Hydroxynorleucine as a glycosyl acceptor is an efficient means for introducing amino acid functionality into complex carbohydrates

Article abstract of DOI:10.1016/S0040-4039(03)00517-3

A new approach to the synthesis of biologically relevant glycosyl amino acids using a non-natural amino acid as the glycosyl acceptor is described. The procedure involves a glycosylation reaction of a suitable carbohydrate donor with Fmoc-L-hydroxynorleucine benzyl ester. This reaction results in the direct incorporation of the amino acid moiety. The acceptor can be used for the preparation of α- or β-O-linked glycosides depending upon the nature of the glycosyl donor. This method has been applied in the synthesis of six different tumor-associated carbohydrate antigens.

Full text of DOI:10.1016/S0040-4039(03)00517-3

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 44 (2003) 3413–3416
Hydroxynorleucine as a glycosyl acceptor is an efficient means
for introducing amino acid functionality into complex
carbohydrates
Stacy J. Keding,a Endo Atsushi,a Kaustav Biswas,a Andrzej Zatorski,a Don M. Coltarta and
Samuel J. Danishefskya,b,
*
aLaboratory for Bioorganic Chemistry, Sloan-Kettering Institute for Cancer Research, 1275 York Avenue, New York, NY 10021,
USA
bDepartment of Chemistry, Columbia University, New York, NY 10027, USA
Received 4 February 2003; revised 21 February 2003; accepted 21 February 2003
Abstract—A new approach to the synthesis of biologically relevant glycosyl amino acids using a non-natural amino acid as the
glycosyl acceptor is described. The procedure involves a glycosylation reaction of a suitable carbohydrate donor with Fmoc-L-
hydroxynorleucine benzyl ester. This reaction results in the direct incorporation of the amino acid moiety. The acceptor can be
used for the preparation of a- or b-O-linked glycosides depending upon the nature of the glycosyl donor. This method has been
applied in the synthesis of six different tumor-associated carbohydrate antigens. © 2003 Elsevier Science Ltd. All rights reserved.
It is known that certain types of glycoconjugates found
in normal tissue types are often significantly overex-
pressed in tumors of that tissue.1 Given effective
immunosurveillance, a host would generate an immune
response directed to these tumor-associated antigens.
Unfortunately, initial antibody response to tumor anti-
gens is apparently inefficient in eradicating the disease,
thus tumor progression ensues. The idea of using syn-
thetically derived and cell-free glycoconjugates of
tumor-associated antigens in the development of anti-
tumor vaccines to generate an increased immune
response is attractive. One of the main drawbacks in
exploring the viability of this approach is the limited
availability of purified tumor-associated carbohydrate
antigens. Therefore, total chemical synthesis has
become increasingly important for providing probe
structures to be evaluated in such a program. Our
involvement in numerous synthetic and subsequent pre-
clinical and clinical studies with many different tumor-
associated carbohydrate antigens has intensified our
quest for optimizing the process of vaccine design and
synthesis.2
Based on our earlier investigations, we have come to
prefer the display of tumor antigens on a peptide
backbone. Efforts in our laboratory for introducing
amino acid functionality to tumor-associated carbohy-
drate antigens have included glycosylations with serine
or threonine to form the core-mucin structure,3 trans-
formation of O-pentenyl glycosides via an ozonolysis–
Wittig–asymmetric hydrogenation protocol4 and
cross-metathesis of O-allyl glycosides with allylglycine.5
The first of these methods introduces amino acids of
natural origin while the latter two result in non-natural
amino acid side chains.
Scheme 1. Synthesis of Fmoc-L-hydroxynorleucine benzyl
ester. Reagents: (a) 2,2,2-trichloroethyl chloroformate, pyr.,
CH2Cl2; (b) 4N HCl/dioxane; (c) Fmoc-OSu, NaHCO3, ace-
tone/H2O; (d) BnBr, NaHCO3, DMF; (e) Zn, AcOH/THF.
Our ongoing efforts in the total synthesis of carbohy-
drate-based antitumor vaccines have led us to favor the
non-natural amino acids as components of our peptide-
linked vaccines.6 The use of vaccines incorporating
non-natural amino acids as linkers to carbohydrate
Keywords: hydroxynorleucine; glycosylation; glycosylamino acid; car-
bohydrate-based antitumor vaccines.
* Corresponding author. E-mail: s-danishefsky@ski.mskcc.org
0040-4039/03/$ - see front matter © 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0040-4039(03)00517-3

3414
S. J. Keding et al. / Tetrahedron Letters 44 (2003) 3413–3416
Table 1. Comparison of glycosylations of donor 4 with three different acceptors
Entry
Acceptor
Temperature (°C)
Solvent
Yield (%)
a:ba
1
2
3
4
5
6
4-Penten-1-ol
1
1
1
−10
−10
−40
−78
−78
−78
THF
THF
THF
THF
CH2Cl2
THF
90
39
47
88
80
86
1:1
2:1
2.5:1
4:1
1:9
1:0
1
Fmoc-Ser-OBn
a a:b Ratios were determined by 1H NMR of the crude material.
domains is under continuing investigation. Thus far, at
least, there is no indication that constructs featuring
non-natural linkages are disadvantageous with respect
to immunogenicity. Indeed, in principle such unnatural
linkages might result in an enhanced immune response.
anomeric selectivity was not as good as what was seen
with serine (1:0, entry 6), the results were significantly
better than 4-penten-1-ol (1:1, entry 1). Additionally,
the undesired b-anomer 5 could be removed by silica
gel chromatography. This successful diastereoselective
glycosylation allowed for the efficient transformation of
4 into three different glycosylamino acids, which con-
tain the tumor-associated carbohydrate antigens Tn,
TF and STn (Scheme 2).
Introduction of non-natural linkages utilizes well-estab-
lished principles of glycosylation. Thus, a-1,2 oxido
donors (glycal epoxides) result in primarily b-linked
products as seen in our syntheses of Globo-H and
Lewisy (Ley) pentenyl glycosides.2 However, in the case
of the glycophorins, the a-anomer is desired. The use of
a non-participating azido group at C-2 allows for a-
selectivity with serine or threonine acceptors.3a How-
ever, relatively no selectivity was realized with
4-penten-1-ol as the glycosyl acceptor.4 The mixture of
pentenyl anomers was inseparable and although the
undesired b-anomer could eventually be removed, it
was realized that an alternative and perhaps
diastereoselective method for introduction of the
desired non-natural amino acid functionality would be
useful. Building on the successful results of using serine
and threonine for a-O-linked glycosylamino acids we
chose to investigate the viability of hydroxynorleucine
as an acceptor.
We then explored the utility of a suitably protected
hydroxynorleucine as an acceptor in the glycosylations
of more complex carbohydrates. Four different carbo-
hydrate donors were chosen: Gb3 glycal 6a, Globo-H
hexasaccharide glycal 6b, ABC trisaccharide glycal
fragment of Globo-H 6c, and Ley pentasaccharide gly-
cal 6d (Table 2). The amino acid functionality was
Boc-
sized from
L
-hydroxynorleucine tert-butyl ester 2 was synthe-
-lysine according to literature procedures.7
L
Protecting group manipulations (Scheme 1) provided
the desired suitably protected hydroxynorleucine 1 for
use in the glycosylation reaction.
Reaction of trichloroacetimidate donor 3,3a derived
from tri-O-acetyl-D-galactose, with hydroxy norleucine
acceptor 1 (1.5 equiv.) in the presence of TMSOTf in
THF provided a 2:1 mixture of anomeric products with
the desired a-anomer 4 as the major product (Table 1,
entry 2). Decreasing the temperature increased the yield
and selectivity (88%, 4:1, entry 4) while changing the
solvent to dichloromethane caused a reversal with the
b-anomer 5 predominating (1:9, entry 5). Although the
Scheme 2. Glycosylamino acid 4 was used in the synthesis Tn,
TF and STn glycosylamino acids.

S. J. Keding et al. / Tetrahedron Letters 44 (2003) 3413–3416
3415
Table 2. b-selective glycosylations of complex glycals with acceptor 1
introduced in two different ways. The first method
involved four steps in which the glycal was oxidized
using dimethyldioxirane to the glycal epoxide8 and sub-
sequently opened with ethanethiol to provide the ethyl
sulfide. The C-2 hydroxyl was protected as a benzoate
ester, which acted as a participating group during gly-
cosylations with hydroxynorleucine acceptor 1. This
reaction sequence is very reliable and generates the
desired b-linkage in good yield for Gb3 glycal (26%,
entry 1) and Globo-H (20%, entry 2). Surprisingly, we
also found the glycosyl acceptor could be introduced
selectively in a simpler, two-step procedure. In this case
the glycosyl acceptor was used to open the glycal
epoxide directly. The yields are better for the two-step
compared with the four-step procedure with generation
of the necessary b-linkage. In both methods an excess
of the amino acid acceptor was used, typically 2–10
equiv. Using the shorter synthetic sequence we were
able to make the differentially protected ABC unit of
Globo-H (70%, entry 3) and Ley glycosylamino acids
(65%, entry 4) in high yield of the b-linked anomer.
In conclusion, we describe the use of a new glycosyl
acceptor, hydroxynorleucine, for the incorporation of
amino acid functionality in the context of complex
carbohydrates. This procedure directly affords the N-

3416
S. J. Keding et al. / Tetrahedron Letters 44 (2003) 3413–3416
and C-protected glycosylamino acid for further func-
tionalization of the carbohydrate domain or for depro-
tection and use in the preparation of glycopeptides. We
have synthesized hydroxynorleucine acceptor 1 in large
quantities and successfully generated sufficient amounts
of many glycosylamino acids for incorporation into
vaccine constructs for immunological investigation.
Chem. Soc. 1998, 120, 7760; (b) Kuduk, S. D.; Schwarz, J.
B.; Chen, X.-T.; Glunz, P. W.; Sames, D.; Ragupathi, G.;
Livingston, P. O.; Danishefsky, S. J. J. Am. Chem. Soc.
1998, 120, 12474; (c) Schwarz, J. B.; Kuduk, S. D.; Chen,
X.-T.; Sames, D.; Glunz, P. W.; Danishefsky, S. J. J. Am.
Chem. Soc. 1999, 121, 2662; (d) Glunz, P. W.; Hinter-
mann, S.; Schwarz, J. B.; Kuduk, S. D.; Chen, X.-T.;
Williams, L. J.; Sames, D.; Danishefsky, S. J.;
Kudryashov, V.; Lloyd, K. O. J. Am. Chem. Soc. 1999,
121, 10636.
Acknowledgements
4. Allen, J. R.; Harris, C. R.; Danishefsky, S. J. J. Am.
Chem. Soc. 2001, 123, 1890.
5. Biswas, K.; Coltart, D. M.; Danishefsky, S. J. Tetrahedron
This work was supported by the National Institutes of
Health (CA-28824). Postdoctoral fellowship support is
gratefully acknowledged by S.J.K. (The Texaco Foun-
dation) and D.M.C. (Natural Science and Engineering
Research Council of Canada: PDF-230654-2000 and
Alberta Heritage Foundation for Medical Research:
199901330).
Lett. 2002, 43, 6107.
6. Ragupathi, G.; Coltart, D. M.; Williams, L. J.; Koide, F.;
Kagan, E.; Allen, J.; Harris, C.; Glunz, P. W.; Livingston,
P. O.; Danishefsky, S. J. Proc. Natl. Acad. Sci. USA 2002,
99, 13699.
7. (a) Nevill, C. R.; Angell, P. T. Tetrahedron Lett. 1998, 39,
5671; (b) Yoshifuji, S.; Tanaka, K.; Nitta, Y. Chem.
Pharm. Bull. 1987, 35, 2994.
8. The epoxidation of the peracetylated glycals resulted not
only in the desired a-epoxides, but also b-epoxides (5–
29%) that gave b-mannoside products upon epoxide open-
ing. Upon changing the hydroxyl protecting groups from
acetates to benzyl ethers, the DMDO epoxidation resulted
in exclusively a-epoxide formation.
References
1. Review: Hakomori, S.; Zhang, Y. Chem. Biol. 1997, 4, 97.
2. Review: Danishefsky, S. J.; Allen, J. R. Angew. Chem., Int.
Ed. Engl. 2000, 39, 836.
3. (a) Chen, X.-T.; Sames, D.; Danishefsky, S. J. J. Am.