Convergent synthesis of a fully phosphorylated GPI anchor of the CD52 antigen

Article abstract of DOI:10.1021/ol701870m

A fully phosphorylated GPI anchor (1) of the CD52 antigen was synthesized by a highly convergent strategy. After a trimannose and a phospholipidated pseudodisaccharkte were prepared separately, they were coupled together to form the GPI core, which was th

Full text of DOI:10.1021/ol701870m

ORGANIC
LETTERS
2007
Vol. 9, No. 21
4311-4313
Convergent Synthesis of a Fully
Phosphorylated GPI Anchor of the CD52
Antigen
Xiaoming Wu and Zhongwu Guo*
Department of Chemistry, Wayne State UniVersity, 5101 Cass AVenue,
Detroit, Michigan 48202
zwguo@chem.wayne.edu
Received August 2, 2007
ABSTRACT
A fully phosphorylated GPI anchor (1) of the CD52 antigen was synthesized by a highly convergent strategy. After a trimannose and a
phospholipidated pseudodisaccharide were prepared separately, they were coupled together to form the GPI core, which was then phosphorylated
to introduce two phosphoethanolamine moieties in one step to afford CD52 GPI in its fully protected form. Finally, global deprotection of the
product resulted in 1.
Glycosylphosphatidylinositol (GPI) anchors are a class of
natural glycolipids which are expressed by all eukaryotic
cells.1 These compounds have many important biological
functions, of which anchoring proteins and glycoproteins to
cell membranes is the most obvious.1-5 GPIs share a
remarkably conserved core structure, which is composed of
a tetrasaccharide, a phosphoethanolamine group and an
inositol residue at the nonreducing and reducing ends of the
glycan, respectively, and a phosphatidyl moiety that is
attached to the inositol ring. Owing to the interesting and
complex structures and important biological functions of
GPIs and related molecules, their chemical synthesis has
attracted great attention in the past decade.6,7
Recently, we became especially interested in the CD52
antigen, a GPI-anchored glycopeptide antigen involved in
the human reproduction and human immune recognition
processes.8-11 Because of its simple structure and evident
bioactivity, the CD52 antigen can be a useful model for
studying the functions of GPI anchors. In this regard,
homogeneous GPIs and GPI-anchored molecules, which are
difficult to obtain from nature, are critical.
This paper reports the chemical synthesis of a fully
phosphoryated GPI anchor 1 of CD52 (Scheme 1) having a
long acyl chain attached to the inositol 2-O-position.
As shown in Scheme 1, our overall synthetic plan was to
first assemble the phospholipidated core (2) and then
introduce the two phosphoethanolamine moieties through
two-step phosphorylation. Because the R-glycosylation of
mannose is relatively easy, but not that of glucosamine, a
logical and convergent design for the assembly of the GPI
(1) Ferguson, M. A. J.; Williams, A. F. Ann. ReV. Biochem. 1988, 57,
285-320.
(2) Hwa, K. AdV. Exp. Med. Biol. 2001, 491, 207-214.
(3) Eckert, V.; Gerold, P.; Schwarz, R. T. In Glycosciences; Gabius, H.-
J., Gabius, S., Eds.; Chapman & Hall: Weinheim, Germany, 1997; pp 223-
243.
(4) Cole, R. N.; Hart, G. W. New Compr. Biochem. 1997, 29, 69-88.
(5) Thomas, J. R.; Dwek, R. A.; Rademacher, T. W. Biochemistry 1990,
29, 5413-5422.
(6) See Review: Gigg, R.; Gigg, J. In Glycopeptides and Related
Compounds: synthesis, analysis, and applications; Large, D. G., Warren,
C. D., Eds.; Marcel Dekker, Inc.: New York, 1997; pp 327-392.
(7) See Review: Guo, Z.; Bishop, L. Eur. J. Org. Chem. 2004, 3585-
3596 and references cited therein.
(8) Tsuji, Y.; Clausen, H.; Nudelman, E.; Kaizu, T.; Hakomori, S.-I.;
Isojima, S. J. Exp. Med. 1988, 168, 343-356.
(9) Hale, G.; Xia, M.-Q.; Tighe, H. P.; Dyer, J. S.; Waldmann, H. Tissue
Antigens 1990, 35, 118-127.
(10) Treumann, A.; Lifely, M. R.; Schneider, P.; Ferguson, M. A. J. J.
Biol. Chem. 1995, 270, 6088-6099.
(11) Schroter, S.; Derr, P.; Conradt, H. S.; Nimtz, M.; Hale, G.; Kirchhoff,
C. J. Biol. Chem. 1999, 274, 29862-29873.
10.1021/ol701870m CCC: $37.00
© 2007 American Chemical Society
Published on Web 09/18/2007

an unsaturated palmitoleoyl chain was introduced because,
before global deprotection, its CdC bond could be modified
to produce some reactive intermediates, e.g., oxidative
cleavage of the double bond to form carbonyl compounds,14
which are useful for the preparation of various GPI conju-
gates. On the other hand, during the reductive debenzylation
at the final step, the palmitoleoyl group could be readily
reduced to form a palmitoyl groupsthe natural acyl chain
of the CD52 GPI anchor. Based on the lessons that we had
learned from the synthesis of another CD52 GPI anchor,15
we decided to introduce the bulky phospholipid moiety to
the inositol 1-O-position at this stage, because the presence
of a large acyl group at the inositol 2-O-position might render
late-stage phospholipidation impossible. Oxidative removal
of the 4-methoxybenzyl (PMB) group on the inositol ring
was achieved with ceric ammonium nitrate (CAN), which
was followed by phospholipidation employing 5 in a two-
step, one-pot procedure to give a diastereomeric mixture
Scheme 1. Retrosynthesis of Target GPI Anchor 1
1
(1:1.1) of 13 that was fully characterized by H NMR, 31P
NMR, and MS. Finally, the tert-butyldimethylsilyl (TBS)
group in 13 was removed to afford 14, which was ready to
be coupled with the trimannose fragment 3 for the assembly
of the final synthetic target.
core glycan would be to prepare fragments 3 and 4 first and
then couple them together through a relatively easy glyco-
sylation reaction.
Trimannose 15 was prepared according to a procedure
developed for similar structures,16 but the 2-O-position of
mannose-I was protected by a PMB group. Glycosylation
of 14 using 15 as glycosyl donor and N-iodosuccinamide
and silver triflate (NIS/AgOTf) as promoter would affect the
double bond in the palmitoleoyl group, whereas the reaction
of 14 and 15 with dimethyl(methylthio)sulfonium triflate
(DMTST) or methyl triflate (MeOTf) as promoter could not
yield the desired product either. Therefore, 15 was trans-
formed into 3 following oxidative hydrolysis of the reducing
end by reaction with NIS/TfOH and water and then trichlo-
roacetimidation of resultant hemiacetal. Glycosylation of 14
using a large excess of 3 (3 equiv) with trimethylsilyl triflate
(TMSOTf) as promoter went well, but the isolation of 16
was complicated by the hydrolytic products derived from 3.
Therefore, after 16 was purified briefly on a silica gel
column, it was directly subjected to treatment by BF3‚Et2O
in CH2Cl2 to remove the TBS and PMB groups and give
17, of which the two diastereomers could be separated
through careful silica gel column chromatography. Both
The synthesis of pseudodisaccharide 4, as well as its
phospholipidated derivative 14, is outlined in Scheme 2. First,
an optically pure inositol derivative 6 was prepared according
to a reported method.12 Next, an acetyl group was introduced
to temporarily protect the inositol 2-OH, followed by
deallylation to afford 8, which was ready for installation of
the glucosamine moiety by glycosylation. To establish
reliable conditions to achieve this difficult R-linkage, we have
prepared and tested several glucosamine derivatives as
potential glycosyl donors, such as glycosyl halides, glycosyl
trichloroacetimidate, and thioglycosides of 2-azido-2-deoxy-
glucose. We found that glycosylation reactions under Le-
mieux conditions,13 i.e., using glucosyl bromide 9 as the
glycosyl donor and tetrabutylammonium bromide (TBAB)
as the promoter, gave the R-glycoside 10 stereospecifically
in a good yield (56%). The inositol 2-O-position in 10 was
then deprotected to expose a free hydroxyl group, to which
was introduced a palmitoleoyl group employing dicyclo-
hexylcarbodiimide (DCC) as the condensation reagent. Here,
Scheme 2. Synthesis of the Phospholipidated Pseudodisaccharide Segment
4312
Org. Lett., Vol. 9, No. 21, 2007

Scheme 3. Final Assembly of the Target GPI Anchor 1
stereoisomers were individually characterized by 1H and 31
P
NMR and MS. Then, one isomer, 17b, was phosphorylated
employing 18 by the two-step, one-pot procedure described
above to introduce two phosphoethanolamine groups simul-
taneously, giving fully protected GPI anchor 19. Global
deprotection of 19 was realized in two steps. First, the
2-cyanoethyl group protecting the phospholipid moiety was
removed by DBU treatment for a short period (ca. 5 min) to
give 20, and the product was fully characterized by NMR
and MS. The NMR signals of the anomeric protons and
carbons (500 and 125 MHz, respectively, in CDCl3), which
were clearly identifiable in the HMQC spectrum, suggested
R-configurations of all glycosidic bonds. Finally, the benzyl
and benzyloxycarbonyl (CBz) groups were removed by
hydrogenolysis under an H2 atmosphere using 10% Pd/C as
the catalyst to eventually afford CD52 GPI 1. The 1H NMR
spectra (500 MHz) of 1 obtained in a mixture of CD3OD,
which shows the highest m/z peak at 893 [M + 2H+]. The
MS result further indicates that 1 was obtained as a sodium
salt and that its free amino groups were easily protonated to
give rise to the observed m/z peak at the positive ionization
mode of MALDI MS.
In summary, a highly convergent synthesis of the fully
functionalized GPI anchor (1) of the CD52 antigen has been
described in this paper. Our collaborator, Professor Xiangqun
Zeng at Oakland University, has shown that the synthetic
GPI anchor 1 could interact strongly with GPI-binding
proteins. Our two laboratories are currently studying the
biological activities of 1 in detail, and the results will be
reported in due course.
Acknowledgment. This work was supported by NSF
(CHE-0554777). X.M.W. thanks Dr. Jie Xue of WSU for
many helpful discussions. We also appreciate Dr. B. Shay
and Dr. L. Hryhorczuk of WSU for the MS measurements
and Dr. B. Ksebati of WSU for his help with some NMR
measurements.
1
CDCl3, and D2O (3:3:1) showed no aromatic H signals,
suggesting a complete removal of all benzyl and CBz groups.
This conclusion, as well as the structure of the final product,
was confirmed by 13C NMR (125 MHz) and by MS spectrum
Supporting Information Available: Experimental and
the NMR and MS spectra of 1, 4, 6-14, 17, 19, and 20.
This material is available free of charge via the Internet at
http://pubs.acs.org.
(12) Xue, J.; Guo, Z. Bioorg. Med. Chem. Lett. 2002, 12, 2015-2018.
(13) Lemieux, R. U.; Hendriks, K. B.; Stick, R. V.; James, K. J. Am.
Chem. Soc. 1975, 97, 4056-4062.
(14) Xue, J.; Pan, Y.; Guo, Z. Tetrahedron Lett. 2002, 43, 1599-1602.
(15) Xue, J.; Guo, Z. J. Am. Chem. Soc. 2003, 125, 16334-16339.
(16) Xue, J.; Shao, N.; Guo, Z. J. Org. Chem. 2003, 68, 4020-4029.
OL701870M
Org. Lett., Vol. 9, No. 21, 2007
4313