A general method for synthesis of GPI anchors illustrated by the total synthesis of the low-molecular-weight antigen from Toxoplasma gondii

Article abstract of DOI:10.1002/anie.201103483

Building blocks: A new, general synthetic strategy, which allows the construction of branched glycosylphosphatidylinositols (GPIs), enables the synthesis of parasitic glycolipid 1 from Toxoplasma gondii. In addition, the structure is further confirmed by

Full text of DOI:10.1002/anie.201103483

DOI: 10.1002/anie.201103483
Synthesis of GPI Anchors
A General Method for Synthesis of GPI Anchors Illustrated by the
Total Synthesis of the Low-Molecular-Weight Antigen from
Toxoplasma gondii**
Yu-Hsuan Tsai, Sebastian Gçtze, Nahid Azzouz, Heung Sik Hahm, Peter H. Seeberger,* and
Daniel Varon Silva*
Glycosylphosphatidylinositol (GPI) anchors are a class of
naturally occurring glycolipids with a conserved core struc-
ture:
H2N(CH2)2OPO3H-6Mana1!2Mana1!6Mana1!
4GlcNH2a1!6myo-Ino1-OPO3H-Lipid
(Man = mannose;
GlcNH2 = glucosamine; Ino = inositol; Scheme 1).[1] GPI
anchors are present in all eukaryotic cells and their structures
vary in a species dependent manner. Possible modifications of
the core structure include: an extra mannose at the non-
reducing end of the linear pseudopentasaccharide, branching
at the 3 or 4 position of the mannose I, and additional
phosphorylations. There are many known variations of the
phospholipid moiety of GPI anchors. Additionally, a fatty
acid ester may be present at the 2-position of myo-inositol.[2]
Functions of GPI anchors in higher eukaryotic cells stem
from their ability to anchor attached proteins to cell
membranes and associate them with micro domains, which
enables specific interactions with other membrane proteins.[3]
This GPI-induced localization is essential for certain cell–cell
interactions and signal-transduction processes.[4] Some proto-
zoan parasites, such as Plasmodium falciparum or Toxo-
plasma gondii, use free GPI anchors and GPI-anchored
proteins to modulate the immune system of their host.[5] For
example, GPI 1 (Scheme 2b), also known as the low-
molecular-weight antigen, which activates the human
immune system during the T. gondii infections, is known to
Scheme 1. The conserved core structure of GPI anchors and possible
modifications.
be immunogenic and causes antibody production in patients
who suffer from toxoplasmosis, although its specific role in
these processes is not fully understood yet.[6]
In most cases, the function of GPI anchors is unknown
beyond protein localization on the outer leaflet of the cell
membrane. This lack of knowledge is mainly due to the low
availability of pure GPI samples arising from their amphi-
philic and heterogeneous character, complicating purification
from natural sources. Consequently, detailed structure–activ-
ity relationship studies for GPI anchors are not available. The
only access to sufficient quantities of structurally defined GPI
anchors is by way of chemical synthesis.[7] Although many
GPI anchors have been prepared using various synthetic
methods and protecting-group strategies, a general unifying
route that will enable efficient access to a wide range of GPI
anchors has not yet been developed.[8]
Herein, we present a general synthetic route to branched
GPI structures (Scheme 2a). Using the interchangeable
building blocks 5–7 it is in principle possible to prepare a
library of orthogonally protected glycans 2 in a few synthetic
steps. The protecting-group pattern in the common inter-
mediate 2 enables the synthesis of different GPI anchors and
derivatives suitable for biological assays and physical mea-
surements. Structures with an additional mannose, palmitoy-
lated inositol, and different phosphorylation patterns should
be easily accessible using this strategy.
[*] Y.-H. Tsai, S. Gçtze, Dr. N. Azzouz, H. S. Hahm,
Prof. Dr. P. H. Seeberger, Dr. D. Varon Silva
Department of Biomolecular Systems
Max Planck Institut of Colloids and Interfaces
Am Mꢀhlenberg 1, 14424 Potsdam (Germany)
and
Institute of Chemistry and Biochemistry, Free University of Berlin
Arnimallee 22, 14195 Berlin (Germany)
E-mail: seeberger@mpikg.mpg.de
daniel.varon@mpikg.mpg.de
Homepage: http://www.mpikg.mpg.de/Biomolekulare_Systeme
[**] This work was supported by the Max Planck Society, the Kçrber
foundation (Kçrber European Science Award to P.H.S.) and the
German Academic Exchange Service (doctoral scholarship to Y.-
H.T.). We are grateful for the help of Dr. Ivan Vilotijevic, Xiaoqiang
Guo, and Martha Kozakowski. We thank Prof. Dr. Xinyu Liu
(University of Pittsburgh (USA)) for the discussions, Dr. Jean
Francois Dubremetz (University of Montpellier, France) for the
monoclonal antibodies mAbT54E10 and mAbT33F12, GIANT-
microbes for the T. gondii picture, and Dr. Alexander O’Brien and
Dr. Dominea Rathwell for proofreading this manuscript.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201103483.
Angew. Chem. Int. Ed. 2011, 50, 9961 –9964
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
9961

Communications
Scheme 2. a) General retrosynthetic analysis for GPI anchors. Orthogonal protecting groups for late-stage modifications are highlighted. b) Retrosynthetic
analysis of GPI 1. Roman numerals indicate the assembly sequence. Ac=acetyl; Bn=benzyl; Bu=butyl; Cbz=benzyloxycarbonyl; Lev=levulinyl;
Me=methyl; Nap=2-naphthylmethyl; Ph=phenyl; PMB=para-methoxybenzyl; TBDPS=tert-butyldiphenylsilyl; TIPS=triisopropylsilyl.
To demonstrate the feasibility of this approach we have
prepared the low-molecular-weight antigen 1 of T. gondii
(Scheme 2b) and compared this material to isolated GPI
anchors of T. gondii in binding assays with known monoclonal
antibodies.[9]
The central mannoside 5a is the key building block in our
retrosynthetic analysis of the GPI 1. It enables the assembly
of a trisaccharide subunit (by glycosylations I and II, Sche-
Scheme 3. Synthesis of the key mannoside 5a. Reaction conditions:
a) NapBr, NaH, DMF, 99%; b) 80% AcOH(aq), 658C, 72%; c) 1.
me 2b) which is then elongated with dimmanosyl imidate 6a
and pseudodisaccharide 7a[8c] in a [2+3+2] glycosylation
sequence to yield a heptasaccharide. Selective phosphoryla-
tions with H-phosphonates 3[8f] and 4,[8b] and subsequent
global deprotection affords GPI 1. Initial attempts to directly
glycosylate the pentasaccharide core with a variety of N-
acetylgalactosyl building blocks were unsuccessful, highlight-
ing the importance of the appropriate glycosylation sequence.
The synthesis of building block 5a commenced with the
conversion of d-mannose into the acetal 10[10] The 4-O
hydroxy group of 10 was protected as a 2-naphthylmethyl
ether[11] (Scheme 3). After removal of the isopropylidene
group, diol 12 was selectively protected as a benzyl ether at
the 3-position via a stannylene acetal. Levulinate ester
protection of the remaining free hydroxy group[12] followed
by deprotection of the 2-naphthylmethyl ether afforded
orthogonally protected mannoside 5a.[13] This flexible strat-
egy for the synthesis of 5a also allowed for the preparation of
Bu2SnO, PhMe, reflux; 2. BnBr, TBAB, 99%; d) LevOH, DIC, DMAP,
CH2Cl2, 82%; e) DDQ, H2O, CH2Cl2, 96%. DDQ=2,3-dichloro-5,6-
dicyanobenzoquinone; DIC=N,N’-diisopropylcarbodiimide;
DMAP=4-(dimethylamino)pyridine; DMF=dimethylformamide;
TBAB=tetrabutylammonium bromide; THF=tetrahydrofuran.
compound 5b by switching the protecting groups at the 3- and
4-positions. Building block 5b enables the synthesis of GPI
anchors of Trypanosoma brucei, the only GPI structures with
a branch at the 3-position of mannose I.[14]
Synthesis of the orthogonally protected glycan 22 started
with a one-pot sequence involving glycosylation of mannose
5a with galactosyl phosphate 8 followed by oxidative
deprotection of the 2-naphthylmethyl ether with DDQ to
provide disaccharide 15 in quantitative yield (Scheme 4).
Installation of the a-glucoside using glycosyl phosphate 9[15] in
a mixture of thiophene and toluene as solvent gave trisac-
9962 www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 9961 –9964

Scheme 4. Synthesis of orthogonally protected glycan 22. Reaction conditions: a) 1. TMSOTf, CH2Cl2, ꢀ408C; 2. PBS-buffer, DDQ, quant.; b) TBSOTf,
thiophene/toluene (2:1), 08C, 82% (a:b=6:1); c) HF-pyridine, THF, 82%; d) TBSOTf, 4 ꢁ MS, thiophene/toluene (1:1), 69%; e) Zn, AcOH, 558C, 94%;
f) 1. [Ir(cod)(PPh2Me)2]PF6, H2, THF, 2. HgCl2, HgO, H2O, acetone; g) Cl3CCN, DBU, CH2Cl2, 55% over two steps; h) TMSOTf, toluene, ꢀ408C, 71%;
cod=cyclooctadiene, DBU=1,3-diazabicyclo[5.4.0]undecene; MS=molecular sieves; PBS=phosphate buffered saline; TBS=tert-butyldimethylsilyl; Tf=tri-
fluoromethanesulfonyl; TMS=trimethylsilyl.
charide 16 in 82% yield (a/b ratio 6:1). Use of thiophene,
which forms a b-configured intermediate during the glyco-
sylation reaction, led to preferential formation of the desired
a-glycoside.[16] However, the anomeric mixture was insepa-
rable by column chromatography. Following HF-pyridine
cleavage of the TBDPS ether, glycosylation with imidate 6a
formed branched pentasaccharide 18 (a/b ratio 15:1), which
was isolated from the crude mixture of four diastereomers in
69% yield. After reduction of the trichloroacetamide with
zinc in acetic acid at 558C, the anomeric allyl ether at the
reducing end of glycan 19 was removed by isomerization in
the presence of iridium catalyst, followed by mild hydrolysis
using mercury salts.[8g,17] Hemiacetal 20 was then transformed
into trichloracetimidate 21, which was used in an a-stereose-
lective [5+2]-glycosylation of pseudodisaccharide 7a to
afford the orthogonally protected glycan 22 in 71% yield.
Removal of the allyl ether from glycan 22 provided, in
80% yield, alcohol 23 (Scheme 5). Phosphonylation via the
mixed anhydride prepared in situ from H-phosphonate 3 and
pivaloyl chloride, then oxidation by I2 in aqueous solution[8f,18]
and consequent removal of the TIPS group under acidic
conditions by treatment with scandium triflate gave glycolipid
Scheme 5. Synthesis of GPI 1. Reaction conditions: a) 1. [Ir(cod)-
24 in 64% yield starting from 23. The second phosphorylation
using H-phosphonate 4 and subsequent hydrazinolysis of the
levulinate ester afforded glycolipid 25, which after hydro-
genolysis using palladium on charcoal in a mixture of water,
methanol, and chloroform provided GPI 1 in 58% yield
starting from 24.[19]
(PPh2Me)2]PF6, H2, THF 2. HgCl2, HgO, H2O, acetone, 80%; b) 1. 3,
PivCl, pyridine; 2. I2, H2O; 3. Dowex 50WX8 (Na+), CHCl3, MeOH; 4.
Sc(OTf)3, MeCN, CHCl3, 64%; c) 1. 4, PivCl, pyridine; 2. I2, H2O; 3.
hydrazine, AcOH, pyridine, CHCl3, 94%; d) H2, Pd/C, CHCl3/MeOH/
H2O (3:3:1), 62%. Piv=pivaloyl.
The synthetic carbohydrate moiety of GPI 1 was printed
on appropriately coated glass slides and covalently linked to
the surface.[20] This glycoarray was incubated with monoclonal
antibodies (mAb), obtained by immunization of mice with
T. gondii and subsequent subcloning. The mAbT54E10 spe-
cifically recognizes GPI 1 while mAbT33F12 recognizes the
GPI structure lacking the a-glucose. These experiments were
previously performed exclusively with materials isolated from
parasites.[6,9] Our binding assays showed that mAbT54E10
Angew. Chem. Int. Ed. 2011, 50, 9961 –9964
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 9963

Communications
[7] N. Azzouz, F. Kamena, P. H. Seeberger, OMICS 2010, 14, 445 –
454.
strongly binds the carbohydrate moiety of the GPI 1 while
mAbT33F12 has lower affinity for this structure. Therefore,
we confirmed the structure of the “low-molecular-weight
antigen” of T. gondii and the selectivity of mAbT54E10 and
mAbT33F12 antibodies by total synthesis.[21]
In summary, we have developed a strategy for the
synthesis of branched GPI anchors and demonstrated its
utility in the first total synthesis of the complete surface
antigen 1 of T. gondii. We also showed that the synthetic GPI
1 is specifically recognized by monoclonal antibodies. Based
on these results, it is possible to develop a microarray
diagnostic test to distinguish healthy people from T. gondii
infected patients if polyclonal antibodies from blood also
specifically recognize structure 1.[22]
[8] Review: a) A. V. Nikolaev, N. Al-Maharik, Nat. Prod. Rep. 2011,
28, 970 – 1020; Original work: b) C. Murakata, T. Ogawa,
Carbohydr. Res. 1992, 235, 95 – 114; c) D. K. Baeschlin, A. R.
Chaperon, L. G. Green, M. G. Hahn, S. J. Ince, S. V. Ley, Chem.
Eur. J. 2000, 6, 172 – 186; d) K. Pekari, R. R. Schmidt, J. Org.
Chem. 2003, 68, 1295 – 1308; e) Y.-U. Kwon, X. Liu, P. H.
Seeberger, Chem. Commun. 2004, 1706 – 1707; f) X. Liu, Y.-U.
Kwon, P. H. Seeberger, J. Am. Chem. Soc. 2005, 127, 5004 – 5005;
g) B. M. Swarts, Z. Guo, J. Am. Chem. Soc. 2010, 132, 6648 –
6650.
[9] S. Tomavo, G. Couvreur, M. A. Leriche, A. Sadak, A. A.
Achbarou, B. Fortier, J. F. Dubremetz, Parasitology 1994, 108,
139 – 145.
[10] P. Cmoch, Z. Pakulski, Tetrahedron: Asymmetry 2008, 19, 1494 –
1503.
[11] M. Csꢀvꢀs, A. Borbꢁs, L. Szilagyi, A. Liptꢁk, Synlett 2002, 887 –
890.
[12] J. H. van Boom, P. M. J. Burgers, Tetrahedron Lett. 1976, 17,
4875 – 4878.
Received: May 20, 2011
Revised: July 4, 2011
Published online: September 5, 2011
[13] J. Xia, C. F. Piskorz, J. L. Alderfer, R. D. Locke, K. L. Matta,
Tetrahedron Lett. 2000, 41, 2773 – 2776.
[14] M. J. Ferguson, S. W. Homans, R. A. Dwek, T. W. Rademacher,
Science 1988, 239, 753 – 759.
[15] R. R. Schmidt, M. Stumpp, J. Michel, Tetrahedron Lett. 1982, 23,
405 – 408.
[16] J. Park, S. Kawatkar, J.-H. Kim, G.-J. Boons, Org. Lett. 2007, 9,
1959 – 1962.
Keywords: cell surface antigens · glycosylphosphatidylinositol ·
natural products · synthetic strategy · Toxoplasma gondii
.
[1] R. Gigg, J. Gigg, Glycopeptides and Related Compounds (Eds.:
D. G. Large, C. D. Warren), Marcel Dekker, New York, 1997,
pp. 327 – 392.
[2] Z. Guo, L. Bishop, Eur. J. Org. Chem. 2004, 3585 – 3596.
[3] T. B. Nicholson, C. P. Stanners, J. Cell Biol. 2006, 175, 647 – 659.
[4] S. Hiscox, M. B. Hallett, B. P. Morgan, C. W. van den Berg,
Biochem. Biophys. Res. Commun. 2002, 293, 714 – 721.
[5] M. A. J. Ferguson, J. Cell Sci. 1999, 112, 2799 – 2809.
[6] a) N. Azzouz, H. Shams-Eldin, S. Niehus, F. Debierre-Grocki-
ego, U. Bieker, J. Schmidt, C. Mercier, M.-F. Delauw, J.-F.
Dubremetz, T. K. Smith, R. T. Schwarz, Int. J. Biochem. Cell
Biol. 2006, 38, 1914 – 1925; b) B. Striepen, C. F. Zinecker, J. B. L.
Damm, P. A. T. Melgers, G. J. Gerwig, M. Koolen, J. F. G.
Vliegenthart, J.-F. Dubremetz, R. T. Schwartz, J. Mol. Biol.
1997, 266, 797 – 813.
[17] J. J. Oltvoort, C. A. A. van Boeckel, J. H. de Koning, J. H.
van Boom, Synthesis 1981, 305 – 308.
[18] R. H. Hall, A. Todd, R. F. Webb, J. Chem. Soc. 1957, 3291 – 3296.
[19] Analytical data of GPI 1, see Supporting Information.
[20] T. Horlacher, P. H. Seeberger, Chem. Soc. Rev. 2008, 37, 1414 –
1422.
[21] For pictures of all microarrays see Supporting Information.
[22] F. Kamena, M. Tamborrini, X. Liu, Y.-U. Kwon, F. Thompson, G.
Pluschke, P. H. Seeberger, Nat. Chem. Biol. 2008, 4, 238 – 240.
9964 www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 9961 –9964