Solution-phase synthesis with solid-state workup of an O-glycopeptide with a cluster of cancer-related T antigens

Article abstract of DOI:10.1021/ol0514193

(Chemical Equation Presented) An N-terminal glycopeptide of asialoglycophorin AM with three O-linked T antigens was prepared by "solution-phase synthesis with solid-state workup" using unprotected glycosyl amino acids as building blocks. For the glycopept

Full text of DOI:10.1021/ol0514193

ORGANIC
LETTERS
2005
Vol. 7, No. 16
3589-3592
Solution-Phase Synthesis with
Solid-State Workup of an
O-Glycopeptide with a Cluster of
Cancer-Related T Antigens
Ning Shao† and Zhongwu Guo*,†,‡
Department of Chemistry, Case Western ReserVe UniVersity, 10900 Euclid AVenue,
CleVeland, Ohio 44106-7078, and Department of Chemistry, Wayne State UniVersity,
5101 Cass AVenue, Detroit, Michigan 48202
zxg5@case.edu
Received June 18, 2005
ABSTRACT
An N-terminal glycopeptide of asialoglycophorin AM with three O-linked T antigens was prepared by “solution-phase synthesis with solid-
state workup” using unprotected glycosyl amino acids as building blocks. For the glycopeptide assembly, all reactions were conducted in
homogeneous NMP solutions, while the product of each reaction was readily isolated as solid precipitates upon addition of diethyl ether. In
the preparation of building blocks, a robust approach was established to selectively
r-glycosylate Ser and Thr derivatives.
The chemical synthesis of complex glycopeptides is a
significant challenge in organic chemistry.1 To confront it,
we have recently explored a new synthetic strategy known
as “solution-phase synthesis with solid-state workup” using
glycosyl amino acids with unprotected glycans as the key
building blocks.2,3 This strategy has several useful features.
For example, while all reactions to construct glycopeptides
were achieved in homogeneous solutions in a polar solvent,
the product of each step could be precipitated and separated
easily in the solid state, as glycopeptides with free glycans
are insoluble in most organic solvents. The use of unprotected
glycosyl amino acids could also eliminate the need to
deprotect carbohydrates at the final stage and overcome the
problem of chemical incompatibility of some peptides to
carbohydrate deprotection reactions.4 Additionally, the free
carbohydrate moieties are more stable to acids than the
corresponding O-benzylated forms,5 which will help acidic
deprotection of peptide chains, as well as the preparation of
glycopeptides with acid-labile glycans.
Even though this strategy has been utilized successfully
to synthesize a number of N-glycopeptides of considerable
complexity,2,3,6 its applicability to O-glycopeptides has yet
to be explored. Nonetheless, we anticipate that, because of
the high efficiency and simple intermediate purification of
† Case Western Reserve University.
(4) Guo, Z.; Nakahara, Y.; Nakahara, Y.; Ogawa, T. Carbohydr. Res.
1997, 303, 373-378.
‡ Wayne State University.
(1) Kunz, H. Pure Appl. Chem. 1993, 65, 1223-1232.
(2) Wen, S.; Guo, Z. Org. Lett. 2001, 3, 3773-3776.
(3) Xue, J.; Guo, Z. J. Org. Chem. 2003, 68, 2713-2720.
(5) Kunz, H.; Unverzagt, C. Angew. Chem., Int. Ed. Engl. 1988, 27,
1697-1699.
(6) Shao, N.; Xue, J.; Guo, Z. J. Org. Chem. 2003, 68, 9003-9011.
10.1021/ol0514193 CCC: $30.25
© 2005 American Chemical Society
Published on Web 07/13/2005

the new strategy, it should be especially applicable to
O-linked glycopeptides that often contain glycan clusters.
Moreover, homogeneous reaction conditions can avoid the
decrease of reaction rate and yield associated with solid-
phase synthesis for these compounds. To test the hypothesis,
the strategy was used to synthesize N-terminal glycopeptide
1 of asialoglycophorin AM containing three O-linked cancer-
related T antigens (Scheme 1).
and Thr derivatives 2, 3, and 4 that contain a free R-linked
T antigen. The preparation of R-linked glycosyl amino acids
is one of the critical issues in O-glycopeptide synthesis. Many
studies have been directed to solve this problem,17-20 but
there is still no general method to realize the reaction with
predictable stereochemistry, especially when oligosaccharide
donors are used. Consequently, our first task was to develop
an approach for the preparation of free glycosyl amino acids
2, 3, and 4.
To obtain the desired R-linkage during the glycosylation
of Ser and Thr, the amino group of galactosamine had to be
protected by a nonparticipating group. For this purpose, an
azido group was used as a latent amino group.17 A partially
protected derivative 5 of D-galatal was used as the starting
material to prepare the disaccharide donor 12 (Scheme 2),
Scheme 1
Scheme 2
Asialoglycophorin AM is the desialylated structure of
glycophorin AM, a major transmembrane sialoglycoprotein
of the human erythrocyte.7 Their N-terminal sequence
contains numerous Ser and Thr residues that are modified
by T and sialylated T antigens, respectively. Because
asialoglycophorin AM is relevant to oncological changes,8-10
its structure has gained significant interest as a research target
for cancer therapy. For example, short glycopeptide segments
of asialoglycophorin AM have been coupled with proteins,
and the resultant conjugates were investigated as potential
cancer vaccines.11-14 Indeed, chemical synthesis of glyco-
phorin A or asialoglycophorin A glycopeptides by itself is
an important subject.15,16
As shown in Scheme 1, the synthesis of glycopeptide 1
required three key building blocks, i.e., the glycosylated Ser
(7) Tomita, M.; Marchesi, V. T. Proc. Natl. Acad. Sci. U.S.A. 1975, 72,
2964-2968.
(8) Springer, G. F.; Desai, P. R.; Banatwala, I. J. Natl. Cancer Inst. 1975,
54, 335-339.
(9) Dausset, J.; Moullec, J.; Bernard, J. Blood 1959, 14, 1079-1093.
(10) Springer, S. Science 1984, 224, 1198.
(11) Kunz, H.; Birnbach, S. Angew. Chem., Int. Ed. Engl. 1986, 25, 360-
363.
(12) Kunz, H.; Birnbach, S. Carbohydr. Res. 1990, 202, 207-223.
(13) Kuduk, S. D.; Schwarz, J. B.; Chen, X.; Glunz, P. W.; Sames, D.;
Ragupathi, G.; Livingston, P. O.; Danishefsky, S. J. J. Am. Chem. Soc.
1998, 120, 12474-12485.
(14) Klich, G.; Paulsen, H.; Meyer, B.; Meldal, M.; Bock, K. Carbohydr.
Res. 1997, 299, 33-48.
(15) See review: Nakahara, Y. Trends Glycosci. Glycotechnol. 2003,
15, 257-273.
(16) (a) Bencomo, V. V.; Sinay, P. Glycoconjugate J. 1984, 1, 5-8. (b)
Ferrari, B.; Pavia, A. A. Tetrahedron 1985, 41, 1939-1944. (c) Paulsen,
H.; Merz, G.; Peters, S.; Weichert, U. Liebigs Ann. Chem. 1990, 1165-
1173. (d) Nakahara, Y.; Iijima, H.; Sibayama, S.; Ogawa, T. Tetrahedron
Lett. 1990, 31, 6897-6900. (e) Ciommer, M.; Kunz, H. Synlett 1991, 593-
595. (f) Nakahara, Y.; Iijima, H.; Shibayama, S.; Ogawa, T. Carbohydr.
Res. 1991, 216, 211-225. (g) Iijima, H.; Nakahara, Y.; Ogawa, T.
Tetrahedron Lett. 1992, 33, 7907-7910. (h) Nakahara, Y.; Iijima, H.;
Ogawa, T. Tetrahedron Lett. 1994, 35, 3321-3324. (i) Klich, G.; Paulsen,
as the transformation of glycals to corresponding 2-azido
sugars is well documented.13,21
The equatorial 3-OH of 5 was more reactive than 4-OH,
so it was possible to regioselectively glycosylate 5 at -78
°C with an R-trichloroacetimidate (6) as the glycosyl donor22
and a catalytic amount of trimethylsilyl triflate (TMSOTf,
H.; Meyer, B.; Meldal, M.; Bock, K. Carbohydr. Res. 1997, 299, 33-48.
(j) Schwarz, J. B.; Kuduk, S. D.; Chen, X.; Sames, D.; Glunz, P. W.;
Danishefsky, S. J. J. Am. Chem. Soc. 1999, 121, 2662-2673. (k) Ando, S.;
Nakahara, Y.; Ito, Y.; Ogawa, T. Carbohydr. Res. 2000, 329, 773-780.
(17) See review: Arsequell, G.; Valencia, G. Tetrahedron: Asymmetry
1997, 8, 2839-2876.
(18) Herzner, H.; Reipen, T.; Schultz, M.; Kunz, H. Chem. ReV. 2000,
100, 4495-4537.
(19) Seitz, O. ChemBioChem 2000, 1, 214-246 and references therein.
(20) Davis, B. G. Chem. ReV. 2002, 102, 579-601.
(21) Lemieux, R.; Ratcliffe, R. M. Can. J. Chem. 1979, 57, 1244-1251.
3590
Org. Lett., Vol. 7, No. 16, 2005

0.05 equiv) as the promoter. The desired disaccharide 7 was
obtained in 69% yield with an ortho ester (10%) being
identified as the major side product. No O-4 glycosylation
was observed. However, it is worth noting that the use of
0.1 equiv of TMSOTf under the same condition resulted in
degradation of galactal, probably as a result of the relatively
poor stability of the enol ether under stronger acidic
conditions. After the tert-butyldiphenylsilyl (TBDPS) group
was detached, the diol was acetylated to afford 8. Azidoni-
tration of 8 was realized by the reported process21 to form
nitrate 9 (65%). Then, the nitrate was selectively removed,
and the resultant hemiacetal 10 was transformed smoothly
into the glycosyl fluoride 11 using diethylaminosulfur
trifluoride (DAST). Thereafter, all hydroxyl groups were
benzylated following deacetylation to produce the disaccha-
ride 12 that was ready as a glycosyl donor for being coupled
with amino acids. Azidonitration of the perbenzylated
disaccharide 13 was also investigated, but we found that the
reaction was very complex and no desired product could be
isolated. It well indicates the severe influence of protecting
groups on the reactivity of substrates. Overall, the reactions
involved in the synthesis of 12 (Scheme 2) afforded good to
excellent yields, with glycosylation and azidonitration being
the key transformations.
ionic reaction mechanism, while the preferential formation
of R-anomers may result from anomeric effect.
Intriguingly, when R-trichloroimidate 18 was used as the
donor, the glycosylation of 14 gave 17â predominantly,
whereas the reaction of 15 showed no stereoselectivity. These
reactions must have gone through a different mechanism
from that of 12.
It was even more interesting to compare our results with
those of the literature.13,25 Danishefsky and co-workers13
reported that when peracylated disaccharides 19-21 were
used as glycosyl donors, the glycosylation of Ser and Thr
derivatives under a variety of conditions did not offer
meaningful stereoselectivity, although the chemical yields
were generally good. Some reactions were even in favor of
the â-anomer. It is particularly worth noting that under
conditions similar to ours the reaction of glycosyl fluoride
21 did not show preference for R-anomeric products either.
On the other hand, Ogawa and co-workers24 did find that
the glycosylation of Ser by a benzylated disaccharide 22
under these conditions gave mainly R-anomers (R/â: 3/1).
Next, we carefully studied the glycosylation of Ser and
Thr by 12. The reactions were carried out at -78 to 0 °C in
dichloromethane (DCM) with Cp2ZrCl2/AgClO4 as the
promoter.23,24 As shown in Scheme 3, these reactions gave
These findings clearly indicated that in addition to the
leaving groups at the anomeric center of the donor, the
protecting groups also had a decisive influence on the
stereochemistry of a glycosylation reaction. We initially
suspected that the cyclic acetal in 22 might play a significant
role in fostering its stereoselective reactions. However, after
comparing our results with those of literature, we think that
the electronic effects may be a major factor affecting the
stereochemical outcome, instead of the molecular geometry.
It seems that electron-withdrawing protecting groups favor
the formation of the â-anomer and electron-donating groups
favor the R-anomer.
Scheme 3
Once the azido group in the carbohydrate chain fulfilled
its missions, it was converted to an acetamino group via a
one-pot reduction-acylation protocol employing thioacetic
acid26-28 (Scheme 4). Catalytic debenzylation of the glycans
of 23 and 24 proceeded well, but a fraction of the product
had its Fmoc group removed as shown by TLC, which may
be caused by the catalyst, i.e., Pd(OH)2. Nevertheless, upon
addition of H2O to the reaction mixture, the reactions gave
25 and 4 as the only products (99%). Compound 4 was one
of the desired building blocks. To obtain 2 and 3, 25 and a
part of 4 were subjected to N-acylation and acid-catalyzed
C-terminal debutylation. To our surprise, the reaction of 25
b
a Ratio of separated products. Ratio determined by NMR.
the desired R-glycosyl amino acids (16R, 17R) as the major
products with good stereoselectivity (R/â 3/1 to 9/1). The
R- and â-glycosylation products were readily separable by
silica gel column chromatography. Since the glycosylation
of 14 and 15 by 12R, 12â or their mixture offered similar
results, these reactions must have involved the same inter-
mediate, and the results suggest an SN1 or solvent-separated
(25) Nakahara, Y.; Nakahara, Y.; Ogawa, T. Carbohydr. Res. 1996, 292,
71-81.
(22) Schmidt, R. R.; Kinzy, W. AdV. Carbohydr. Chem. Biochem. 1994,
50, 21-56.
(26) Rosen, T.; Lico, I. M.; Chu, D. T. W. J. Org. Chem. 1988, 53,
1580-1582.
(23) Suzuki, K.; Maeta, H.; Matsumoto, T. Tetrahedron Lett. 1989, 30,
4853-4856.
(27) Shangguan, N.; Katukojvala, S.; Greenberg, R.; Williams, L. J. J.
Am. Chem. Soc. 2003, 125, 7754-7755.
(28) Fazio, F.; Wong, C.-H. Tetrahedron Lett. 2003, 44, 9083-9085.
(24) Suzuki, K.; Maeta, H.; Matsumoto, T.; Tsuchihashi, G. Tetrahedron
Lett. 1988, 29, 3571-3574.
Org. Lett., Vol. 7, No. 16, 2005
3591

Scheme 4
Scheme 5
and 4 with Fmoc-OSu formed two types of products: the
acylation products 26 and 27 and the desired free acids 2
and 3 resulting from removal of the tert-butyl group. As the
reactions were performed under almost neutral conditions
and tert-butyl esters proved to be stable to the byproduct
N-hydroxysuccinimide (HOSu) in the synthesis of N-linked
glycopeptides,2,3 it was not clear how this debutylation
occurred. However, this reaction did not bother us, since both
26 and 27 were eventually converted into free acids 2 and 3
following treatment with 20% trifluoroacetic acid (TFA) in
DCM. Overall, 2 and 3 were obtained from 25 and 4 in an
excellent yield (90%).
The assemblies of amino acids and glycosyl amino acids
to build the target glycopeptide 1 followed the principles of
solution-phase synthesis with solid-state workup with 2, 3
and 4 as the key building blocks (Scheme 5). N-Methylpyr-
rolidinone (NMP) was the principal solvent to perform
homogeneous reactions, and diethyl ether was the precipita-
tion solvent to isolate products. For the coupling reaction
between 3 and 4, because both substrates contain a free
carbohydrate chain, neither should be used in excess.
Otherwise, the excess substrate would be coprecipitated with
the product. Thus, after 3 was dissolved in NMP and treated
with N,N′-dicyclohexylcarbodiimide (DCC) and N-hydroxy-
benzotriazole (HOBt) to form the activated ester in situ, 1
equiv of 4 was introduced. The solution was stirred at room
temperature overnight, and TLC showed that it gave only
one product. Diethyl ether was then added to the reaction
mixture, and the precipitates were isolated through filtration
and further washed with ether to afford 28 in 90% yield.
The 1H NMR spectra of the crude product suggested that it
was quite homogeneous. It was thereby dissolved in 20%
piperidine in NMP to remove the Fmoc group. Again the
product was isolated by precipitation and washed with diethyl
ether. Similar protocols were used for the coupling of 2 and
29 to obtain 30 and for removing the Fmoc group from 30,
as well as for the introduction of a Ser residue to the
N-terminus of the resultant glycopeptide using its commercial
pentafluorophenyl (Pfp) ester. Finally, after purification by
size exclusion chromatography, the target glycopeptide 1 was
obtained in a 70% overall yield based on 4.
In brief, a glycopeptide segment (1) of asialoglycophorin
AM containing three O-linked T antigens was synthesized
from 2, 3, and 4 via solution-phase synthesis with solid-
state workup. In the preparation of key building blocks 2, 3,
and 4, a robust method was developed for R-selective
glycosylation of amino acids. It was observed that both the
leaving groups and protecting groups of glycosyl donors
could affect the stereochemical outcome of glycosylation
reactions, with electron-donating protecting groups in favor
of producing the R-anomer. The new strategy proved to be
convenient and efficient for the synthesis of O-linked
glycopeptides with glycan clusters. The synthetic target 1 is
ready for being linked to various carrier molecules through
its N- or C-terminus to form multivalent glycoconjugates of
biological importance, e.g., as cancer vaccines.
Acknowledgment. This work was supported in part by
a grant from National Institutes of Health (R01 CA95142),
and the authors thank Mr. J. Faulk for MS measurements.
Supporting Information Available: Experimental data
and selected NMR spectra. This material is available free of
charge via the Internet at http://pubs.acs.org.
OL0514193
3592
Org. Lett., Vol. 7, No. 16, 2005