross peaks from Ha (X)
to Ha (Pro) and a chemical shift difference of around
31 ppm between Ca and Cb.[35] Additionally, the differences
in 13C chemical shifts of Cb–Cg were around 4.5 ppm which
indicated trans-peptide bonds.[36,37]
In contrast, Schuman et al.[28] observed two sets of reso-
nances in a MUC1 derived 9-mer suggesting that cis/trans
isomerization occurred at one of the X-Pro peptide bonds.
Apopulation ratio of 5:1 was inferred under the experimen-
tal conditions (58C, pH 5.1). The low signal to noise ratio of
the cis isomer precluded further structural analysis. In con-
trast to our results, glycosylation of the 9-mer shifted the cis/
trans equilibrium toward the trans isomer for residues at the
C-terminal side of the glycosylation. These differing obser-
vations might be correlated to different experimental condi-
tions concerning the peptide (sequence, length), the carbo-
hydrates (type, complexity) and the environmental parame-
ters (temperature, pH, solvent system).
Anumber of side-chain methylene proton pairs exhibited
a spectral dispersion that indicated conformational preferen-
ces even in the unglycosylated MUC1 peptide. The Hb pro-
tons of Ser12 in the unglycosylated peptide 10 showed a Dd
value of 0.15 ppm indicating a non-averaged conformation
of the c1 dihedral. Upon glycosylation, large chemical shift
dispersion was measured within the pair of NH protons of
the two Thr residues Thr11 and Thr16, the first of them
being glycosylated in 13. In this glycopeptide the difference
increased to more than 0.7 ppm. Conversely, Grinstead
et al.[38] observed for a MUC1 hexadecapeptide degenerate
Hb resonances of Ser which changed into well resolved
peaks only after glycosylation. Again, the experimental con-
ditions may explain these differences.
Figure 1. Chemical shift deviations (CSD) of Ha for the glycopeptide 13.
The CSD relative to the values of the unglycosylated peptide 10 is shown
in bright grey, while the dark gray color represents the CSD relative to
random coil values (helical: ꢀ0.39 ppm, extended: +0.37 ppm).
at Thr11 is clearly seen when the differences between the
chemical shifts of 13 and 10 are displayed (Figure 1). The
comparison with random coil values (d(13)ꢀd(random coil))
shows a deviation at Ala13, Pro14, Thr16 and Arg17 which
can be explained by sequence effects for instance induced
by the proline residues Pro14 and Pro18.
Temperature coefficients: The temperature dependence of
the NH proton chemical shifts in partially folded peptides is
a function of at least two variables: the temperature de-
pendent equilibrium between the folded and random coil
states, and the degree of structuring of the folded state at
the lower temperature.[40] For unfolded regions, temperature
coefficients (Dd/DT) are expected to be between 6 and
10 ppb per K, indicating that the backbone is freely solvated
by water and that no hydrogen bonds are present which
would protect the backbone amides from proton exchange.
In glycopeptide 13 the values of the temperature coefficients
were below 5 ppbKꢀ1 between residues Val10 and Arg17,
suggesting at least partial shielding from solvent and/or hy-
drogen bonding (Figure 2). The negative Dd/DT value ob-
served for Thr11 correlated with a downfield shift of the
HN resonance of Thr11 in the glycosylated peptide when
compared with the non-glycosylated peptide which could be
interpreted as a hydrogen bond effect.
Chemical shift deviation: Generally, Ha chemical shift devi-
ations (CSD, DdHa or Ca =dobserved ꢀ drandom coil) exhibit a mean
shift of ꢀ0.39 ppm when the residue is placed in a helical
conformation while a mean shift of +0.37 ppm is observed
when the residue is found in an extended conformation.[35,39]
The CSD values, dHa and dCa, for the unglycosylated
MUC1 derived 20-mer peptide 10 were close to random coil
(DdHa ꢁ 0.02 ppm and DdCa ꢁ 0.4 ppm) for all residues
except Ala4 and Pro5 at the N-terminus, and Ala13, Arg17
and Ala19 at the C-terminus. In addition, the Ha resonances
for residues near the site of glycosylation in the glycosylated
peptide 13 (Figure 1) showed significant downfield shifts for
Val10 and Thr11 (+0.104 and +0.235 ppm) whereas Ala13
and Asp15 Ha resonances were shifted upfield by ꢀ0.154
and ꢀ0.123 ppm, respectively. This is in agreement with an
increase in the population of extended structures in the
GVTSAregion and an ordering effect of the glycan for the
PDTR region. The influence of the position of glycosylation
Coupling constants: b Turns are characterized by a dihedral
angle of fi+1 =ꢀ608, which is consistent with a coupling con-
3
stant of JNa between 4 and 5 Hz, assuming a b turn that is
100% populated. Inverse g turns show a fi+2 of approxi-
mately ꢀ808 with a coupling constant between 6 and 8 Hz.
Because unstructured regions also display values within this
range, inverse g turns cannot be distinguished from random
coil on the basis of this coupling constant alone.[41] The high
4986
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 4981 – 4993

Glycosylated Peptides
FULL PAPER
signals near the glycosylation site was detected. They in-
clude ROE signals between the backbone NH proton of
Thr11 and the methyl as well as NH protons of the N-acetyl
group of GalNAc, and ROE signals between the b proton
and g proton of Thr11 and the anomeric H1 proton of
GalNAc. These ROE interactions suggested that rotation
about the a-glycosidic linkage is hindered.
Similar peptide–sugar connectivities have been observed
in other NMR studies of a-GalNAc O-glycosylated pepti-
des.[27,32,43] All ROE values used for structure calculation are
listed in Table S1 of the Supporting Information.
Conformational analysis of the (2,6)-sialyl-T glycoeicosa-
peptide 13 by restrained MD calculations: Atotal of 107
ROE signals were collected for the (2,6)-sialyl-T glycoeico-
sapeptide 13 and classified into four groups according to
their integrated intensities. These distance information in
combination with peptide bonds restrained in the trans con-
formation were used as input for a restrained MD simula-
tion. After energy minimization, a final set of eight low
energy structures (< 15 kcalmolꢀ1, < 0.05 nm distance re-
straint violation) was selected.
Two overlapping peptide fragments, GVTSAand
APDTR comprising the glycosylated Thr11 with flanking
residues, and the immunodominant region PDTR of MUC1,
were selected for cluster analysis to determine structural ef-
fects of glycosylation. For the GVTSAsegment the mean
pairwise RMSD for the heavy atoms of these conformers
and the corresponding average structure was equal to
0.78 Š. The structural model of 13 exhibits a clearly defined
extended, rod-like conformation for the sequence GVTSA
(Figure 4).
The directly O-linked GalNAc is positioned along one
side of an extended b strand formed by the sequence
GVTSA. Such positioning is consistent with strong contacts
between the N-acetyl NH proton of the GalNAc moiety and
the amide proton of the glycosylated Thr11 residue and be-
tween the methyl group of GalNAc and the Ha and Hb pro-
tons of Ala13. Coltart et al.[30] found a similar methyl group
association in their study of the glycosylated N-terminal
fragment STTAV of the cell surface glycoprotein CD43.
3
Figure 2. ROE connectivities, JNa and temperature coefficients for glyco-
peptide 13. The line thickness corresponds to the ROE intensity. In the
case of proline, NH refers to d protons.
value of 10.0 Hz observed for Thr11 in 13 defined a restrain-
ed dihedral at the site of glycosylation (Figure 2). The se-
quence from Val10 to Arg17 exhibited values above 7.5 Hz
which indicates either the presence of possibly interconvert-
ing turn structures and/or the presence of extended confor-
mations.[38] Considering together the CSD, J coupling and
temperature coefficients, it appears that the four residues in
13 around the site of glycosylation (Val10, Thr11, Ser12,
Ala13) and the neighboured sequence PDTR have a high
propensity for an extended and turn structure.
ROE cross peaks: The ROE connectivities were found to be
very similar for both peptides except for the glycosylated
region in 13, in which significant differences were observed.
In this area, a large number of strong consecutive daNACHTREU(NG i,i+1)
connectivities indicated the predominance of extended back-
bone conformations.[42] The important observation of ROE
signals between GalNAc and Gal that clearly indicate the ri-
gidity of the carbohydrate substituent in 13, prompted us to
further investigate this glycopeptide in detail. For this struc-
ture, exclusively interresidual ROE contacts are depicted in
Figure 3. Asignificant number of peptide–saccharide ROE
Figure 3. ROE contacts of the glycopeptide 13.
Chem. Eur. J. 2006, 12, 4981 – 4993
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4987

H. Kunz, C. Griesinger et al.
to the anti-MUC1 antibody MY.1E12 was higher than for
the analogous glycopeptide without sialylic acid substitution.
The latter two findings clearly advocate using complex
sugars as epitope-relevant structures.
For the second immunogenically important domain,
PDTR, several structural models have been proposed: a
type I b turn formed by the residues PDTR[8] and a type II
b turn formed by residues APDT were proposed as key ele-
ment of a knob-like structure.[38,50] The data presented by
Kinarsky et al.[32] did not provide direct NMR evidences
supporting the existence of either type I or type II b turn.
These authors proposed that the PDTR sequence adopts
two overlapping inverse g turns, the first spanning Pro-Asp-
Thr and the second Asp-Thr-Arg suggesting a S-shaped
bend for the PDTR fragment rather than a knob-like motif.
Figure 4. GVTSAsequence of glycopeptide 13 exhibits a rod-like secon-
dary structure. Of the trisaccharide, only GalNAc is shown.
It was suggested by Kirnarsky et al.[32] that the N-acetyl
group of the GalNAc moiety interacts directly with the pep-
tide backbone, possibly through hydrogen bonding. An in-
tramolecular hydrogen bonding between the amide proton
of GalNAc and the carbonyl oxygen of the O-linked threo-
nine residue was also proposed by Naganagowda et al.[44] as
the key stabilizing element, opposed to the O-linked serine
analogue for which this interaction seemed to be missing.
Our model based on ROE cross peaks clearly indicates a
hydrogen bond between the amide proton of GalNAc and
the carbonyl oxygen of Thr11 (N–O distance below 2.5 Š,
NHO angle > 1208) [ROEs between the NH proton of
GalNAc und Ha of Ser12 (medium intensity), Ha of Thr11
(weak intensity) and Hb of Thr11 (weak intensity)].
The strong dNNAHCTRE(UNG i,i+1) connectivities observed by Schuman
et al.[28] in their study of 9-residue peptides argued against
the existence of two overlapping inverse g turns, as this ar-
rangement would give rise to only weak dNNACHTRE(UGN i,i+1) cross
peaks between Asp and Thr and between Thr and Arg, cor-
responding to distances of 3.8 Š in each g turn. In contrast,
the observed NOEs were diagnostic of a type I b turn span-
ning Pro-Asp-Thr-Arg within the PDTRP peptide epitope
region.
With our experimental data obtained under typical bioas-
say conditions we calculated turn-like structures for the
PDTR sequence of 13 (Figure 5). The RMSD of 2.2 Š sup-
ports the view of a well-ordered secondary structure in close
vicinity to the extended, rod-like conformation of GVTSA.
The structuring effect of the carbohydrates could be ex-
plained by the observation that b-branched amino acids
favor extended conformations, due to both steric clashes
with neighboring side chains and steric clashes with main-
chain atoms.[45] Similarly, the attached carbohydrates on
Thr11 could act as extremely bulky side chains and influence
conformational equilibria of side chain as well as main-chain
dihedrals. Schuman et al.[46] concluded in their study of
serine trimers substituted by sialyl aACHTRE(UNG 2,6) GalNAc that clus-
tering of more complex carbohydrates shift the conforma-
tional equilibrium of the underlying peptide backbone
toward a more extended and rigid state.
Figure 5. APDTR sequence of glycopeptide 13 exhibits a turn-like struc-
ture.
The analysis of solely monoglycosylated peptides might
explain the differing results from Kirnarsky et al.[47] In a
recent study on a 21-mer glycosylated with GalNAc, they
identified several structural clusters for the GVTSAP se-
quence by NMR-based molecular modeling comprising
turn-like and extended conformations. In a previous report
they demonstrated a mostly extended structural shape,
termed “g-turn-like” to indicate that this turn does not fold
the peptide chain back.
The significance of our rod-like model of a complex gly-
copeptide lies in the observation that the inclusion of the
tumor-associated Tn carbohydrates at Thr3 and Ser4 up-
stream from the PDTRP core peptide epitope increased
B27.29 antibody binding affinity through direct carbohy-
drate–antibody interactions.[48] Additionally, Takeuchi
et al.[49] showed that the affinity of a sialylated glycopeptide
Aclear indication of a b-turn structure according to the
7 Š criterion[41] is the ProCa–ArgCa distance of 5.3 Š in a
representative structure of 13. Although the dihedrals for
the i+1 and i+2 residues of an ideal b turn are not fulfilled
[Asp
A
G
N
AHCTRE(UNG yi+1)=+107], which correlates to the equatorial position of
both side chains in contrast to an axial position for residue
i+2 in an ideal b turn, the short distance between the Ca of
Pro14 and Arg17 and the back folding of the backbone justi-
fy the classification as a b turn. Aclear distinction between
type I and type II can be made by the HaACTHERNGU(i+1)–NHACHTREUNG(i+2)
distance (3.5 and 2.1 Š type II). In the representative struc-
ture this distance amounts to 3.6 Š indicating a type I b turn
for the sequence PDTR.
4988
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 4981 – 4993

Glycosylated Peptides
FULL PAPER
the reaction mixture was allowed to cool to 408C and was diluted by the
addition of toluene (25 mL). The solution was concentrated in vacuo and
co-evaporated with toluene (5”25 mL). The resulting crude product was
purified by flash chromatography (silica gel; cyclohexane/ethyl acetate
1:4; column: h=19 cm, 1=3 cm) to give the title compound as a color-
less, amorphous solid (755 mg, 0.77 mmol, 82%). Rf =0.10 (cyclohexane/
ethyl acetate 1:4); [a]D22 = 34.9 (c=1.00, CHCl3); 1H NMR (300 MHz,
Anumber of different ROE values have been reported.
Schuman et al.[28] observed strong NH
A
ACHTRE(UGN i+2), NH-
ACHTREUNG(i+2)–NHACHTREUNG(i+3) and HbACHERT(UNG i+1)–NHACHTRENUG
residue peptide In the case of 13 the NH ROEs were not
strong and the last ROE was missing. Moreover, additional
HbACHTREUNG(i+1)–NHACHTREUNG(i+3) ROE values were measured for Ala13–
CDCl3): d
= 7.75 (d, J3,4 =J5,6 =7.4 Hz, 2H, H4-, H5-Fmoc), 7.59 (d,
Asp15 and Pro14–Thr16 which were also not observed by
J
H1,H2 =JH8,H7 =7.4 Hz, 2H, H1, H8), 7.46–7.15 (m, 9H, H2-, H3-, H6-,
Kirnarsky et al.[47] studying a 21-mer glycopeptide. The Ha-
H7-Fmoc; 5H, Har-Bn), 5.95 (d, JNH,H2 =8.5 Hz, 1H, NH-GalNAc), 5.51
(d, JNH,Ta =9.2 Hz, 1H, NH-Fmoc), 5.39 (d, JH3’,H4’ =2.9 Hz, 1H, H4’),
5.23–5.08 (m, 1H, H2’), 5.00–4.90 (m, 1H, H3’), 4.80 (brs, 1H, H1), 4.64–
4.33 (m, 6H, H1’, H2, CH2-Fmoc, CH2-Bn), 4.29–3.98 (m, 5H, H9-Fmoc,
Ta, Tb, H4), 3.94–3.56 (m, 5H, H6b, H6a, H5’, H3, H5), 3.55–3.35 (m, 2H,
H6’a, H6’b), 2.05, 2.04, 1.98, 1.95 (4”s, 12H, CH3-Ac), 1.43 (s, 9H, CH3-
tBu), 1.24 (brs, 3H, Tg); 13C NMR (75.5 MHz, CDCl3, BB): d = 170.19,
170.10, 169.54 (C=O), 156.37 (C=O urethane), 143.65 (C1a-, C8a-Fmoc),
141.32 (C4a-, C5a-Fmoc), 137.31 (Cq-Bn), 128.51, 128.0 (Car-Bn), 127.78
(C3-, C6-Fmoc), 127.10 (C2-, C7-Fmoc), 125.18, 124.91 (C1-, C8-Fmoc),
120.07 (C4-, C5-Fmoc), 101.57 (C1’), 100.03 (C1), 83.15 (Cq-tBu), 76.23
(Tb), 76.01 (C3), 73.58 (CH2-Bn), 72.34 (C5’), 70.85 (C3’), 69.82, 69.54
(C4, C6), 68.72 (C2’), 67.84 (C6’), 67.44 (C4’), 66.83 (CH2-Fmoc), 62.86
(C5), 59.02 (Ta), 47.59 (C2), 47.24 (C9-Fmoc), 27.95 (CH3-tBu), 23.23
(CH3-NHAc), 20.71, 20.61, 20.55 (3”CH3-OAc), 18.74 (Tg); HR-ESI-
TOF-MS (positive ion mode): m/z: calcd for C50H62N2O18Na: 979.4076;
found: 979.4094 [M+Na]+.
ACHTREUNG(i+1)–NHACHTREUNG
(i+3) cross peak observed by Schuman et al.[28]
was not seen in 13. These differences could be attributed to
different test molecules and measurement conditions. Since
the glycosylated peptide 13 representing the full length
MUC1 repeat sequence was substituted by a complex carbo-
hydrate typically found in cancer-associated cells and was
studied under physiological conditions, we are confident to
reproduce the conditions relevant for tumour immunothera-
py assays. Consequently, our proposed model of a MUC1-
derived glycopeptide may give a sound basis for modeling
approaches in antibody design.
Experimental Section
N-(9H-Fluoren-9-yl)methoxycarbonyl-O-(2-acetamido-2-deoxy-3-O-
[2,3,4-tri-O-acetyl-6-O-benzyl-b-d-galactopyranosyl]-6-O-[benzyl-(5-acet-
amido-4,7,8,9-tetra-O-acetyl-3,5-dideoxy-a-d-glycero-d-galacto-2-nonulo-
General methods: Solvents for moisture-sensitive reactions (acetonitrile,
methanol, and dichloromethane) were distilled and dried prior to use ac-
cording to standard procedures.[51] DMF (amine free, for peptide synthe-
sis) was purchased from Roth, acetic anhydride and pyridine in p.a. quali-
ty from Acros. Reagents were purchased at highest available commercial
quality and used without further purification unless outlined otherwise.
Fmoc-protected amino acids were purchased from Novabiochem. Rapp
TentaGel was used as a resin for the solid-phase synthesis. Reactions
were monitored by thin-layer chromatography with pre-coated silica gel
60 F254 aluminium plates (Merck KGaA, Darmstadt). Flash column chro-
matography was performed with silica gel (40–63 mm) purchased form
Merck KGaA, Darmstadt. Optical rotations [a]D were measured with a
Perkin–Elmer polarimeter 241. RP-HPLC analyses were carried out on a
Knauer HPLC system with Phenomenex Luna C18(2) (250”4.6 mm, 5 m)
and Phenomenex Jupiter C18 columns (250”4.6 mm, 5 m) at a pump rate
of 1 mLminꢀ1. Preparative HPLC separations were performed on a
Knauer HPLC system with a Phenomenex Luna C18(2) column (250”
50 mm, 10 mm) and a pump rate of 20 mLminꢀ1. Semipreparative HPLC
separations were carried out on a Knauer HPLC system with Phenomen-
ex Luna C18(2) (250”21.20 mm, 10 m) and Phenomenex Jupiter (250”
21.20 mm, 5 m) columns at a flow rate of 10 mLminꢀ1. Water and CH3CN
were used as solvents. 1H, 13C, and 2D NMR spectra were recorded on
Bruker AC-300, AM-400, ARX-400 or DRX-600 spectrometers. Proton
chemical shifts are reported in ppm relative to residual CHCl3 (d=7.24),
DMSO (d=2.49) or water (d=4.76). Multiplicities are given as s (sin-
glet), d (doublet), t (triplet), q (quartet), m (multiplet). 13C chemical
shifts are reported relative to CDCl3 (d=77.0) or DMSO (d=39.5). As-
signment of proton and carbon signals was achieved by COSY, TOCSY,
HMQC and HMBC experiments when noted. For 1H and 13C signals of
the saccharide portions the following denominations were used: N-
acetyl-d-galactosamine (no apostrophe); d-galactose (’); N-acetyl-neura-
minic acid (’’). MALDI-TOF mass spectra were acquired on a Micromass
Tofspec E spectrometer while ESI-mass spectra were obtained on a The-
moQuest-Navigator spectrometer. HR-ESI mass spectra were recorded
on a Micromass Q-TOF Ultima spectrometer (matrix: DHB: dihydro-
benzoic acid).
pyranosyl)onat]-a-d-galactopyranosyl)-l-threonine-tert-butylester
(7):
Fmoc-Thr(bAc3-6-Bn-Gal
AHCTREUNG
and aAc4NeuNAcCOOBnXan[17] (6; 1.10 g, 1.64 mmol, 2.5 equiv) were
dissolved in a mixture of dry acetonitrile and dry dichloromethane
(60 mL, 2:1). The solution was stirred for 1 h in a Schlenk flask (brown
glass) in the presence of flame-dried molecular sieves (3 g, powdered,
3 Š) under an argon atmosphere and the exclusion of moisture. After
cooling to ꢀ658C, dry silver triflate (421 mg, 1.64 mmol) and a pre-
cooled (ꢀ108C) solution of methyl sulfenyl bromide in dry 1,2-dichloro-
ethane[18] (1.03 mL of a 1.6m solution, 1.64 mmol) were added slowly
over 25 min. The reaction mixture was stirred at ꢀ688C for 4 h, subse-
quently neutralized with Huenigꢁs base (0.33 mL) and allowed to warm
to room temperature. The suspension was diluted with dichloromethane
(40 mL), filtered through Hyflo Super Cel and concentrated in vacuo. Pu-
rification of the crude product by flash chromatography (silica gel; ethyl
acetate, column: h=20 cm, 1=3 cm) gave an anomeric mixture which
was separated by preparative RP-HPLC (Phenomenex LUNA, acetoni-
trile/water 55:45
!
80:20, 60 min; 100:0, 30 min; l=254 nm, tR (a
anomer)=63.4 min, tR (b-anomer)=79.0 min) to yield the desired
a
anomer as a colorless amorphous solid (606 mg, 0.40 mmol, 61%, conver-
sion: 63%). In addition, the b anomer (20 mg, 0.08 mmol, 12%) as the
minor component as well as fractions of unreacted 5 (26 mg, 0.027 mmol,
4%) were isolated. a Anomer: Rf =0.21 (ethyl acetate); [a]2D3 = 12.2 (c=
1.00, CHCl3); tR =31.8 min (Phenomenex LUNA, acetonitrile/water 55:45
1
! 75:25, 40 min; 100:0, 20 min, l=254 nm); H NMR (600 MHz, CDCl3,
COSY, HMQC, HMBC): d = 7.79–7.72 (m, 2H, H4-, H5-Fmoc), 7.60 (d,
J
H1,H2 =JH8,H7 =7.8 Hz, 2H, H1-, H8-Fmoc), 7.44–7.17 (m, 14H, H3-, H6-,
H2-, H7-Fmoc; 10H, Har-Bn), 5.87 (d, JNH,H2 =8.6 Hz, 1H, NH-GalNAc),
5.50–5.39 (m, 2H, TNH {5.45}, H4’ {5.42}), 5.38–5.24 (m, 2H, H8’’ {5.32},
H7’’ {5.28}), 5.23–5.06 (m, 4H, CH2-COOBn {5.18, 5.15}, H2’ {5.13}, NH-
NeuNAc {5.09}), 4.95 (dd, JH3’,H2’ =10.3, JH3’,H4’ =3.0 Hz, 1H, H3’), 4.86–
4.77 (m, 1H, H4’’), 4.71 (d, JH1,H2 =3.0 Hz, 1H, H1), 4.66–4.44 (m, 5H,
H1’ {4.59, d, JH1’,H2’ =8.1 Hz}, H2 {4.49}, CH2-Fmoc {4.51, 4.47}, CH2a-Bn
{4.47}), 4.43–4.31 (m, 1H, CH2b-Bn), 4.30–4.19 (m, 2H, H9a’’ {4.27, dd,
JH9a’’, H9b’’ =12.3, JH9a’’,H8’’ =2.3 Hz}, H9-Fmoc {4.24}), 4.18–4.00 (m, 5H, Ta
{4.14}, Tb {4.11}, H9b’’ {4.05}, H5’’ {4.04}, H6’’ {4.03}), 3.94–3.73 (m, 4H,
H4 {3.90}, H6a {3.89}, H5’ {3.84}, H5 {3.79}), 3.66–3.57 (m, 1H, H3 {3.61}),
3.55–3.46 (m, 2H, H6b {3.52}, H6a’ {3.48}), 3.45–3.38 (m, 1H, H6b’ {3.42}),
2.58 (dd, 1H, JH3eq’’,H3ax’’ =12.6, JH3eq’’,H4’’ =4.3 Hz, H3eq’’), 2.09, 2.08, 2.05,
2.03, 1.99, 1.98, 1.95 (7”s, 24H, 8”CH3-Ac), 1.90 (m, 1H, H3ax’’), 1.84 (s,
N-(9H-Fluoren-9-yl)methoxycarbonyl-O-(2-acetamido-2-deoxy-3-O-
[2,3,4-tri-O-acetyl-6-O-benzyl-b-d-galactopyranosyl]-a-d-galactopyrano-
syl)-l-threonine-tert-butylester (5): Asolution of Fmoc-Thr( bAc3-6-Bn-
Gal
A
ous acetic acid (80%, 25 mL) was stirred at 808C for 1 h. Subsequently,
Chem. Eur. J. 2006, 12, 4981 – 4993
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4989

H. Kunz, C. Griesinger et al.
3H, CH3-Ac), 1.42 (s, 9H, CH3-tBu), 1.26 (d, JTg,Tb =6.1 Hz, 3H, Tg);
13C NMR (CDCl3, chemical shifts obtained from HMQC, HMBC): d =
170.2, 169.5, 169.4, 169.2, 166.7 (C=O), 166.4 (C1’’), 155.6 (C=O ure-
thane), 143.0 (C1a-, C8a-Fmoc), 140.5 (C4a-, C5a-Fmoc), 136.7 (Cq-Bn
(C6’)), 134.1 (Cq-Bn (C1’’)), 127.8, 127.4 (Car-Bn), 127.2 (C2-, C7-Fmoc),
127.1 (Car-Bn), 126.3 (C3-, C6-Fmoc), 124.1 (C1-, C8-Fmoc), 119.2 (C4-,
C5-Fmoc), 100.8 (C1’), 99.5 (C1), 98.1 (C2’’), 82.4 (Cq-tBu), 76.7 (C3),
75.8 (Tb), 72.7 (CH2-Bn (C6’)), 71.9 (C6’’), 71.4 (C5’), 70.0 (C3’), 68.4
(C8’’), 68.2 (C5), 68.1 (C4’’), 68.0 (C2’), 67.5 (C4), 66.9 (CH2-Bn (C1’’)),
66.6 (C4’), 66.6 (C7’’), 66.5 (C6’), 66.3 (CH2-Fmoc), 63.1 (C6), 61.6 (C9’’),
58.1 (Ta), 48.6 (C5’’), 46.9 (C2), 46.6 (C9-Fmoc), 36.8 (C3’’), 27.4 (CH3-
tBu), 22.5, 22.3 (CH3-NHAc), 20.2, 20.0, 19.9, 19.8 (CH3-OAc), 17.9 (Tg);
HR-ES-TOF-MS (positive ion mode): m/z: calcd for C76H93N3O30Na:
1550.5742, found: 1550.5731 [M+Na]+.
mixture of Ac2O (0.5m), DIPEA(0.125 m) and HOBt (0.015m) in NMP
(10 min vortex). Attachment of the glycosylated amino acids was per-
formed manually as described in the procedures for the corresponding
glycopeptides.
Ac-Gly-Ser-Thr-Ala-Pro-Pro-Ala-His-Gly-Val-Thr-Ser-Ala-Pro-Asp-Thr-
Arg-Pro-Ala-Pro-OH (10): Starting from Fmoc-Pro-O-Trt preloaded
Tentagel S resin[21] 9 (520 mg, 0.094 mmol, loading: 0.18 mmolgꢀ1), the as-
sembly of the eicosapeptide was performed according to the automated
standard protocol. After coupling of the final amino acid, Fmoc-Gly-OH,
the Fmoc group was cleaved with piperidine (20%) in NMP, and the N-
terminus was acetylated with capping reagent on the resin. For the cleav-
age procedure under simultaneous removal of the acid-labile side-chain
protecting groups, the resin was placed into a Merrifield glass reactor,
washed with dichloromethane (3”15 mL) and treated with a mixture of
trifluoroacetic acid (15.0 mL), distilled water (0.9 mL) and triisopropylsi-
lane (0.9 mL) for 2 h. After filtration, the resin was washed with tri-
fluoroacetic acid (3”3 mL), and the combined filtrates were concentrat-
ed in vacuo and co-evaporated with toluene (3”15 mL). The peptide was
precipitated by addition of cold (08C) diethyl ether (15 mL) to furnish a
colorless solid, which was washed with diethyl ether (3”10 mL), dis-
solved in distilled water and lyophilized. The crude product was purified
by preparative RP-HPLC (Phenomenex LUNAC18, acetonitrile/water
+ 0.1% TFA5:95 ! 45:55; 60 min; l=212 nm, tR =40.1 min) to give
the title compound (120 mg, 0.062 mmol, 66%) as a colorless solid after
lyophilization. [a]2D2 = ꢀ148.8 (c=1.00, H2O); tR = 14.3 min (Phenom-
enex Jupiter C18, CH3CN/H2O + 0.1% TFA5:95 ! 45:55, 30 min; l =
212 nm); 1H NMR (600 MHz, H2O/D2O 9:1, NaH2PO4/Na2HPO4 buffer,
50 mm, pH 6.50, COSY, TOCSY, 15N HSQC, 13C-HSQC, HMBC,
N-(9H-Fluoren-9-yl)methoxycarbonyl-O-(2-acetamido-2-deoxy-3-O-
[2,3,4-tri-O-acetyl-6-O-benzyl-b-d-galactopyranosyl]-6-O-[benzyl-(5-acet-
amido-4,7,8,9-tetra-O-acetyl-3,5-dideoxy-a-d-glycero-d-galacto-2-nonulo-
pyranosyl)onat]-a-d-galactopyranosyl)-l-threonine (8): Asolution of pro-
tected trisaccharide 7 (580 mg, 0.394 mmol) in a mixture of TFA(5 mL),
dichloromethane (5 mL) and anisole (0.5 mL) was stirred at ambient
temperature for 2 h. The reaction mixture was then diluted with toluene
(25 mL) and the solvent was removed in vacuo. The resulting residue was
co-evaporated with toluene (3”25 mL) and purified by flash chromatog-
raphy (silica gel; ethyl acetate/ethanol 4:1; column: h=20 cm, 1=3 cm)
and subsequently by preparative RP-HPLC (Phenomenex LUNA, aceto-
nitrile/water 60:40 ! 70:30, 70 min; l=254 nm, tR =46.5 min) to yield
compound 8 (492 mg, 0.334 mmol, 85%) as a colorless, amorphous solid.
Rf =0.51 (EE/EtOH 2:1); tR =17.1 min (Phenomenex LUNA, acetoni-
trile/water + 0.1% TFA, 55:45 ! 75:25, 30 min; l=254 nm); [a]D22
=
ROESY): d
=
8.55–8.47 (m, 2H, H(8)Im-H2 {s, 8.52}, D(15)NH {8.50}),
1
8.44–8.09 (m, 12H, H(8)NH {8.38, d, JNH,Ha =7.6 Hz}, G(9)NH {8.35},
A(19)NH {8.33}, A(7)NH {8.32}, S(2)NH {8.28}, A(13)NH {8.27}, T(11)NH
{8.24}, S(12)NH {8.22}, G(1)NH {8.21}, R(17)NH {8.17, d, JNH,ra =6.9 Hz},
A(4)NH {8.15}, T(3)NH {8.14}), 8.06 (d, JNH,Va =7.6 Hz, 1H, V(10)NH}), 7.97
(d, JNH,Ta =7.6 Hz, 1H, T(16)NH), 7.23 (s, 1H, H(8)Im-H4), 7.14–7.06 (m,
1H, R(17)NH-Gua), 4.66–4.42 (m, 8H (signal intensity reduced by H2O sup-
pression), D(15)a {4.64}, H(8)a {4.62}, P(5)a {4.61}, R(17)a {4.56}, A(13)a
{4.52}, A(4)a {4.51}, A(19)a {4.48}, S(2)a {4.46}), 4.39–4.21 (m, 8H, S(12)a
{4.37}, T(11)a {4.33}, P(18)a {4.33}, P(14)a {4.32}, P(20)a {4.31}, P(6)a
{4.29}, T(3)a {4.27}, T(16)a {4.24}), 4.20–4.09 (m, 5H, T(3)b {4.16}, A(7)a
{4.15}, V(10)a {4.14}, T(11)b {4.14}, T(16)b {4.13}), 3.98–3.65 (m, 13H,
G(9)aa {3.92}, G(1)a {3.89}, S(12)ba {3.89}, G(9)ab {3.87}, S(2)ba {3.82},
S(2)bb {3.78}, S(12)bb {3.75}, P(5)da {3.74}, P(6)da {3.73}, P(14)da {3.72},
P(18)da {3.73}, P(20)da {3.69}, P(14)db {3.53}), 3.63–3.45 (m, 4H, P(20)db
24.7 (c=1.00, CHCl3); H NMR (400 MHz, CDCl3, COSY, HMQC): d =
7.78–7.70 (m, 2H, H4-, H5-Fmoc), 7.63–7.52 (m, 2H, H1-, H8-Fmoc),
7.42–7.13 (m, 14H, H3-, H6-, H2-, H7-Fmoc, Har-Bn (10H)), 6.22 (d,
J
NH,H2 =8.2 Hz, 1H, NH-GalNAc), 5.73 (d, JNH,Ta =7.4 Hz, 1H, NH-
Fmoc), 5.44–5.23 (m, 3H, H4’ {5.40}, H8’’ {5.33}, H7’’ {5.26}), 5.23–5.05
(m, 3H, CH2-COOBn {5.17, 5.10}, H2’ {5.11}), 5.01–4.72 (m, 3H, H3’
{4.94}, H1 {4.81}, H4’’ {4.80}), 4.71 (d, 1H, H1, JH1,H2 =3.0 Hz), 4.62–4.52
(m, 1H, H1’), 4.51–4.42 (m, 3H, CH2-Fmoc {4.46}, CH2a-Bn {4.47}), 4.39–
4.25 (m, 5H, CH2b-Bn {4.36}, H2 {4.35}, H9a’’ {4.30}, Ta {4.31}, Tb {4.28}),
4.24–4.17 (m, 1H, H9-Fmoc), 4.10–3.72 (m, 7H, H9b’’ {4.04}, H6’’ {4.04},
H5’’ {4.02}, H4 {3.89}, H6a {3.88}, H5’ {3.82}, H5 {3.79}), 3.71–3.60 (m, 1H,
H3), 3.56–3.46 (m, 2H, H6b {3.50}, H6a’ {3.48}), 3.45–3.32 (m, 1H, H6b’),
2.57 (dd, JH3eq’’,H3ax’’ =8.8, JH3eq’’,H4’’ =3.9 Hz, 1H, H3eq’’), 2.15, 2.05, 2.04,
2.02, 1.98, 1.96, 1.94, 1.93 (8”s, 24H, 8”CH3-Ac), 1.88 (m, 1H, H3ax’’),
1.85 (s, 3H, CH3-Ac), 1.21 (d, JTg,Tb =6.3 Hz, 3H, Tg); 13C NMR
(100.6 MHz, CDCl3, BB, HMQC): d = 172.70 (COOH), 170.80, 170.27,
170.15, 169.87, 169.64 (C=O), 167.30 (C1’’), 155.82 (C=O urethane),
143.78 (C1a-, C8a-Fmoc), 141.25 (C4a-, C5a-Fmoc), 137.25 (Cq-Bn an
C6’), 134.83 (Cq-Bn an C1’’), 128.73, 128.67, 128.50, 128.43, 128.32, 127.78,
127.67 (Car-Bn), 127.12 (C2-, C7-Fmoc), 125.03 (C3-, C6-Fmoc), 124.92
(C1-, C8-Fmoc), 119.99 (C4-,C5-Fmoc), 102.11 (C1’), 101.30 (C1), 98.68
(C2’’), 78.74 (Tb), 77.14 (C3), 73.48 (CH2-Bn (C6’)), 73.30 (C6’’), 72.81
(C5’), 70.82 (C3’), 69.25 (C8’’), 68.92 (C5), 68.49 (C4’’), 68.17 (C2’), 67.79
(C4), 67.73 (C7’’), 67.45 (C6’), 67.02 (CH2-Fmoc), 67.35 (C4’), 66.70
(CH2-Bn (C1’’)), 63.99 (C6), 62.42 (C9’’), 58.66 (Ta), 49.25 (C5’’), 48.40
(C2), 47.20 (C9-Fmoc), 37.46 (C3’’), 23.01, 22.77 (3”CH3-NHAc), 21.02,
20.70, 20.58, 20.53, 20.47 (7”CH3-OAc), 18.31 (Tg); HR-ESI-TOF (posi-
tive ion mode): m/z: calcd for C72H85N3O30Na: 1494.5116, found:
1494.5117 [M+Na]+.
{3.58}, P(6)db {3.58}, P(5)db {3.53}, P(18)db {3.53}), 3.21 (dd, 1H, H8ba
Hba,Hbb =15.6 Hz, JHb,Ha =5.9 Hz), 3.17–3.04 (m, 3H, R(17)d {3.13}, H(8)bb
{3.11}), 2.87 (dd, JDba,Dbb =16.9, JHb,Ha =6.6 Hz, 1H, D(15)ba), 2.79 (dd,
Dba,Dbb =17.1, JHb,Ha =6.9 Hz, 1H, D(15)bb), 2.32–2.13 (m, 5H, P(5)ba
,
J
J
{2.26}, P(14)ba {2.19}, P(18)ba {2.19}, P(20)ba {2.22}, P(6)ba {2.18}), 2.07–1.99
(m, 1H, V(10)b {2.02}), 1.99–1.70 (m, 19H, AcNHterminal (1.97, s), P(5)g
{1.95}, P(20)g {1.95}, P(14)g {1.95}, P(6)g {1.93}, P(18)g {1.93}, P(20)bb
{1.92}, P(14)bb {1.83}, P(5)bb {1.79}, P(18)bb {1.80}, P(6)bb {1.76}, R(17)ba
{1.76}), 1.70–1.51 (m, 3H, R(17)bb {1.66}, R(17)g {1.58}), 1.33–1.19 (m,
12H, A(19)b {1.28}, A(13)b {1.28}, A(4)b {1.25}, A(7)b {1.24}), 1.15–1.04
(m, 9H, T(11)g {1.12}, T(3)g {1.11}, T(16}g {1.10}), 0.85 (t, 6H, V(10)g,
JVg,Vb =6.3 Hz); 13C NMR (chemical shifts taken from 13C-HSQC and
HMBC): d
=
176.21 (P(20)C=O), 174.89 (A(7)C=O), 174.05 (P(14)C=O),
174.04 (P(6)C=O), 173.78 (V(10)C=O), 173.54 (P(18)C=O), 172.61
(D(15)COOH), 172.18 (S(2)C=O), 172.01 (H(8)C=O), 171.87 (P(5)C=O), 171.64
(T(11)C=O), 171.31 (T(16)C=O), 171.19 (T(3)C=O), 171.22 (R(17)C=O), 171.20
(G(9)C=O), 133.67 (H(8)Im-C2), 128.39 (H(8)Im-C5), 117.42 (H(8)Im-C4), 67.07
(T(11)b), 67.07 (T(3)b), 67.07 (T(16)b), 61.25 (S(12)b), 61.12 (S(2)b), 60.35
(P(18)a), 60.32 (P(14)a), 60.06 (P(6)a), 59.85 (P(20)a), 59.57 (V(10)a),
58.99 (T(16)a), 58.97 (T(11)a), 58.83 (T(3)a), (P(5)a)§, 55.50 (S(2)a), 55.32
(S(12)a), (H(8)a)§, (D(15)a)§, 51.08 (R(17)a), 47.85 (A(13)a), 49.70
(A(7)a), 47.82 (A(4)a), 47.65 (A(19)a), 48.51 (P(6)d), 47.73# (P(5)d,
P(18)d), 47.65 (P(14)d), 47.44 (P(20)d), 42.45 (G(1)a), 40.61 (R(17)d),
35.19 (D(15)b), 30.18 (V(10)b), 29.29 (P(14)b), 29.28 (P(6)b), 29.28
(P(18)b), 28.88 (P(20)b), 27.98 (P(5)b), 27.46 (R(17)b), 26.34 (H(8)b),
24.55# (P(5)g, P(6)g, P(14)g, P(18)g, P(20)g), 23.98 (R(17)g), 21.61
General procedure for the automated solid-phase glycopeptide synthesis:
Peptide syntheses were performed according to the Fmoc protocol in an
automated Perkin–Elmer ABI 433 A peptide synthesizer using Fmoc-
Pro-PHB preloaded Tentagel resins.[21] In iterative cycles the peptide se-
quences were assembled by sequential coupling of the corresponding
amino acids. In every coupling step, the N-terminal Fmoc group was re-
moved by treatment of the resin (3”2.5 min) with 20% piperidine in N-
methylpyrrolidone. Amino acid couplings were carried out using Fmoc-
protected amino acids (1 mmol) activated by HBTU/HOBt[22] (1 mmol
each) and DIPEA(2 mmol) in DMF (20–30 min vortex). After every
coupling step, unreacted amino groups were capped by treatment with a
4990
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 4981 – 4993

Glycosylated Peptides
FULL PAPER
(AcNHterminal), 18.83 (T(16)g), 18.74 (T(3)g), 18.74 (T(11)g), 18.39
(V(10)ga), 17.61 (V(10)gb), 16.40 (A(7)b), 15.38 (A(13)b), 15.36 (A(4)b),
15.16 (A(19)b); §: annihilated by H2O suppression, #: signals overlapped;
MALDI-TOF-MS (DHB, positive ion mode): m/z: calcd for
C82H130N25O29: 1930.1, found: 1930.2 [M+H]+, 1952.2 [M+Na]+, 1968.2
[M+K]+, 1974.3 [M+2NaꢀH]+.
treated with a solution of 1% sodium methoxide in methanol until a pH
of 9.5 was reached. The reaction mixture was stirred for 16 h at ambient
temperature and was then neutralized by the addition of acetic acid
(0.05 mL). The solvent was removed in vacuo, and the resulting residue
was purified by preparative RP-HPLC (Phenomenex Jupiter C18, gradi-
ent: acetonitrile/water + 0.1% TFA5:95 ! 30:70, 60 min, l=212 nm,
tR =24.8 min), the deprotected MUC1 glycoeicosapeptide 13 (34 mg,
0.013 mmol, 55% over two steps) was isolated as colorless lyophilizate.
[a]2D3 =ꢀ107.3 (c=1.00, H2O); tR =14.2 min (Phenomenex Jupiter C18,
Fmoc-SerACHTREUNG(tBu)-Ala-Pro-AspCAHTRE(UNG OtBu)-ThrACHTUER(GN tBu)-ArgAHCTRE(UGN Pmc)-Pro-Ala-Pro-
Trt-Tg (11): The resin-bound peptide fragment was prepared in a 41 mL
reaction vessel of peptide synthesizer according to the standard proce-
dure starting from the Fmoc-Pro-O-Trt preloaded Tentagel S resin 9a[21]
gradient: acetonitrile/water
+
0.1% TFA5:95 ! 30:70, 30 min, l=
212 nm); 1H NMR (600 MHz, [D6]DMSO, COSY, TOCSY, HMQC,
HMBC, ROESY, NOESY): d = 8.93 (s, 1H, H(8)Im-H2), 8.27–8.20 (m,
2H, D(15)NH {8.24}, G(9)NH {8.22}), 8.17–8.06 (m, 10H, S(12)NH {8.12},
A(19)NH {8.11}, T(11)*NH {8.10}, H(8)NH {8.08}, G(1)NH {8.12}, A(7)NH
{8.04}, NH-NeuNAc {8.02}, S(2)NH {7.96}, V(10)NH {7.94}, A(13)NH {7.88}),
(1.38 g, 0.28 mmol, loading: 0.20 mmolgꢀ1
)
and using the FastMoc
(0.25 mmol) protocol for amino acid couplings. After coupling of the last
amino acid, Fmoc-Ser(tBu)-OH, the resin was thoroughly washed with
ACHTREUNG
NMP and dichloromethane and dried under a nitrogen flow. Residual
solvent was removed in high vacuum to give 1.55 g dry resin, which was
used in the syntheses of different glycopeptide structures.
7.82 (d, 1H, A(4)NH, JNH,Aa =6.9 Hz), 7.71 (d, 1H, T(16)NH, JNH,Ta
=
8.5 Hz), 7.49 (d, 1H, R(17)NH, JNH,Ra =3.8 Hz), 7.36 (s, 1H, H(8)Im-H4),
7.32 (d, 1H, T(3)NH, JNH,Ta =7.9 Hz), 7.21 (sb, 1H, OH), 7.13 (d, 1H, OH,
J=1.5 Hz), 7.06–6.99 (m, 2H, NH-GalNAc, OH), 4.82 (d, 1H, H1,
Ac-Gly-Ser-Thr-Ala-Pro-Pro-Ala-His-Gly-Val-Thr(bAc3BnGal-(1!3)-
[aAc4NeuNAcCOOBn-(2!6)]-aGalNAc)-Ser-Ala-Pro-Asp-Thr-Arg-
Pro-Ala-Pro-OH (12): Aportion of the functionalized Tentagel resin 11
(282 mg, max. 0.05 mmol) was treated in the peptide synthesizer with pi-
peridine (20%) in NMP to remove the temporary Fmoc-protecting
group. Subsequently, a solution of the (2,6)-sialyl-T threonine building
J
H1,H2 =2.1 Hz), 4.60–4.41 (m, 7H, H(8)a {4.57}, V(10)a {4.52}, D(15)a
{4.50}, A(4)a {4.49} R(17)a {4.48}, A(19)a {4.46}, T(11)*a {4.45}), 4.40–4.25
(m, 6H, 3”Pa {4.34, 4.29, 4.26}, A(13)a {4.37}, S(2)a {4.37}, S(12)a {4.28}),
4.23–4.07 (m, 8H, 2”Pd {4.22, 4.17}, T(3)a {4.19}, H1’ {4.18}, T(16)a {4.17},
T(11)*b {4.16}, H2 {4.10}, A(7)a {4.15}), 4.07–4.01 (m, 2H, T(16)b {4.04},
OH), 3.97–3.91 (m, 2H, T(3)b {3.95}, H4’ {3.93}), 3.90–3.81 (m, 5H,
G(9)aa {3.86}, H5 {3.84}, H4 {3.83}, OH), 3.80–3.27 (m, 31H, G(9)ab
{3.76}, G(1)a {3.73}, H6a {3.73}, H3 {3.67}, 5”Pd {3.64, 3.63, 3.57, 3.50,
3.44}, H7’’ {3.61}, H4’’ {3.55}, S(2)ba {3.61}, H9a’’ {3.61}, S(12)b {3.53},
S(2)bb {3.52}, H6a’ {3.52}, H5’’ {3.48}, H6b {3.44}, H6b’ {2.47}, H9b’’ {3.38},
H6’’ {3.33}, H2’ {3.30}, H8’’ {3.30}, H5’ {3.30}), 3.25–3.20 (m, 1H, H3’),
3.15–3.02 (m, 3H, H(8)ba {3.11}, R(17)d {3.08}), 3.00–2.92 (m, 1H, H8bb),
2.72 (dd, 1H, D(15)bb, JDba,Dbb =16.4 Hz, JDb,Da =5.9 Hz), 2.54–2.46 (m,
2H (partially covered by DMSO signal), D(15)bb {2.51), H3eq’’ {2.49}),
2.17–2.08 (m, 2H, 2”Pba {2.13}, {2.12}), 2.06–1.65 (m, 29H, 2”Pba {2.01,
1.98}, V(10)b {1.95}, 5”Pg {1.89, 1.88, 1.87, 1.86, 1.84}, Pb {1.83}, 4”Pbb
{1.82, 1.77, 1.76, 1.75}, AcNH’’ {s, 1.87, 3H), AcNHterminal (s, 1.85, 3H),
AcNHGalNAc (s, 1.82, 3H), Rba {1.68}), 1.56–1.46 (m, 3H, R(17)bb {1.51},
R(17)g {1.51}, H3ax’’ {1.50}), 1.22–1.10 (m, 15H, A(7)b {1.19}, A(19)b
{1.18}, A(4)b {1.16}, T(16)g {1.15}, A(13)b {1.14}), 1.01 (d, 3H, T(11}*g,
JTg,Tb =6.1 Hz), 0.99 (d, 3H, T(3)g, JTg,Tb =6.1 Hz), 0.90 (d, 3H, V(10)ga,
JVg,Vb =6.4 Hz), 0.84 (d, 3H, V(10)gb, JVg,Vb =6.7 Hz); 13C NMR (chemical
shifts obtained from 13C-HSQC and HMBC): d = 173.35, 172.64, 171.93,
171.86, 171.14, 170.50, 170.43, 170.39, 170.35, 170.00, 169.66, 169.55,
169.35 (C=O), 170.67 (C1’’), 133.94 (H(8)C4), 117.26 (H(8)C4), 105.01
[24]
block 8 (125 mg, 0.085 mmol, 1.7 equiv), HATU (34 mg, 0.090 mmol,
1.8 equiv), HOAt (12 mg, 0.090 mmol, 1.8 equiv) and N-methylmorpho-
line (19.8 mL, 0.18 mmol, 3.6 equiv) in NMP (2 mL) was added to the
resin. After shaking for 4 h, excess reagents were removed by filtration
and the resin was washed with NMP. Remaining unreacted amino groups
were acetylated with capping reagent. The eicosapeptide sequence was
completed resuming the automated standard procedure according to the
Fmoc protocol. After coupling of the last amino acid, the terminal Fmoc
group was exchanged for an acetyl group and the resin was treated with
a mixture of TFA(15 mL), distilled water (0.9 mL) and triisopropylsilane
(0.9 mL) for simultaneous cleavage of the linker and the acid-labile side-
chain protecting groups. Subsequently, the resin was washed with tri-
fluoroacetic acid (3”3 mL), the combined filtrates were concentrated in
vacuo and co-evaporated with toluene (3”15 mL). The peptide was pre-
cipitated by addition of cold (08C) diethyl ether (15 mL) to give a color-
less solid, which was washed with diethyl ether (3”10 mL), dissolved in
distilled water and lyophilized. The crude material was purified by prepa-
rative RP-HPLC (Phenomenex Jupiter C18, grad.: acetonitrile/water +
0.1% TFA25:75 ! 50:50, 80 min, l=212 nm, tR =28.1 min) to furnish
the partially protected glycopeptide 12 as a colorless lyophilizate (66 mg,
0.022 mmol, 44%). [a]2D3 =ꢀ63.8 (c=1.00, methanol); tR =26.6 min (Phe-
nomenex Luna C18(2), gradient: acetonitrile/water + 0.1% TFA15:85
! 45:55, 30 min, l=212 nm); MALDI-TOF-MS (DHB, positive ion
mode): m/z: calcd for C135H195N27O54: 3060.1, found 3059.9 [M]+, 3082.0
[M+Na]+, 3104.0 [M+2NaꢀH]+.
(2’’), 77.98 (C3), 75.72 (T(11)*b), 75.55 (C5’),
(C1’), 98.95 (C1), 98.29 CACHTREUNG
73.71 (C6’’), 73.22 (C3’), 71.52 (C7’’), 70.80 (C2’), 68.65 (C8’’), 68.34 (C4),
68.32 (C5), 66.95 (C4’), 66.95 (T(3)b), 66.48 (C4’’), 66.45 (T(16)b), 63.76
(C6), 63.17 (C9’’), 61.95 (S(12)b), 61.90 (S(2)b), 60.70 (C6’), 59.54, 59.29,
59.25, 58.76, 58.04 (5”Pa), 58.75 (T(11)*a), 58.60 (T(3)a), 58.09 (T(16)a),
57.38 (V(10)a), 55.10 (S(12)a), 55.05 (S(2)a), 52.48 (C5’’), 51.62 (H(8)a),
50.55 (R(17)a), 50.15 (D(15)a), 48.61 (A(7)a), 48.21 (C2), 46.53 (A(7)a),
46.42 (A(4)a), 46.94, 46.92, 46.80, 46.77, 46.50 (5”P(5)d), 42.22 (G(9)a),
40.78 (C3’’), 40.77 (R(17)d), 35.76 (D(15)b), 31.33 (V(10)b), 29.22, 29.16,
28.85, 28.82, 28.00 (5”Pb), 28.48 (R(17)b), 27.13 (H(8)b), 24.8–24.4 (5”
Pg)#, 24.66 (R(17)g), 22.85 (AcNH’’), 22.69 (AcNHGalNAc), 22.69
(AcNHterminal), 20.00 (T(11)*g), 19.85 (T(3)g), 19.33 (V(10)ga), 18.70
(A(13)b), 18.60 (A(4)b), 18.48 (V(10)gb), 18.42 (T(16)g), 17.81 (A(7)b),
16.87 (A(19)b); 1H NMR (600 MHz, H2O/D2O 9:1, NaH2PO4/Na2HPO4
buffer, 50 mm, pH 6.50, COSY, TOCSY, 15N-HSQC, 13C-HSQC, HMBC,
ROESY): d=8.66 (d, 1H, T*(11)NH, JNH,Ta =7.9 Hz), 8.51–7.87 (m, 16H,
S(12)NH {8.45}, A(13)NH {8.43}, D(15)NH {8.38}, H(8)H2 {8.38}, G(9) {8.33},
H(8)NH {8.33}, A(7)NH {8.32}, S(2)NH {8.28}, A(19)NH {8.24}, G(1) {8.22},
Ac-Gly-Ser-Thr-Ala-Pro-Pro-Ala-His-Gly-Val-Thr
A
(1!3)-
[aAc4NeuNAc
COOH-(2!6)]-aGalNAc)-Ser-Ala-Pro-Asp-Thr-Arg-
Pro-Ala-Pro-OH: For the removal of the benzyl groups, the (2,6)-sialyl-T
glycopeptide 12 (73 mg, 0.024 mmol) was dissolved in anhydrous metha-
nol (20 mL), and a catalytic amount of 5% palladium on activated char-
coal was added under argon. The reaction flask was subsequently purged
with H2 and the suspension was stirred for 21 h under H2 atmosphere.
The charcoal was removed by filtration through Hyflo Super Cell which
was washed with methanol (50 mL) afterwards. The combined filtrates
were concentrated in vacuo, dissolved in distilled water (5 mL) and
lyophilized to give the debenzylated glycopeptide (65 mg, max.
0.223 mmol) as a colorless lyophilizate, which was employed for the final
deprotection without further purification. tR =12.5 min (Phenomenex Ju-
piter C18, gradient: acetonitrile/water
+
0.1% TFA15:85 ! 45:55,
R(17)NH {8.18}, A(4)NH {8.15}, T(3)NH {8.15}, V(10)NH {8.02, d, JNH,Va
6.9 Hz}, T(16)NH {7.95}, NH-NeuNAc {7.95}), 7.54 (d, 1H, NH-GalNAc,
NH,H1 =9.9 Hz), 7.28–7.13 (m, 2H, R(17)NH-Gua {7.24}, H(8)H4 {7.19}), 4.86
=
30 min, l=212 nm); MALDI-TOF-MS (DHB, positive ion mode): m/z:
calcd for C121H184N27O54: 2880.9, found: 2880.8 [M+H]+, 2902.5 [M+Na]+,
2918.5 [M+K]+, 2924.4 [M+2NaꢀH]+.
J
Ac-Gly-Ser-Thr-Ala-Pro-Pro-Ala-His-Gly-Val-Thr
(bGal-
G
(m, 1H, H1 (partially covered by H2O signal)), 4.63–4.35 (m, 10H (parti-
ally covered by H2O signal), P(5)a {4.61}, H(8)a {4.59}, T(11)*a {4.56},
R(17)a {4.53}, D(15)a {4.52}, A(4)a {4.51}, A(19)a {4.48}, S(2)a {4.45},
S(12)a {4.39}, A(13)a {4.37}), 4.34–4.20 (m, 8H, H1’ {4.33}, P(18)a {4.31},
NAcCOOH-(2!6)]-aGalNAc)-Ser-Ala-Pro-Asp-Thr-Arg-Pro-Ala-Pro-
OH (13): The crude, debenzylated (2,6)-sialyl-T eicosapeptide (65 mg,
max. 0.223 mmol) was dissolved in anhydrous methanol (20 mL) and
Chem. Eur. J. 2006, 12, 4981 – 4993
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4991

H. Kunz, C. Griesinger et al.
P(14)a {4.31}, P(6)a {4.29}, T(3)a {4.27}, V(10)a {4.24}, T(16)a {4.23},
T(11)*b {4.22}), 4.19–3.97 (m, 7H, T(3)b {4.15}, A(7)a {4.14}, P(20)a {4.14},
T(16)b {4.14}, H2 {4.13}, H4 {4.09}, H5 {4.01}), 3.97–3.32 (m, 34H, G(9)aa
{3.91}, H3 {3.90}, G(1)a {3.89}, G(9)ab {3.84}, H6a {3.83}, S(2)ba {3.82}, H4’
{3.80} H7’’ {3.79}, S(2)bb {3.78}, H9a’’ {3.77}, S(12)ba {3.76}, H5’’ {3.73},
P(5)da {3.73}, P(14)da {3.73}, P(18)da {3.73}, P(6)da {3.72}, S(12)bb {3.69},
H6a’ {3.68}, P(20)da {3.65}, H6b’ {3.63}, H6’’ {3.60}, P(14)db {3.57}, H4’’
{3.56}, H9b’’ {3.55}, P(20)db {3.54}, P(6)db {3.53}, P(5)db {3.53}, H5’ {3.53},
P(18)db {3.53}, H3’ {3.50}, H8’’ {3.48}, H6b {3.48}, H2’ {3.41}), 3.25–3.02 (m,
4H, H(8)ba {3.19}, R(17)d {3.13}, H(8)bb {3.10}), 2.70–2.51 (m, 3H, D(15)ba
{2.64}, H3eq’’ {2.58}, D(15)bb {2.57}), 2.31–2.08 (m, 5H, P(5)ba {2.26},
P(14)ba {2.22}, P(18)ba {2.19}, P(6)ba {2.18}, P(20)ba {2.12}), 1.96 (s, 3H,
AcNHterminal), 1.93 (s, 3H, AcNH’’), 1.92 (s, 3H, AcNHGalNAc), 2.05–1.45
(m, 23H, V(10)b {2.01}, P(14)g {1.96}, P(5)g {1.94}, P(6)g {1.92}, P(18)g
{1.91}, P(20)g {1.89}, P(14)bb {1.84}, P(5)bb {1.80}, P(18)bb {1.81}, P(20)bb
{1.81}, P(6)bb {1.75}, R(17)ba {1.75}, R(17)bb {1.68}, R(17)g {1.59}, H3ax’’
{1.54}), 1.36–1.15 (m, 15H, A(13)b {1.30}, A(19)b {1.28}, A(4)b {1.25},
A(7)b {1.24}, T(11)*g {1.22}), 1.15–1.01 (m, 6H, T(3)g {1.11}, T(16}g {1.10}),
0.89 (d, 3H, V(10)ga, JVg,Vb =6.9 Hz), 0.87 (d, 3H, V(10)gb, JVg,Vb =6.6 Hz);
13C NMR (chemical shifts obtained from 13C-HMQC and HMBC): d =
174.93, 174.90, 173.95, 173.76, 173.66, 173.61, 173.53, 173.34, 171.61,
171.25, 170.94, 170.62 (C=O), 173.24 (C1’’), 134.15 (H(8)C2), 117.63
fied empirically as strong, medium, weak and very weak, were applied as
biharmonic restraints with lower and upper bounds of 0.20–0.25 Š, 0.20–
0.35 Š, 0.20–0.4 Š and 0.20–0.5 Š, respectively. Likewise, due to the de-
tected trans configuration of all peptide bonds, the w dihedral angle was
restrained to 1808. According to a simulated annealing approach, the re-
sulting starting molecules were heated to 600 K initially, subsequently
cooled and finally, after MD at 300 K, subjected to an energy minimiza-
tion using both steepest descent and conjugate gradient methods succes-
sively. Fifty structures were sampled. Eight structures within energy inter-
vals of 15 kcalmolꢀ1 and with maximum violation of upper limits less
than 0.5 Š were selected. The tightness of this family of conformers was
characterized by the mean pairwise RMSD for the heavy atoms or back-
bone atoms of the conformers and the corresponding average structure
of the structural family.
Acknowledgements
This work was supported by the Deutsche Forschungsgemeinschaft (SFB-
416, Project ZS and Ku 394-18) and by the Stiftung Rheinland-Pfalz für
Innovation. S.D. is grateful for a doctoral fellowship provided by the
Fonds der Chemischen Industrie and the Bundesministerium für Bildung
und Forschung (BMBF).
(H(8)C4), 104.70 (C1’), 100.07 C(2’’), 99.31 (C1), 77.44 (T(11)*b), 77.16
ACHTREUNG
(C3), 74.91 (C5’), 72.63 (C3’), 72.49 (C6’’), 71.77 (C7’’), 70.65 (C2’), 69.63
(C8’’), 68.99 (C4), 68.64 (C4’), 68.26 (C5), 68.26 (C4’’), 66.95 (T(3)b),
66.95 (T(16)b), 63.85 (C9’’), 62.65 (C6), 61.89 (P(20)a), 61.50 (S(12)b),
61.05 (C6’), 61.04 (S(2)b), 60.3 (P(18)a), 60.2 (P(14)a), 60.05 (P(6)a), 59.28
(V(10)a), 58.88 (T(16)a), 58.87 (T(3)a), 58.57 (P(5)a), 57.11 (T(11)*a),
55.49 (S(2)a), 54.96 (S(12)a), 52.49 (H(8)a), 51.85 (C5’’), 51.4 (D(15)a),
51.38 (R(17)a), 48.26 (C2), 48.20 (A(7)a), 47.75 (A(13)a), 47.68 (A(4)a),
47.8 (P(5)d, P(14)d)#, 47.6 (P(6)d), 47.6 (P(18)d), 47.46 (A(19)a), 47.4
(P(20)d), 42.52 (G(1)a), 40.56 (R(17)d), 40.15 (C3’’), 38.08 (D(15)b), 30.13
(V(10)b), 29.22 (P(18)b), 29.19 (P(20)b), 29.11 (P(6)b), 29.11 (R(17)b),
27.82 (P(5)b), 26.59 (H(8)b), 24.50 (P(5)g), 24.50 (P(6)g), 24.50 (P(14)g),
24.50 (P(18)g), 24.39 (P(20)g), 23.81 (R(17)g), 22.03 (AcNHGalNAc), 21.90
(AcNH’’), 21.67 (AcNHterminal), 18.65 (T(3)g), 18.70 (T(16)g), 18.42
(V(10)ga), 18.27 (T(11)*g), 17.67 (V(10)gb), 16.28 (A(7)b), 15.21 (A(4)b),
15.13 (A(19)b), 14.96 (A(13)b); #: signals overlapped; MALDI-TOF-MS
(DHB, positive): m/z: calcd for C107H169N27O47: 2584.17, found: 2585.5
[M]+, 2607.1 [M+Na]+, 2629.1 [M+2NaꢀH]+.
[1] J. N. Blattman, P. D. Greenberg, Science 2004, 305, 200–205.
[2] P. Moingeon, Vaccine 2001, 19, 1305–1326.
[3] S. J. Danishefsky, J. A. Allen, Angew. Chem. 2000, 112, 882–911;
Angew. Chem. Int. Ed. 2000, 39, 836–863.
[4] J. Taylor-Papadimitriou, J. M. Burchell, D. W. Miles, M. Dalziel, Bio-
chim. Biophys. Acta 1999, 1455, 301–313.
[5] J. Taylor-Papadimitriou, J. M. Burchell, T. Plunkett, R. Graham, I.
Correra, D. Miles, M. Smith, J. Mammary Gland Biol. Neoplasia
2002, 7, 209–221.
[6] A. M. Vlad, J. C. Kettel, N. M. Alajez, C. A. Carlos, O. J. Finn, Adv.
Immunol. 2004, 82, 249–293.
[7] J. M. Burchell, A. Mungul, J. Taylor-Papadimitriou, J. Mammary
Gland Biol. Neoplasia 2001, 6, 355–364.
[8] M. J. Scanlon, S. D. Morley, D. E. Jackson, M. R. Price, S. J. B. Ten-
dler, Biochem. J. 1992, 284, 137–144.
Structural analysis—Material and methods
[9] J. M. Burchell, S. J. Gendler, J. Taylor-Papadimitriou, A. Girling, A.
Lewis, R. Millis, D. Lamport, Cancer Res. 1987, 47, 5476–5482.
[10] S. Dziadek, A. Hobel, E. Schmitt, H. Kunz, Angew. Chem. 2005,
117, 7803–7808; Angew. Chem. Int. Ed. 2005, 44, 7630–7635.
[11] S. Dziadek, C. Brocke, H. Kunz, Chem. Eur. J. 2004, 10, 4150–4162.
[12] B. Liebe, H. Kunz, Angew. Chem. 1997, 109, 629–631; Angew.
Chem. Int. Ed. Engl 1997, 36, 618–621.
[13] B. Liebe, H. Kunz, Helv. Chim. Acta 1997, 80, 1473–1482.
[14] M. Lergenmüller, Y. Ito, T. Ogawa, Tetrahedron 1998, 54, 1381–
1394.
Structural NMR spectra
NMR spectra were recorded on Bruker 400 and 600 Avance spectrome-
ters. All spectra were recorded in H2O/D2O 9:1, pH 6.5 at 298 K with
peptide concentrations of 10 mm. The assignments were carried out with
the help of standard DQF-COSY, TOCSY, ROESY, 13C-HSQC, 15N-
HSQC and 13C-HMBC experiments. Typically 2k data points in F2 and
512 experiments in F1 were acquired. The spectra were acquired with 16
transients and a relaxation delay of 2 s except the ROESY experiments
with 80 transients. For ROESY experiments, a spinlock field of 2.8 kHz
was used with a mixing time of 300 ms. The TOCSY experiments were
performed with a spinlock field of 4.5 kHz using the MLEV17 sequence
with mixing times of 40 ms and 80 ms. The data were zero filled and
processed as 4k”1k matrix. PE COSY experiments were processed as
8k”2k matrix. To obtain the temperature coefficients of the amide
proton chemical shifts, TOCSY spectra were recorded between +15 and
+458C. The HN–Ha coupling constants were determined by the 1D
proton and DQF-COSY, PE COSY spectra.
[15] B. Helferich, F. Wedemeyer, Liebigs Ann. Chem. 1949, 563, 139–
145.
[16] A. Marra, P. Sinaꢂ, Carbohydr. Res. 1990, 195, 303–308.
[17] S. Keil, C. Claus, W. Dippold, H. Kunz, Angew. Chem. 2001, 113,
379–382; Angew. Chem. Int. Ed. 2001, 40, 366–369.
[18] F. Dasgupta, P. J. Garegg, Carbohydr. Res. 1988, 177, c13-c17.
[19] A. Hasegawa, H. Ohki, T. Nagahama, H. Ishida, M. Kiso, Carbo-
hydr. Res. 1991, 212, 277–281.
[20] L. A. Carpino, G. Y. Han, J. Am. Chem. Soc. 1970, 92, 5748–5749.
[21] E. Bayer, W. Rapp, Chem. Pept. Proteins 1986, 3, 3–8; all peptide
syntheses were carried out in a Perkin–Elmer ABI A 433 peptide
synthesizer (Applied Biosystems).
[22] V. Dourtoglou, B. Gross, V. Lambropoulou, C. Zioudrou, Synthesis
1984, 572.
[23] W. Kçnig, R. Geiger, Chem. Ber. 1970, 103, 788–798.
[24] L. A. Carpino, D. Ionescu, A. El-Faham, J. Org. Chem. 1996, 61,
2460.
Molecular dynamics
All molecular mechanics/dynamics simulations were performed with DIS-
COVER of the InsightII package (Accelrys) on a Silicon Graphics
Octane workstation.[53] The simulations were done using the consistent
valence force field (CVFF) that proved to account for solution NMR
data to a satisfactory extent.[54] Adielectric constant ( e=78) was used.
The molecular structures were first minimized with a gradient criterion
of less than 0.01 kcalmolꢀ1. The energy-minimized structures were then
used for MD runs. Pseudo-atoms were used for a number of methylene
proton pairs. Distance restraints derived from ROE-cross peaks, classi-
[25] G. ZemplØn, A. Kunz, Chem. Ber. 1923, 56, 1705–1710.
4992
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 4981 – 4993

Glycosylated Peptides
FULL PAPER
[26] P. Braun, G. M. Davies, M. R. Price, P. M. Williams, S. J. B. Tendler,
H. Kunz, Bioorg. Med. Chem. 1998, 6, 1531–1545.
[42] H. J. Dyson, P. E. Wright, Annu. Rev. Phys. Chem. 1996, 47, 369–
395.
[27] M. Leuck, H. Kunz, J. Prakt. Chem. 1997, 339, 322–334.
[28] J. Schuman, A. P. Campbell, R. R. Koganty, B. M. Longenecker, J.
Pept. Res. 2003, 61, 91–108.
[43] D. H. Live, L. J. Williams, S. D. Kuduk, J. B. Schwarz, D. W. Glunz,
X.-T. Chen, D. Sames, R. A. Kumar, S. J. Danishefsky, Proc. Natl.
Acad. Sci. USA 1999, 96, 3489–3493.
[29] M. Liu, B. Acres, J.-M. Balloul, N. Bizouarne, S. Paul, P. Slos, P.
Squiban, Proc. Natl. Acad. Sci. USA 2004, 101, 14567–14571.
[30] D. M. Coltart, A. K. Royyuru, L. J. Williams, P. W. Glunz, D. Sames,
S. D. Kuduk, J. G. Schwarz, X.-T. Chen, S. J. Danishefsky, D. H.
Live, J. Am. Chem. Soc. 2002, 124, 9833–9844.
[44] G. A. Naganagowda, T. L. Gururaja, J. Satyanarayana, M. J. Levine,
J. Pept. Res. 1999, 54, 290–310.
[45] D. Pal, P. Chakrabarti, Acta Crystallogr. Sect. D 2000, 56, 589–594.
[46] J. Schuman, D. Qiu, R. R. Koganty, B. M. Longenecker, A. P. Camp-
bell, Glycoconjugate J. 2000, 17, 835–848.
[31] S. Mensdorff-Pouilly, L. Kirnarsky, K. Engelmann, S. E. Baldus,
R. H. Verheijen, M. A. Hollingsworth, V. Pisarev, S. Sherman, F.-G.
Hanisch, Glycobiology 2005, 15, 735–746.
[47] L. Kirnarsky, G. Suryanarayanan, O. Prakash, H. Paulsen, H. Clau-
sen, F.-G. Hanisch, M. A. Hollingsworth, S. Sherman, Glycobiology
2003, 13, 929–939.
[32] L. Kirnarsky, O. Prakash, S. M. Vogen, M. Nomoto, M. A. Hollings-
worth, S. Sherman, Biochemistry 2000, 39, 12076–12082.
[33] L. Kirnarsky, M. Nomoto, Y. Ikematsu, H. Hassan, E. P. Bennett,
R. L. Cerny, H. Clausen, M. A. Hollingsworth, S. Sherman, Bio-
chemistry 1998, 37, 12811–12817.
[34] Reference [114] in S. Dziadek, C. G. Espinola, H. Kunz, Aust. J.
Chem. 2003, 56, 519–543.
[35] D. S. Wishart, D. A. Case, Methods Enzymol. 2001, 338, 3–34.
[36] D. E. Dorman, F. A. Bovey, J. Org. Chem. 1973, 38, 2379–2383.
[37] N. G. Delaney, V. Madison, J. Am. Chem. Soc. 1982, 104, 6635–
6641.
[38] J. S. Grinstead, R. R. Koganty, M. J. Krantz, B. M. Longenecker,
A. P. Campbell, Biochemistry 2002, 41, 9946–9961.
[39] D. S. Wishart, B. D. Sykes, F. M. Richards, J. Mol. Biol. 1991, 222,
311–333.
[40] N. H. Andersen, J. W. Neidigh, S. M. Harris, G. M. Lee, Z. Liu, H.
Tong, J. Am. Chem. Soc. 1997, 119, 8547–8561.
[48] J. S. Grinstead, J. T. Schuman, A. P. Campbell, Biochemistry 2003,
42, 14293–14305.
[49] H. Takeuchi, K. Kato, K. Denda-Nagai, F.-G. Hanisch, H. Clausen,
T. Irimura, J. Immunol. Methods 2002, 270, 199–209.
[50] J. D. Fontenot, S. V. Mariappan, P. Catasti, N. Domenech, O. J. Finn,
G. Gupta, J. Biomol. Struct. Dyn. 1995, 13, 245–260.
[51] D. D. Perrin, W. L. F. Armarego, Purification of Laboratory Chemi-
cals, Pergamon Press, Oxford, 3 ed., 1988.
[52] L. A. Carpino, H. Imazumi, A. El-Faham, F. J. Ferrer, C. Zhang, Y.
Lee, B. M. Foxmann, P. Henklein, C. Hanay, C. Mügge, H. Wen-
schuh, J. Klose, M. Beyermann, M. Bienert, Angew. Chem. 2002,
114, 457–461; Angew. Chem. Int. Ed. 2002, 41, 441–445.
[53] P. Dauber-Osguthorpe, V. A. Roberts, D. J. Osguthorpe, J. Wolff, M.
Genest, A. T. Hagler, Proteins Struct. Funct. Genet. 1988, 4, 31–47.
[54] M. Martꢃn-Pastor, J. F. Espinosa, J. L. Asensio, J. JimØnez-Barbero,
Carbohydr. Res. 1997, 298, 15–49.
[41] G. D. Rose, L. M. Gierasch, J. A. Smith, Adv. Protein Chem. 1985,
37, 1–109.
Received: February 2, 2006
Published online: April 27, 2006
Chem. Eur. J. 2006, 12, 4981 – 4993
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4993