Convergent synthesis of a fully lipidated glycosylphosphatidylinositol anchor of Plasmodium falciparum

Article abstract of DOI:10.1021/ja042374o

A highly convergent strategy for the synthesis of fully lipidated GPI anchors of malarial origin is reported. This strategy utilized three orthogonal protecting groups, which can be chemoselectively deprotected and functionalized in the late stage of the

Full text of DOI:10.1021/ja042374o

Published on Web 03/17/2005
Convergent Synthesis of a Fully Lipidated Glycosylphosphatidylinositol
Anchor of Plasmodium falciparum
Xinyu Liu, Yong-Uk Kwon, and Peter H. Seeberger*
Laboratory for Organic Chemistry, ETH Zu¨rich, Wolfgang-Pauli-Str. 10, HCI F315, 8093 Zu¨rich, Switzerland
Received December 19, 2004; E-mail: seeberger@org.chem.ethz.ch
Malaria is a devastating parasitic disease that threatens 40% of
the world’s population and claims more than two million lives each
year, mostly young children in developing countries.1 Using
chemical synthesis, we identified the glycan part of the malaria
toxin, a cell surface glycosylphosphatidylinositol (GPI) from P.
falciparum that is responsible for mortality by malaria2 and
demonstrated that a synthetic glycan can serve as an effective
antitoxin vaccine in a rodent model.3 The glycan alone, while
sufficient for vaccine development, is not toxic. To elucidate the
role of GPIs in malaria pathology and signal transduction and to
investigate GPI biosynthesis, access to the pure, naturally occurring,
lipidated malarial GPI will be of paramount importance.
GPIs are among the most complex classes of natural products
Figure 1. Consensus structure and retrosynthetic analysis of lipidated P.
as they combine lipids, carbohydrates, and peptides. The highly
complex nature of GPIs renders their isolation a daunting task.
Diligent work has yielded a proposed consensus structure (Figure
1) although the exact nature of the lipid portion remains elusive.4
To determine the exact structure of the malarial GPI responsible
for the toxic effect and establish a structure-function relationship,
the synthesis of GPI glycans containing a variety of lipids will be
required.5
falciparum GPIs.
Scheme 1. Preparation of Glucosamine-Inositol Disaccharide 7a
Here, we report the first total synthesis of fully lipidated malarial
GPI 1 using a convergent yet versatile synthetic approach.6 Three
variable sites exist on the GPI glycan backbone; inositol is lipidated
and serves as attachment site for phospholipids, and the C6 hydroxyl
group of the penultimate mannose contains a phosphate ethanol-
amine group. Mindful of the structural diversity points, the hallmark
of our synthetic approach is the convergent assembly of a glycan
containing three differentially protected hydroxyl groups in antici-
pation of late stage acylation and phosphorylation (Figure 1). Thus,
rapid access to various GPIs with different lipid moieties can be
achieved.
GPI 1 will be derived by [4 + 2] coupling of a tetramannoside
and a glucosamine-inositol pseudodisaccharide. On the basis of
our experience with the synthesis of GPI glycans,7 triispropylsilyl
ether and allyl ether groups were chosen to mark the two
phosphorylation sites. Selection of the protecting group for the C2
hydroxyl group of inositol revealed the complexities associated with
the synthesis of highly complex glycans; a DDQ-labile p-(3,4-
dimethoxyphenyl)benzyl ether that had served well in the synthesis
of a disaccharide model could not be selectively removed when
operating on a hexasaccharide.8 Extensive trials established the
p-methoxybenzyl (PMB) group as the most reliable mode of
protection for the inositol C2 hydroxyl in the context of complex
molecule construction.
a Reagents and yields: (a) PMBCl, NaH, TBAI, DMF, 68%; (b) i. 4,
TMSOTf, CH2Cl2, -30 °C, ii. NaOMe/MeOH, 89% (R:â )4: 1); (c) i.
PhCH(OMe)2, CSA, CH3CN, ii. BnBr, NaH, DMF, 74%; (d) NaCNBH3,
HCl/Et2O, THF, 43% (R-isomer).
trichloroacetimidate 4 to furnish preferentially the desired R-linked
disaccharide (R:â ) 4:1). The three acetate groups on the
glucosamine moiety, crucial in directing the stereochemistry of the
glycosylation reaction, were removed to give disaccharide 5 in 89%
yield (two steps). Installation of the appropriate protecting group
pattern by formation of the 4,6-O-benzylidene group and benzyla-
tion of the C3 hydroxyl was completed by the regioselective opening
of the 4,6-O-benzylidene to afford the differentially protected
pseudodisaccharide 7. At this stage, the R- and â-isomers can be
readily separated by silica gel column chromatography.
The union of disaccharide 7 and tetramannoside 8,7 readily
available in multigram quantities, furnished hexasaccharide 9 in
excellent yield under complete stereocontrol by virtue of a
neighboring benzoyl ester protecting group. Having served their
purpose, the ester groups in 9 were subsequently replaced by benzyl
ethers to afford nearly 1 g of hexasaccharide 11. Hexasaccharide
11 is the key for the preparation of fully lipidated GPI anchors as
it contains three orthogonal protecting groups for further elaboration.
Decoration of glycan backbone 11 with lipids, phospholipids,
and phosphate ethanolamine can be performed in varying order.
Careful evaluation of the order of functionalization regarding
transformation efficiency and ease of purification revealed that
removal of the PMB group and lipidation, followed by deallylation
Following the strategic decisions, the assembly of GPI 1
commenced with the stereoselective synthesis of the R-linked
glucosamine-inositol pseudodisaccharide. Regioselective alkylation
with 4-methoxybenzyl chloride resulted in the differentiation of
the two hydroxyl groups in inositol 2 to afford 3. The notoriously
difficult glycosylation of 3 was carried out with glucosamine
9
5004
J. AM. CHEM. SOC. 2005, 127, 5004-5005
10.1021/ja042374o CCC: $30.25 © 2005 American Chemical Society

C O M M U N I C A T I O N S
Scheme 2. Assembly of the Hexasaccharide Backbonea
phosphoramidite reagents.5f Removal of the silyl ether was most
efficiently achieved by exposure to scandium (III) triflate in the
presence of traces of water. Finally, the phosphate ethanolamine
was introduced by coupling H-phosphonate 12 and glycolipid 18
and in situ oxidation. Thus, fully protected, lipidated GPI hexa-
saccharide 19 was obtained in pure form as the bistriethylammo-
nium salt in 94% yield.9
After establishing the structural integrity of 19, hydrogenolysis
using Pearlmann’s catalyst in a mixture of solvents required 32 h
at ambient temperature to ensure the complete removal of all of
the benzyl ether groups. After lyophilization, lipidated GPI 1 was
harvested as an amorphous white solid in excellent yield. Final
product 1 is very poorly soluble in organic solvents other than
DMSO, but methods to ascertain identity and purity of the product
were established.9
In conclusion, we have developed a highly convergent synthetic
strategy to access fully functionalized GPI anchors, as demonstrated
by the first total synthesis of the P. falciparum GPI. Strategic
placement of three orthogonal protecting groups for the late stage
installation of three side chains on the GPI hexasaccharide backbone
was key to the success of the synthesis. The synthetic GPI 1 is
currently used as a molecular probe for the study of malarial
pathogenesis and aspects of fundamental immunology. The strategy
reported here is the basis for the preparation of various GPIs with
different lipid side chains for biological applications.
a Reagents and conditions: (a) 7, TMSOTf, CH2Cl2, -40 °C, 94%; (b)
NaOMe/MeOH, 50 °C, 77%; (c) NaH, BnBr, DMF, 91%.
Scheme 3. Chemoselective Functionalization of Hexasaccharide
11 and Synthesis of GPI Anchor 1a
Acknowledgment. This research was supported by ETH Zu¨rich
and Korea Science and Engineering Foundation (postdoctoral
fellowship for Y.U.K.). We thank Simone Bufali and Christian Noti
for acquiring high field NMR spectra.
Supporting Information Available: Experimental procedures and
spectral copies (1H, 13C, 31P, HSQC, MS) of all new compounds and
full citation for ref 4b. This material is available free of charge via the
Internet at http://pubs.acs.org.
References
a Reagents and conditions: (a) CAN, CH3CN/PhMe/H2O, 76%; (b)
C15H31COOH, DCC, DMAP, CH2Cl2, 98%; (c) PdCl2, NaOAc, AcOH/
H2O, 52%; (d) i. 13, PivCl, pyridine, ii. I2, Py/H2O, 72%; (e) Sc(OTf)3,
CH3CN, H2O, 40 °C, 77%; (f) i. 12, PivCl, pyridine, ii. I2, pyridine/H2O,
94%; (g) H2, Pd(OH)2/C, CHCl3/MeOH/H2O, 94%.
(1) State of the art of new vaccines (research and development) World
Health Organization, http://www.who.int/vaccine_research/documents/
new_vaccines/en/.
(2) Schofield, L.; Vivas, L.; Hackett, F.; Gerold, P.; Schwarz, R. T.; Tachado,
S. Annu. Trop. Med. Parasitol. 1993, 87, 617-626.
(3) Schofield, L.; Hewitt, M. C.; Evans, K.; Slomos, M.-A.; Seeberger, P. H.
Nature 2002, 418, 785-788.
(4) (a) Gerold, P.; Schofield, L.; Blackan, M. J.; Holder, A. A.; Schwarz, R.
T. Mol. Biochem. Parasitol. 1996, 75, 131. (b) Gowda, D. C. et al. J.
Exp. Med. 2000, 192, 1563-1575.
and introduction of the phospholipid prior to the removal of the
silyl ether followed by installation of the phosphate ethanolamine
was best. The PMB ether was cleaved using cerium ammonium
nitrate (CAN) to afford hexasaccharide 14 in 76% yield.9 The C2
hydroxyl group of inositol was esterified with palmitic acid by DCC
activation to furnish 15. The following deallylation proved to be
more challenging, and careful control of the reaction time was
crucial to the success of this transformation. Treatment with a large
excess of PdCl2 in acetate buffer furnished 16 in moderate yield,10
but it completely suppressed acyl migration. Installation of the
phospholipid chain was achieved by phosphorylation using H-
phosphonate 13. Coupling of 13 by activation with pivaloyl chloride
was followed by in situ oxidation to furnish hexasaccharide 17 in
72% yield. This mode of phosphorylation was compatible with the
azide group and avoided the problems encountered by others using
(5) (a) Murakata, C.; Ogawa, T. Carbohydr. Res. 1992, 235, 95-114. (b)
Mayer, T. G.; Kratzer, B.; Schmidt, R. R. Angew. Chem., Int. Ed. Engl.
1994, 33, 2177-2181. (c) Campbell, A. S.; Fraser-Reid, B. J. Am. Chem.
Soc. 1995, 117, 10387-10388. (d) Baeschlin, D. K.; Chaperon, A. R.;
Green, L. G.; Hahn, M. G.; Ince, S. J.; Ley, S. V. Chem.sEur. J. 2000,
6, 172-186. (e) Pekari, K.; Schmidt, R. R. J. Org. Chem. 2003, 68, 1295-
1308. (f) Xue, J.; Guo, Z. J. Am. Chem. Soc. 2003, 125, 16334-16339.
(6) During the course of our studies, the synthesis of a malaria GPI model
lacking one mannose and containing shortened lipids was reported: Lu,
J.; Jayaprakash, K. N.; Schlueter, U.; Fraser-Reid, B. J. Am. Chem. Soc.
2004, 126, 7540-7547.
(7) Seeberger, P. H.; Soucy, R. R.; Kwon, Y. U.; Snyder, D. A.; Kanemitsu,
T. Chem. Commun. 2004, 1706-1707.
(8) Liu, X.; Seeberger, P. H. Chem. Commun. 2004, 1708-1709.
(9) For details, see Supporting Information.
(10) This result is in contrast to the report by: Lu, J.; Jayaprakash, K. N.;
Fraser-Reid, B. Tetrahedron Lett. 2004, 45, 879-882.
JA042374O
9
J. AM. CHEM. SOC. VOL. 127, NO. 14, 2005 5005