Saccharide-induced peptide conformation in glycopeptides of the recognition region of LI-cadherin

Article abstract of DOI:10.1002/anie.200602494

(Chemical Equation Presented) The type and the size of the saccharide side chain can affect the conformation of the peptide portion of glycopeptides. This was shown by conformational analysis in solution for a series of glycoundecapeptides with systematic

Full text of DOI:10.1002/anie.200602494

Communications
DOI: 10.1002/anie.200602494
Glycopeptides
Saccharide-Induced Peptide Conformation in Glycopeptides of the
Recognition Region of LI-Cadherin**
Axel Kuhn and Horst Kunz*
Dedicated to Professor Dieter Hoppe on the occasion of his 65th birthday
Glycoproteins of the outer cell membranes are components of
the glycocalix involved in fundamental recognition process-
es.[1] Among them, the cadherins are crucial to cell adhesion
and to the formation of tissues. The epithelial E-cadherin,
having five extracellular consensus repeat domains,[2] also
plays an important role in cell differentiation.[3] Its interaction
is homotypic and homophilic. The homophilic recognition site
identified by X-ray analysis and multidimensional NMR
spectroscopy is located in the N-terminal domain and consists
of three antiparallel b sheets—bC, bF, and bG—arranged
around the central recognition motif HAV (His-Ala-Val,
Figure 1a).[4,5]
Glycopeptide partial sequences of the homophilic recog-
nition region of E-cadherin were shown to induce differ-
entiation in transformed but E-cadherin-expressing keratino-
cytes.[6] The effect proved dependent upon the conformation
of the glycopeptide and was only observed if the saccharide
forced the peptide backbone to adopt a turnlike conformation
(Figure 1a). Except for the amino acid sequence not much is
known about the structure of LI-cadherin,[7] which was
discovered in 1994 in rat liver and intestine. LI-cadherin
mediates the Ca2+-dependent cell adhesion[8] and is involved
Figure 1. Homophilic recognition region of a) E-cadherin (based on
X-ray crystal structure) and b) LI-cadherin (postulated, based on
sequence homologies).
in intestine development as well as in differentiation process-
es.[9] In the N-terminal domain a recognition motif Ala-Ala-
Leu was identified.[7] It can be concluded from sequence
comparison that LI-cadherin has a homophilic recognition
region similar to that in E-cadherin, in which a turn sequence
(SQG) C-terminally follows the recognition motif AAL
(Figure 1b). To prove this hypothesis, glycopeptide sequences
I (L95 AALDSQGAIV105) of the LI-cadherin recognition
region[7] were synthesized and subjected to conformational
analysis by NMR spectroscopy.
The saccharide parts of these glycopeptides were system-
atically varied in the form of tumor-associated mucin antigens
in order to determine the effect of type and size of the
saccharide upon the peptide conformation. This appeared
attractive, as no conclusions regarding a saccharide-depend-
ant effect could be drawn from previous conformational
studies on single mucin glycopeptides.[10–15]
The 9-fluorenylmethoxycarbonyl (Fmoc)-protected O-
glycosyl serine building blocks were synthesized from Fmoc-
serine tert-butyl ester[16] via the TN antigen serine derivative
1[17] (Schemes 1 and 2). Its selective deprotection led to
compound 3,[18] which served as the substrate for the
construction of all extended saccharide side chains. All
glycosylation reactions and manipulations of protecting
groups must be carried out preserving both the base-labile
Fmoc group and the acid-labile tert-butyl ester. The 4,6-
benzylidene acetal 4 was formed under controlled acid
catalysis. Subsequent Helferich glycosylation using the gal-
actosyl bromides 5 and 6 afforded the T antigen serine
derivatives 7 and 8, respectively. Selective hydrolysis of the
benzylidene acetal 7 to give 7a was achieved by treatment
with hot aqueous acetic acid. Acetylation and treatment with
trifluoroacetic acid yielded the protected T antigen serine
building block 7b (Scheme 1).
[*] Dr. A. Kuhn, Prof. Dr. H. Kunz
Institut für Organische Chemie
Universität Mainz
Duesbergweg 10–14, 55128 Mainz (Germany)
Fax: (+49)6131-392-4786
E-mail: hokunz@mail.uni-mainz.de
Analogous acidolysis (step a in Scheme 1) converted 1
into the TN antigen serine derivative 2. Xanthate 9 of sialic
acid benzyl ester was coupled regio- and stereoselectively to
[**] This work was supported by the Volkswagen-Stiftung.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
454
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 454 –458

Angewandte
Chemie
the partially unprotected
TN antigen and T antigen
serine derivatives 3 and 7a
at ꢀ658C to give the sialyl-
TN antigen (10) and 2,6-
sialyl-T antigen (12) struc-
tures, respectively. After O-
acetylation and acidolysis
of the tert-butyl esters the
sialyl-TN (11) and 2,6-sialyl-
T antigen (13) serine build-
ing blocks were obtained
(Scheme 1).
Alternative
formation of the 2,6-sialyl-
T structure by Helferich
glycosylation of 10 (step k
in Scheme 1) afforded 12
with lower yield than that
obtained for the sialylation
of 7a. The 2,3-sialyl-T
derivative required for the
complete set of tumor-asso-
ciated mucin saccharide
antigens was synthesized
from
precursor
8
(Scheme 2). Its O-deacety-
lation with catalytic sodium
methoxide in methanol
proceeded very slowly
even at pH 9.5,[19] showing
that more rapid transesteri-
fication reactions (e.g. of 1)
proceed first by O-acetyl
migration to the 6-position.
Scheme 1. a) TFA/CH2Cl2 1:1, anisole, 6 h, RT, 73% (HPLC); b) NaOCH3, CH3OH, pH 8.5, RT, 72%;
c) benzaldehyde dimethyl acetal, p-toluenesulfonic acid, pH 4, CH3CN, RT, 63–77%; d) 5/6, Hg(CN)2, CH2Cl2/
CH3NO2, 14 h, RT, 66% (7), 91% (8); e) AcOH/H2O 4:1, 858C, 1 h, 64–90%; f) 1. Ac2O/pyridine, 08C/4 h!
14 h, RT, 97%; 2. TFA, anisole, 1.5 h, RT, 53% (HPLC); g) 9, AgOTf, CH3SBr, CH3CN/CH2Cl2, ꢀ658C, 38–
44%; h) Ac2O/pyridine, 08C/4 h!14 h, RT, 99%; i) TFA/CH2Cl2 1:1, anisole, 6 h, RT, 99%; j) 9, AgOTf,
CH3SBr, CH3CN/CH2Cl2, ꢀ658C, 20% (R=Ac), 71% (R=Bn); k) 5, Hg(CN)2, CH2Cl2/CH3NO2, 14 h, RT,
44%; l) R=Ac: 1. Ac2O/pyridine, 14 h, RT, 99%; 2. TFA, anisole, 1.5 h, RT, 70% (HPLC); R=Bn: TFA, anisole,
1.75 h, RT, 61% (HPLC). Bn=benzyl, OTf=trifluoromethanesulfonate, TFA=trifluoroacetic acid.
Scheme 2. a) NaOCH3, CH3OH, pH 9.5, 24 h, RT, 39%; b) 9, AgOTf, CH3SBr, CH3CN/CH2Cl2, 3 h/ꢀ508C!14 h/ꢀ308C, 49%; c) 1. Ac2O/pyridine,
14 h, RT, 95%; 2. AcOH/H2O 4:1, 858C, 1 h, 76%; d) Ac2O/pyridine, 14 h, RT, 99%; e) TFA, anisole, 1 h, RT, 85% (HPLC); f) 9, AgOTf, CH3SBr,
CH3CN/CH2Cl2, 3 h/ꢀ508C!14 h/ꢀ308C, 36%; g) TFA, anisole, 1.5 h, RT, 93%.
Angew. Chem. Int. Ed. 2007, 46, 454 –458
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
455

Communications
Table 1: Yields of solid-phase synthesis of peptide 19 and glycopeptides
20–25.[a]
The subsequent regio- and stereoselective sialylation with
the xanthate 9 at ꢀ508C stereoselectively produced the 2,3-
sialyl-T serine derivative 14. After acetylation of the free
hydroxy groups of the galactose portion and solvolysis of the
benzylidene acetal, the partially unprotected galactosamine
derivative 15 was furnished. Its O-acetylation and acidolysis
gave the 2,3-sialyl-T serine building block 16. By selective
sialylation of 15, again using xanthate 9 at ꢀ508C, the most
complex structure, the glycophorin antigen 17, was formed.
Subsequent acidolysis of the tert-butyl ester resulted in the
formation of building block 18. The solid-phase syntheses of
LI-cadherin glycopeptides I were carried out using a Wang
anchor[20] on Tentagel resin[21] preloaded with Fmoc valine
according to published procedures[22] (Scheme 3).
Building block
–
2
7b
11
13
16
18
(Glyco-)Peptide
Yield [%][b]
Yield (overall) [%]
19
99
99
20
78
57
21
46
23
23
61
24
24
84
57
22
23
17
25
22
21
[a] The glycopeptides were prepared by solid-phase synthesis using the
glycosyl serine building blocks 2, 7b, 11, 13, 16, and 18. [b] After cleavage
from solid phase.
NOE contacts,[24] a three-dimensional structure of each
(glyco)peptide (Figure 2) was determined by molecular
dynamics calculations and energy minimization by an MM2
force field method.[25,26] Accord-
ing to published procedures, the
found dihedral angles within the
turn region (Table 2) were com-
pared to the ideal dihedral angles
reported for the known turn
types.[27] The deviations from
these turn types are given in
Table 2as the sums of the differ-
ences in these angles.
From these values the follow-
ing conclusions were drawn
(Figure 2). Peptide 19 adopts a
looplike conformation (NH con-
tacts in the sequence DSQ) in the
vicinity of the turn sequence
DSQG. It does not resemble
any ideal turn conformation
(contacts from NH-A9 to a- and
g-CH-L4). The structure is remi-
niscent of a beginning helix that
is not continued towards the C
terminus. The C terminus folds
back to the turn sequence (a-
CH-Q7 to a-CH-V11), resulting in
a kind of double loop. The sit-
uation for the TN-antigen glyco-
peptide 20 is similar. Neverthe-
less, comparison of the dihedral
angles (Table 2) suggests a b III’
turn formed within the sequence
Scheme 3. Solid-phase syntheses of glycopeptides 20–25 containing the amino sequence LAALDSQ-
GAIV.
For comparison, the non-glycosylated LI-cadherin pep-
tide 19 was also synthesized using the same procedures. The
yields of products isolated after acidolytic cleavage from the
solid support as well as the yields of peptide 19 and
glycopeptides 20–25[23] after cleavage of the remaining
benzyl groups by hydrogenolysis, removal of the O-acetyl
groups by transesterification, and purification by HPLC are
listed in Table 1.
To analyze their preferred conformation, peptide 19 and
glycopeptides 20–25 were investigated by NOESY and
ROESY NMR spectroscopy in [D6]DMSO solution. For
discussion of the most important NOE contacts, the amino
acids of the LI-cadherin sequence I are designated from the
N terminus L1 to the C terminus V11. Based on the observed
LDSQ, which, however, cannot assume a stable b-sheet
structure (contacts between a-CH-L4 and a-CH-V11, and
between g-CH-L4 and a-CH-I10).
The conformation of the sequence DSQG of T-antigen
glycopeptide 21 is most like a b-II turn (Table 2, Figure 2).
The saccharide apparently stabilizes this conformation (con-
tacts between CH-NAc and a-CH-D5, a-CH-S6, and NH-D5).
The N-terminal sequence and the C-terminal sequence adopt
a conformation similar to a b sheet (g-CH-L4 and g-CH-I10, b-
CH-L4 and NH-V11, and NH-D5 and b-CH-I10).
In sialyl-T antigen LI-cadherin glycopeptides 22, 24, and
25 the larger saccharides exhibit a marked influence on the
conformation of the peptide backbone, which according to
NOE contacts apparently forms an increasingly closer loop
456
www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 454 –458

Angewandte
Chemie
Table 2: Dihedral angles of (glyco)peptides within the amino acid sequence LAALDSQGAIV.[a]
Antigen
Turn sequence
Calculated dihedral angles
Deviation from the ideal turn (type b)
(type g)
g g’
Fi+1
Yi+1
Fi+2
Yi+2
I
I’
II
II’
III
III’
20 TN
23 STN
21
LDSQ
DSQ
DSQG
43.8
39.1
69.7
28.7
361
74
224
354
362
41
115 125
20 260
119 277
158 283
45 239
34 245
77.1 ꢀ63.0
T
ꢀ19.3
160.4
121.9
ꢀ2.9
66.2
59.0
ꢀ98.0 ꢀ168.8
ꢀ5.3
ꢀ0.7
354
397
326
225
200
165
446
375
62
365
646
545
473
357
292
136
345
397
321
272
218
168
449
376
24 (2!6)ST DSQG
22 (2!3)ST
25 S2T
LDSQ
DSQG
81.9 ꢀ36.8
45.1 ꢀ79.5 ꢀ159.5
0.9
[a] The smallest sums of deviation from the ideal dihedral angles are printed in bold.
V11, between b-CH-L4 and g-CH-I10; 25: between b-CH-L4
and g- and d-CH-I10). It is characteristic for the conformations
of these compounds that only few direct interactions between
the saccharide and peptide parts are observable in most cases
and restricted to the GalNAc group (22: between AcNH-
GalNac and a-CH-A9, between H-2-GalNAc and b-CH-S6;
24: between CHAc-GalNAc and a-CH-D5/a-CH-A9; 25:
between H-2-GalNAc and NH-G8).
The described conformational analyses of LI-cadherin
glycopeptides not only prove that the saccharide side chains
exert a marked influence upon the conformation of the
peptide chain (Figure 2) but also show that this effect is
dependent on the type and size of the saccharide. Beyond this
case,[7] it is important to note that this conformational
influence was investigated for the tumor-associated mucin
saccharide antigens. With a glycopeptide vaccine from tumor-
associated mucin MUC1 a specific immune response in mice
could be induced. The induced antibody reacted neither with
the non-glycosylated peptide of identical sequence nor with
the tumor-associated saccharide antigen (sialyl-TN), but only
with the glycopeptide,[28] probably indicating that the peptide
chain adopts the recognized conformation only within the
glycopeptide.
Received: June 21, 2006
Revised: September 20, 2006
Published online: December 8, 2006
Figure 2. Preferred conformations of peptide 19 and glycopeptides 20–
25 of LI-cadherin determined by NOESY/ROESY experiments in
[D6]DMSO. The saccharide (more intensively colored) is positioned to
the right of the corresponding peptide (C gray, O red, N blue).
Keywords: cadherins · conformational analysis · glycopeptides ·
.
solid-phase synthesis
with antiparallel alignment of the N- and C-terminal sequen-
[1] a) N. Jentoft, Trends Biochem.Sci. 1990, 15, 291; b) J. Dekker,
J. W. A. Rossen, H. A. Büller, A. W. C. Einerhand, Trends
Biochem.Sci. 2002, 27, 126.
ces. Contacts between NHAc and A9 (in 22 and 24) and
8
between H-2(GalNAc) and NH-G (in 25) suggest that the
[2] M. Takeichi, Science 1991, 251, 1451.
saccharide is oriented parallel to this loop (in Figure 2, right).
From the deviations in the dihedral angles (Table 2) one can
conclude that a g-turn-like conformation that adopts an
expanded and twisted form is most likely. This especially
applies to 2,6-sialyl antigen glycopeptide 24. A g-turn
conformation can also be assumed for the sequence DSQ of
sialyl-TN antigen glycopeptide 23, with backfolding of the C
terminus (NOE contacts between a-CH-Q7 and g-CH-I10, and
between NH-G8 and b-A9) similar to 20. For the sialyl-T
glycopeptides, the interactions between the peptide sequen-
ces in the N- and C-terminal regions are typical (22: between
a-CH-A3 and NH-I10/NH-V11; 24: between NH-L4 and a-CH-
[3] a) O. Huber, C. Bierkamp, R. Kemler, Curr.Opin.Cell Biol.
1996, 8, 685; b) A. W. Koch, S. Pokutta, A. Lustig, J. Engel,
Biochemistry 1997, 36, 3507.
[4] M. Overduin, T. S. Harvey, S. Bagby, K. I. Tong, P. Yan, M.
Takeichi, M. Ikura, Science 1995, 267, 386.
[5] B. Nagar, M. Overduin, M. Ikura, J. M. Rini, Nature 1996, 380,
360.
[6] J. Habermann, K. Stüber, T. Skripko, T. Reipen, R. Wieser, H.
Kunz, Angew.Chem. 2002, 114, 4423; Angew.Chem.Int.Ed.
2002, 41, 4249.
[7] D. Berndorff, R. Gessner, B. Kreft, N. Schnoy, A.-M. Lajous-
Petter, N. Loch, W. Reutter, M. Hortsch, R. Tauber, J.Cell Biol.
1994, 125, 1353.
Angew. Chem. Int. Ed. 2007, 46, 454 –458
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
457

Communications
[8] B. Kreft, D. Berndorff, A. Böttinger, S. Finnemann, D. Wedlich,
[18] T. Reipen, H. Kunz, Synthesis 2003, 2487.
M. Hortsch, R. Tauber, J.Cell Biol. 1997, 136, 1109.
[9] a) D. G. Silberg, G. P. Swain, E. R. Suh, P. G. Traber, Gastro-
enterology 2000, 119, 961; b) C. Grötzinger, J. Kneifel, D.
Patschan, N. Schnoy, I. Anagnostopoulos, S. Faiss, R. Tauber,
B. Wiedenmann, R. Geßner, Gut 2001, 49, 73; c) M. Takamura,
M. Sakamoto, Y. Ino, T. Shimamura, T. Ichida, S. Hirohashi,
Cancer Sci. 2003, 94, 425.
[10] a) M. J. Scanlon, S. D. Morley, D. E. Jackson, M. R. Price, S. B. J.
Tendler, Biochem.J. 1992, 284, 137; b) P. Braun, G. M. Davies,
M. R. Price, P. M. Williams, S. B. J. Tendler, H. Kunz, Bioorg.
Med.Chem. 1998, 6, 1531.
[11] a) L. Kinarsky, M. Nomoto, Y. Ikematsu, H. Hassan, E. P.
Bennett, R. L. Cerny, H. Clausen, M. A. Holtingworth, S.
Sherman, Biochemistry 1998, 37, 12811; b) S. Mensdorff-Pouilly,
L. Kinarsky, K. Engelmann, S. E. Baldus, R. H. Verheien, M. A.
Hollinworth, V. Pisarev, S. Sherman, F.-G. Hanisch, Glycobiol-
ogy 2005, 15, 735.
[12] D. M. Coltart, A. K. Royyuru, L. J. Williams, P. W. Glunz, D.
Sames, S. D. Kuduk, J. B. Schwarz, X.-T. Chen, S. J. Danishefsky,
D. H. Live, J.Am.Chem.Soc. 2002, 124, 9833.
[19] C. Brocke, H. Kunz, Synlett 2003, 2052.
[20] S. S. Wang, J.Am.Chem.Soc. 1973, 95, 1328.
[21] E. Bayer, W. Rapp, Chem.Pept.Proteins 1986, 3, 3.
[22] C. Brocke, H. Kunz, Synthesis 2004, 525.
[23] [a] values (c in DMSO) and 13C NMR (103.0 MHz, [D6]DMSO):
20: [a] = ꢀ1.1 degcm3 gꢀ1 dmꢀ1 (c = 2gcmꢀ3), d = 97.6 ppm
(C1); 21: [a] = + 8.1 degcm3 gꢀ1 dmꢀ1 (c = 1.5 gcmꢀ3), d =
104.2ppm (C1 ’’, Gal), 97.8 ppm (C1); 22: [a] =
ꢀ9.5 degcm3 gꢀ1 dmꢀ1 (c = 0.9 gcmꢀ3); d = 57.3 (Vala), 57.2
(Sera), 22.5 (NAc, GalNAc), 22.4 ppm (NAc, NeuNAc). 23:
[a] = ꢀ0.5 degcm3 gꢀ1 dmꢀ1 (c = 1.2gcm ꢀ3), d = 97.5 (C1),
56.8 ppm (Vala); 24: [a] = + 6.3 degcm3 gꢀ1 dmꢀ1 (c =
3.6 gcmꢀ3), d = 103.7 (C1, Gal), 97.5 ppm (C1, GalNAc); 25:
[a] = 8.3 degcm3 gꢀ1 dmꢀ1 (c = 1 gcmꢀ3), d = 101.3 (C1; Gal),
97.7 ppm (C1, GalNAc).
[24] Contacts used for calculations are listed in the Supporting
Information. The influence of the solvent (D2O/[D6]DMSO) is
also discussed. Since the NMR spectra of the glycopeptide Ac-
AALDSC(GalNAc)QGAI-OH (similar to compound 20) in
D2O and in [D6]DMSO showed no important differences, all
NMR experiments of 19–25 were carried out in [D6]DMSO
solution.
[13] a) J. Schuman, A. P. Campbell, R. R. Koganly, B. M. Longe-
necker, J.Pept.Res.
2003, 61, 91; b) J. S. Grꢀnstead, J. T.
Schuman, A. P. Campbell, Biochemistry 2003, 42, 14293.
[14] M. Liu, B. Acress, J.-M. Balloul, N. Bizonarne, S. Paul, P. Slos, P.
Squiban, Proc.Natl.Acad.Sci.USA 2004, 101, 14567.
[15] a) S. Dziadek, C. G. Espinola, H. Kunz, Aust.J.Chem. 2003, 56,
519; b) S. Dziadek, C. Griesinger, H. Kunz, U. M. Reinscheid,
Chem.Eur.J. 2006, 12, 4981.
[16] H. Kunz in Preparative Carbohydrate Chemistry (Ed.: S.
Hanessian), Marcel Dekker, New York, 1997, pp. 265 – 281, in
particular pp. 274–275.
[25] U. Burkert, N. L. Allinger, Molecular Mechanics American
Chem.Soc. , Washington, DC, 1982.
[26] Chem3D. Version 9.0, CambridgeSoft 1986–2004, Cambridge,
MA, USA.
[27] a) J. S. Richardson, Adv.Protein Chem. 1981, 34, 167; b) D. S.
Wishart C. G. Bigam, A. Holm, R. S. Hodges, B. D. Sykes, J.
Biomol.NMR 1995, 5, 67.
[28] S. Dziadek, A. Hobel, E. Schmitt, H. Kunz, Angew.Chem. 2005,
117, 7803; Angew.Chem.Int.Ed. 2005, 44, 7630.
[17] H. Paulsen, K. Adermann, Liebigs Ann.Chem. 1989, 751.
458
www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 454 –458