Total synthesis of the Bacteroides fragilis zwitterionic polysaccharide a1 repeating unit

Article abstract of DOI:10.1021/ja1087375

Nearly all bacteria capsular polysaccharides are T-cell-independent antigens that do not promote immunoglobulin class switching from IgM to IgG nor memory responses. In contrast, zwitterionic polysaccharides activate T-cell-dependent immune responses by m

Full text of DOI:10.1021/ja1087375

Published on Web 12/10/2010
Total Synthesis of the Bacteroides fragilis Zwitterionic
Polysaccharide A1 Repeating Unit
Rajan Pragani and Peter H. Seeberger*
Max Planck Institute of Colloids and Interfaces, Department of Biomolecular Systems, Am
Mu¨hlenberg 1, 14476 Potsdam, Germany, and Freie UniVersita¨t Berlin, Institute for Chemistry
and Biochemistry, Arnimallee 22, 14195 Berlin, Germany
Received September 28, 2010; E-mail: peter.seeberger@mpikg.mpg.de
Abstract: Nearly all bacteria capsular polysaccharides are T-cell-independent antigens that do not promote
immunoglobulin class switching from IgM to IgG nor memory responses. In contrast, zwitterionic
polysaccharides activate T-cell-dependent immune responses by major histocompatability complex class
II presentation, a mechanism previously believed to be reserved for peptidic antigens. The best studied
zwitterionic polysaccharide, polysaccharide A1 (PS A1) is found on the capsule of the commensal bacteria
Bacteroides fragilis. Its potent immunomodulatory properties have been linked to postoperative intra-
abdominal abscess formation. Here, we report the synthesis of the PS A1 tetrasaccharide repeating unit
(2) as a tool to investigate the biological role of this polysaccharide. A modular synthetic strategy originating
from the reducing end of the PS A1 repeating unit was unsuccessful and illustrated the limitations of
glycosylation reactions between highly armed glycosylating agents and poor nucleophiles. Thus, a [3 + 1]
glycosylation relying on trisaccharide 5 and pyruvalated galactose 6 was used to complete the first total
synthesis of the PS A1 repeating unit (2).
Introduction
Zwitterionic polysaccharides (ZPSs) are a unique class of
immunomodulatory agents that can activate a major histocom-
patability complex class II (MHCII)-mediated T-cell-dependent
immune response in the absence of protein.1,2 The promise of
ZPSs as immunotherapeutic agents is slowly being realized.
Several semisynthetic polysaccharide-derived ZPSs have been
constructed that exhibit potent immunostimulatory activity,3 and
a cancer vaccine candidate composed entirely of carbohydrate
moieties has been developed using a conjugate of ZPS and the
carbohydrate hapten, Tn.4 Finally, it has been demonstrated that
structurally different ZPSs appear to stimulate distinct im-
munological responses, including a host memory immune
response.1-5
Figure 1. Structure of native PS A1 (1).
The best characterized natural ZPS is polysaccharide A1 (PS
A1),6 which is found on the capsule of the commensal bacteria
Bacteroides fragilis (Figure 1). Although B. fragilis accounts
for less than 0.5% of the normal colonic microflora in humans,
it is the predominant obligate anaerobe isolated from postopera-
tive intra-abdominal abscesses.7 Studies have determined that
PS A1’s ability to strongly stimulate CD4+ T-cells to produce
cytokines (e.g., IL-2, IL-10, IL-12, IL-17, interferon-γ, and TNF-
R) and chemokines strengthens the localized immune response
to form abscesses around foreign infectious agents. In addition
to T-cell stimulation via MHCII presentation, PS A1 can also
initiate an innate immune response through Toll-like receptor
2 (TLR2) signaling.8 Curiously, the anti-inflammatory properties
of PS A1, mediated by IL-10 production, can also prevent
(1) For selected reviews on zwitterionic polysaccharides, see: (a) Cobb,
B. A.; Kasper, D. L. Cell. Microbiol. 2005, 7, 1398. (b) Mazmanian,
S. K.; Kasper, D. L. Nat. ReV. Immunol. 2006, 6, 849. (c) Avci, F. Y.;
Kasper, D. L. Annu. ReV. Immunol. 2010, 28, 107.
(2) (a) Cobb, B. A.; Wang, Q.; Tzianabos, A. O.; Kasper, D. L. Cell 2004,
117, 677. (b) Cobb, B. A.; Kasper, D. L. Glycobiology 2008, 18, 707.
(3) (a) Gallorini, S.; Berti, F.; Mancuso, G.; Cozzi, R.; Tortoli, M.; Volpini,
G.; Telford, J. L.; Beninati, C.; Maione, D.; Wack, A. Proc. Natl.
Acad. Sci. U.S.A. 2009, 106, 17481. (b) Meng, C.; Peng, X.; Shi, X.;
Wang, H.; Guo, Y. Acta Biochim. Biophys. Sin 2009, 41, 737. (c)
Abdulamir, A. S.; Hafidh, R. R.; Abubaker, F. Curr. Ther. Res. Clin.
Exp. 2010, 71, 60.
(4) De Silva, R. A.; Wang, Q.; Chidley, T.; Appulage, D. K.; Andreana,
P. R. J. Am. Chem. Soc. 2009, 131, 9622.
(6) Baumann, H.; Tzianabos, A. O.; Brisson, J.-R.; Kasper, D. L.; Jennings,
H. J. Biochemistry 1992, 31, 4081.
(5) (a) Groneck, L.; Schrama, D.; Fabri, M.; Stephen, T. L.; Harms, F.;
Meemboor, S.; Hafke, H.; Bessler, M.; Becker, J. C.; Kalka-Moll,
W. M. Infect. Immun. 2009, 77, 3705. (b) Meemboor, S.; Mertens, J.;
Flenner, E.; Groneck, L.; Zingarelli, A.; Gamsta¨tter, T.; Bessler, M.;
Seeger, J. M.; Kashkar, H.; Odenthal, M.; Kalka-Moll, W. M. Innate
Immun. 2010, 16, 310.
(7) Tzianabos, A. O.; Onderdonk, A. B.; Rosner, B.; Cisneros, R. L.;
Kasper, D. L. Science 1993, 262, 416.
(8) Wang, Q.; McLoughlin, R. M.; Cobb, B. A.; Charrel-Dennis, M.;
Zaleski, K. J.; Golenbock, D.; Tzianabos, A. O.; Kasper, D. L. J. Exp.
Med. 2006, 203, 2853.
9
102 J. AM. CHEM. SOC. 2011, 133, 102–107
10.1021/ja1087375  2011 American Chemical Society

Polysaccharide A1 Repeating Unit Synthesis
A R T I C L E S
Scheme 1. Retrosynthetic Analysis of PS A1 Repeating Unit 2
abscess formation against a B. fragilis challenge in germ-free
mice.9 Furthermore, by stimulation of IL-10 secretion, PS A1
can modulate surgical fibrosis,9 inhibit intestinal inflammatory
disease caused by Helicobacter hepaticus,10 and protect against
central nervous system (CNS) demyelinating disease.11 Beyond
its anti-inflammatory activities, PS A1 plays a role in the
development and the maintenance of a balanced mammalian
immune system.8,12 In germ-free mice, administration of PS A1-
bearing B. fragilis corrected systemic T-cell deficiencies, Th1/
Th2 imbalances, and directed lymphoid organogenesis.
not adopt this conformation and are unable to bind to MHCII
and stimulate T-cell expansion.17
The challenging structural features of PS A1 (1) and the lack
of chemically defined PS A1 fragments for use as mechanistic
probes to study B. fragilis provided the impetus to synthesize
the PS A1 repeating unit (2). Currently, only the fully protected
PS A1 repeating unit has been synthesized.18
Results and Discussion
Initial Retrosynthetic Considerations. Previous studies18
highlighted the challenges associated with synthesizing the
repeating unit (2) via trisaccharide nucleophile 4 (Scheme 1,
path A), when an AAT thioglycoside or lactol served as the
glycosylating agent. The low nucleophilicity of 4 and highly
electron-rich nature of the glycosylating agents were most likely
culpable for the low-yielding coupling. However, we had
demonstrated in earlier studies19 that AAT building block 3 can
partake in high-yielding glycosylation reactions with a C4-OH
galactosamine nucleophile. Thus, we were optimistic that
glycosyl N-phenyl trifluoroacetimidate 3 would be able to
overcome the deficiencies of previously used AAT glycosylating
agents. The modular nature of path A was particularly attractive
in light of a future automated synthesis.20
A second retrosynthetic disconnection was considered, in case
path A was not viable. Path B involved the [3 + 1] glycosylation
between trisaccharide glycosylating agent 5 and pyruvalated
galactose nucleophile 6 (Scheme 1). This disconnection appeared
attractive due to the involvement of a more electron-deficient
glycosylating agent (5) and a more nucleophilic C3-OH
galactose residue (6).
The interesting biological profile of native PS A1 (1) is
complemented by a unique structural architecture (Figure 1).
Indeed, the pyruvalated galactose, galactofuranose, and 2-ac-
etamido-4-amino-2,4,6-trideoxy-D-galactose (AAT) are novel
residues often found in immunodominant epitopes.1,13,14 The
zwitterionic charge motif of PS A1 is crucial for the activation
of a T-cell-dependent immune response, since chemical removal
of either the positive or negative charge eliminates MHCII
binding.1,2 The polysaccharides are around 120 repeating units
long. Sp1 and PS A2, ZPSs closely related to PS A1, adopt an
extended right-handed helix15 with two repeating units per turn
and a pitch of 20 Å. This R-helical architecture is necessary
for MHCII binding. Circular dichroism studies16 have shown
that PS A1 fragments exhibit a similar helical architecture;
however, fragments with fewer than three repeating units do
(9) Ruiz-Perez, B.; Chung, D. R.; Sharpe, A. H.; Yagita, H.; Kalka-Moll,
W. M.; Sayegh, M. H.; Kasper, D. L.; Tzianabos, A. O. Proc. Natl.
Acad. Sci. U.S.A. 2005, 102, 16753.
Building Block Synthesis. Our general strategy relied on
N-phenyl trifluoroacetimidate glycosylating agents,21 since we19
and others22 had found them to perform better than glycosyl
trichloroacetimidates in glycosylations involving electron-rich
deoxysugars. Most of the monosaccharide building blocks
(10) Mazmanian, S. K.; Round, J. L.; Kasper, D. L. Nature 2008, 453,
620.
(11) Ochoa-Repa´raz, J.; Mielcarz, D. W.; Wang, Y.; Begum-Haque, S.;
Dasgupta, S.; Kasper, D. L.; Kasper, L. H. Mucosal Immunol. 2010,
3, 487.
(12) Mazmanian, S. K.; Liu, C. H.; Tzianabos, A. O.; Kasper, D. L. Cell
2005, 122, 107.
(17) Kalka-Moll, W. M.; Tzianabos, A. O.; Wang, Y.; Carey, V. J.; Finberg,
R. W.; Onderdonk, A. B.; Kasper, D. L. J. Immunol. 2000, 164, 719.
(18) van den Bos, L. J.; Boltje, T. J.; Provoost, T.; Mazurek, J.; Overkleeft,
H. S.; van der Marel, G. A. Tetrahedron Lett. 2007, 48, 2697.
(19) Pragani, R.; Stallforth, P.; Seeberger, P. H. Org. Lett. 2010, 12, 1624.
(20) Seeberger, P. H. Chem. Soc. ReV. 2008, 37, 19.
(13) Bennett, L. G.; Bishop, C. T. Immunochemistry 1977, 14, 693.
(14) Suzuki, E.; Toledo, M. S.; Takahashi, H. K.; Straus, A. H. Glycobiology
1997, 7, 463.
(15) (a) Wang, Y.; Kalka-Moll, W. M.; Roehrl, M. H.; Kasper, D. L. Proc.
Natl. Acad. Sci. U.S.A. 2000, 97, 13478. (b) Choi, Y.-H.; Roehrl,
M. H.; Kasper, D. L.; Wang, J. Y. Biochemistry 2002, 41, 15144.
(16) Kreisman, L. S. C.; Friedman, J. H.; Neaga, A.; Cobb, B. A.
Glycobiology 2006, 17, 46.
(21) Yu, B.; Tao, H. Tetrahedron Lett. 2001, 42, 2405.
(22) Comegna, D.; Bedini, E.; Di Nola, A.; Iadonisi, A.; Parrilli, M.
Carbohydr. Res. 2007, 342, 1021.
9
J. AM. CHEM. SOC. VOL. 133, NO. 1, 2011 103

A R T I C L E S
Pragani and Seeberger
Scheme 3. Synthesis of Building Blocks 6, 18, and 20a
Scheme 2. De Novo Synthesis of AAT Building Block 3a
a Reagents and conditions (a) i., AcCl, MeOH, 23 °C; ii., Ac2O, NEt3,
DMAP, CH2Cl2, 0 to 23 °C, 95%, two steps; (b) i., LHMDS, THF, -78 to
23 °C; ii., NaHCO3, Me2SO4, acetone, 23 °C, 73%, two steps; (c) DIBAL,
THF, -78 °C, then H+; (d) NaBH4, CeCl3 ·7H2O, MeOH, -78 °C, 77%,
two steps; (e) Ac2O, NEt3, DMAP, CH2Cl2, 23 °C, 90%; (f) CAN, NaN3,
CH3CN, -20 °C, 67% (3.5:1 dr); (g) i., p-TolSH, DIPEA, CH3CN, 23 °C;
ii., F3CC(NPh)Cl, Cs2CO3, CH2Cl2, 23 °C, 69%, two steps.
a Reagents and conditions (a) i., FmocCl, py., 0 °C, 79%; ii., BzCl, py.,
0 °C, 82%; iii., TsOH, MeOH, 40 °C; iv., CH3C(O)CO2Me, BF3 ·OEt2,
CH3CN, 23 °C, 39%, two steps; (b) i., i-PrOH, NIS, AgOTf, CH2Cl2, 0 °C;
ii., NEt3, CH2Cl2, 23 °C, 49%, two steps; (c) LevOH, DIPC, DMAP, CH2Cl2,
0 to 23 °C, 90%; (d) i., NIS, THF, H2O, 23 °C; ii., F3CC(NPh)Cl, Cs2CO3,
CH2Cl2, 0 °C, 83%, two steps; (e) i., NBS, EtOAc, H2O, 23 °C, 91%; ii.,
F3CC(NPh)Cl, Cs2CO3, CH2Cl2, 23 °C, 87%.
needed to prepare the PS A1 repeating unit (2) are readily
accessed by using procedures modified from previously pub-
lished methods. The synthesis of the AAT building block would
benefit from a more efficient route when compared to known
procedures.18,23
from known benzylidene galactoside 14.30 First, the C3 and C2
hydroxyl groups of 14 were selectively protected with Fmoc
(79% yield)31 and Bz (82% yield) groups, respectively. Cleavage
of the benzylidene acetal and BF3 ·OEt2 mediated pyruvate ketal
formation32 afforded ketal 15 as a single diastereomer in 39%
yield over two steps. The high diastereoselectivity of the
thermodynamically driven ketalization is ascribed to the interac-
tion of the low-lying unoccupied σ* orbital with the axial lone
pairs of the ketal moiety in the six-membered ring of 15. This
favorable anomeric interaction33 would not be possible for the
other diastereomer. Finally, NIS/AgOTf promoted glycosylation
of 15 with isopropanol followed by Fmoc cleavage furnished
galactose 6. At this stage, the stereochemistry at the pyruvate
quaternary carbon in galactose 6 was verified through a nuclear
Overhauser effect (NOE) interaction between the equatorial C4
hydrogen and the pyruvate methyl ester.
Galactosamine building block 18 was synthesized in a
straightforward fashion from selenide 16.34 Levulinoylation of
the C3-OH provided 17 in 90% yield. Selenide 17 was further
treated with NIS to furnish an intermediate lactol that was
converted into N-phenyl trifluoroacetimidate 18 in 83% yield
over two steps. The final building block, galactofuranose 20,
was synthesized from known thiol 1935 by conversion to the
lactol with N-bromosuccinimide (NBS) followed by N-phenyl
trifluoroacetimidate formation. In storage, building block 20 was
more stable than the trichloroacetimidate analog.36
The C3-C6 portion of the AAT building block (3) was
mapped back to L-threonine (Scheme 2). As a Cbz group was
deemed appropriate for C4 amine protection, N-Cbz-L-threonine
7 was the starting point of the de novo AAT building block
synthesis.19 Conversion of 7 to a methyl ester followed by
acetylation provided acetate ester 8 in 95% yield over two steps.
Lithium bis(trimethylsilylamide) (LHMDS)-mediated Dieck-
mann cyclization24,25 and K2CO3/Me2SO4 methylation then
afforded enoate 9 in 73% yield over two steps. The 1,2-
diisobutylaluminium hydride (DIBAL) reduction26 of the enoate
followed by acidic workup gave intermediate 10, which was
reduced under Luche conditions27 to afford allylic alcohol 11
in 77% yield over two steps. Acetylation of the alcohol set the
stage for the introduction of the C2 azide. Azidonitration23b,28
of glycal 12 furnished nitrate 13. Nitrate 13 was transformed
into AAT building block 3 by cleavage of the anomeric nitrate
with p-TolSH/DIPEA29 followed by N-phenyl trifluoroacetimi-
date formation in 69% yield over two steps.
Having established a route for the AAT monosaccharide (3),
the remaining three monosaccharide building blocks were
pursued (Scheme 3). Pyruvalated galactose 6 was synthesized
Forays Toward the PS A1 Repeating Unit. With the building
blocks in hand, retrosynthetic path A (Scheme 1), involving a
modular approach that begins from the reducing end, was
reduced to practice (Scheme 4). Pyruvalated galactose 6 was
(23) (a) Hermans, J. P. G.; Elie, C. J. J.; van der Marel, G. A.; van Boom,
J. H. J. Carbohydr. Chem. 1987, 6, 451. (b) Smid, P.; Jo¨rning,
W. P. A.; van Duuren, A. M. G.; Boons, G. J. P. H.; van der Marel,
G. A.; van Boom, J. H. J. Carbohydr. Chem. 1992, 11, 849. (c) Liang,
H.; Grindley, T. B. J. Carbohydr. Chem. 2004, 23, 71. (d) Cai, Y.;
Ling, C.-C.; Bundle, D. R. J. Org. Chem. 2009, 74, 580. (e) Pedersen,
C. M.; Figueroa-Perez, I.; Lindner, B.; Ulmer, A. J.; Za¨hringer, U.;
Schmidt, R. R. Angew. Chem., Int. Ed. 2010, 49, 2585.
(30) Dhe´nin, S. G. Y.; Moreau, V.; Morel, N.; Nevers, M.-C.; Volland,
H.; Cre´minon, C.; Djeda¨ıni-Pilard, F. Carbohydr. Res. 2008, 343, 2101.
(31) Mogemark, M.; Elofsson, M.; Kihlberg, J. J. Org. Chem. 2003, 68,
7281.
(24) Ge, P.; Kirk, K. L. J. Org. Chem. 1996, 61, 8671.
(25) Ren, F.; Hogan, P. C.; Anderson, A. J.; Myers, A. G. Org. Lett. 2007,
9, 1923.
(32) Ziegler, T. Tetrahedron Lett. 1994, 35, 6857.
(33) Tschierske, C.; Ko¨hler, H.; Zaschke, H.; Kleinpeter, E. Tetrahedron
1989, 45, 6987.
(26) Kocienski, P.; Narquizian, R.; Raubo, P.; Smith, C.; Farrugia, L. J.;
Muir, K.; Boyle, F. T. J. Chem. Soc., Perkin Trans. 1 2000, 2357.
(27) Luche, J.-L. J. Am. Chem. Soc. 1978, 100, 2226.
(34) Mironov, Y. V.; Sherman, A. A.; Nifantiev, N. E. Tetrahedron Lett.
2004, 45, 9107.
(28) Lemieux, R. U.; Ratcliffe, R. M. Can. J. Chem. 1979, 57, 1244.
(29) Gauffeny, F.; Marra, A.; Shun, L. K. S.; Sinay¨, P.; Tabeur, C.
Carbohydr. Res. 1991, 219, 237.
(35) Completo, G. C.; Lowary, T. L. J. Org. Chem. 2008, 73, 4513.
(36) Gallo-Rodriguez, C.; Gandolfi, L.; de Lederkremer, R. M. Org. Lett.
1999, 1, 245.
9
104 J. AM. CHEM. SOC. VOL. 133, NO. 1, 2011

Polysaccharide A1 Repeating Unit Synthesis
A R T I C L E S
Scheme 4. Attempted Glycosylations of AAT Building Block 3a
a Reagents and conditions (a) i., TMSOTf, CH2Cl2, 0 °C, 72%; ii., N2H4 ·H2O, AcOH, py., CH2Cl2, 23 °C, 89%; (b) i., TMSOTf, CH2Cl2, -60 °C, 57%;
ii., TES, TfOH, 4 Å MS, CH2Cl2, -50 °C, 54%; (c) TMSOTf, CH2Cl2, 0 °C, 0%; (d) i., TMSOTf, CH2Cl2, 0 °C, 72%; ii., TES, BF3 ·OEt2, CH2Cl2, 0 °C,
54%; (e) TMSOTf, CH2Cl2, 0 °C, 0%.
Scheme 5. Successful Glycosylation of AAT Building Block 3a
R-pyruvulated galactose preferentially occupying the space
below the galactosamine residue near the C6 benzyl ether. This
steric crowding likely forces the C6 benzyl ether to the top face
of the galactosamine residue, thereby shielding the already
poorly nucleophilic C4-OH. However, in nucleophile 25, the
equatorial ꢀ-OTBS group is projected away from the ring,
permitting the C6 benzyl ether to rotate below the ring. Thus,
nucleophile 25 constitutes an attractive means for glycosidic
a Reagents and conditions (a) TMSOTf, CH2Cl2, 0 °C, 74%.
bond formation.
Total Synthesis of the PS A1 Repeating Subunit. Due to
difficulties establishing a modular route to the tetrasaccharide
repeat, AAT-containing cassette 26 was used to finish the
synthesis via retrosynthetic path B (Scheme 1). Thus, the use
of AAT building block 3 in a late-stage, sterically demanding
glycosylation would be avoided.
successfully coupled with building block 18 in 72% yield. Lev
deprotection using N2H4 ·H2O furnished disaccharide 21 in 89%
yield. Galactofuranose 20 was then appended to disaccharide
21, and the resulting trisaccharide was reductively ring-opened
with TES/TfOH at -50 °C to provide alcohol 4. Attempted
glycosylation of alcohol 4 with AAT building block 3 provided
no trace of product 22. Since other building blocks, including
an AAT thioglycoside activated with Ph2SO/Tf2O or NIS/
TMSOTf and an AAT lactol activated with Ph2SO/Tf2O, poorly
couple with alcohol 4,18 glycosylation with a less complex
nucleophile was attempted.
It was thought that by swapping out the bulky C3 perben-
zoylated galactofuranose of 4 for a smaller protecting group,
such as a levulinoyl ester, the glycosylation might proceed.
Disaccharide 23 was prepared by glycosylation of building block
18 and nucleophile 6 followed by reductive TES/BF3 ·OEt2
opening of the benzylidene acetal ring. Unfortunately, the
glycosylation between nucleophile 23 and AAT 3 was not
successful.
The C3 naphthalene ether of disaccharide 26 was removed
using 2,3-dichloro-5,6-dicyano-1,4-benzoquinone (DDQ) in 86%
yield (Scheme 6). Glycosylation of resulting alcohol 27 with
galactofuranose N-phenyl trifluoroacetimidate 20 proceeded best
at -30 °C to afford trisaccharide adduct 28 in 90% yield as the
ꢀ-isomer. The observed 13C chemical shift of 106.5 ppm for
theanomericcarbonofthegalactofuranoseishighlysupportive35,38
of a ꢀ-linkage. The anomeric tert-butyl-dimethylsilyl (TBS)
protecting group was removed using tetrabutylammonium
fluoride (TBAF) buffered with AcOH. The resulting lactol was
finally converted into N-phenyl trifluoroacetimidate 29 in 82%
yield over two steps. Thioethyl glycoside 5 was also prepared
from glycosyl imidate 29 in 96% yield for glycosylation trials.
This approach also posed some interesting challenges: the
union of N-phenyl trifluoroacetimidate 29 and pyruvalated
galactose 6 yielded only trace amounts of adduct 22 (Table 1,
entry 1). Steric interactions between the AAT residue, galac-
tosamine C6 benzyl ether, and pyruvalated galactose that
prevented the glycosylation of nucleophiles 4 and 23 with AAT
3 (Scheme 4) were likely also involved here. A breakthrough
The failure of nucleophiles 4 and 23 to couple with AAT
building block 3 was surprising in light of the successful
glycosylation between galactosamine 2519,37 and 3 (Scheme 5).
The primary difference of 4 and 23 with 25 was the nature of
the anomeric substituent. With nucleophiles 4 and 23, molecular
mechanics (MM2) energy minimized models showed the
(38) Beier, R. C.; Mundy, B. P.; Strobel, G. A. Can. J. Chem. 1980, 58,
(37) Grundler, G.; Schmidt, R. R. Liebigs Ann. Chem. 1984, 1826.
2800.
9
J. AM. CHEM. SOC. VOL. 133, NO. 1, 2011 105

A R T I C L E S
Pragani and Seeberger
Scheme 6. Synthesis of AAT-containing Trisaccharide 5a
promoters42 for thioglycosides with less nucleophilic counterions
were tested. Although MeOTf could not activate thioglycoside
5 (Table 1, entry 3), methylsulfenylating agent dimethyl(meth-
ylthio)sulfonium triflate (DMTST)40,43 promoted the glycosy-
lation to furnish tetrasaccharide 22 in 58% yield as the R-anomer
(Table 1, entry 4). The stereochemistry of the newly formed
anomeric linkage was verified by coupling constant analysis
(3JH1,H2 ) 3.0 Hz for the anomeric position of the galactosamine
residue). Low-temperature activation of thioglycoside 5 with
Ph2SO/Tf2O44 followed by addition of 6 was also unsuccessful
in forming 22 (Table 1, entry 5).
The final stages of the synthesis called for the conversion of
the azides in 22 to be converted to acetamides (Scheme 7). Initial
studies using Staudinger conditions45 (PMe3 in THF/H2O)
followed by acetylation gave inconsistent yields of diacetamide
30 with poorer results associated with the absence of water
during the Staudinger reduction. Employing AcSH and pyridine
to effect a reduction-acetylation sequence46,47 allowed for a
one-pot reaction, ensured constant buffering of the moderately
acid-labile deoxy sugars, and provided consistent and clean
reactions. Diacetamide 30 was isolated in 67% yield from
diazide 22.
Global deprotection of acetamide 30 was not entirely
straightforward (Scheme 7). Employing KOH in THF/H2O first
to cleave the ester groups showed gradual formation of a cyclic
carbonate before all benzoates had been cleaved. Hydrogenation
of the intermediate with Pearlman’s catalyst48 afforded solely
cyclic carbonate 31. Since the cleavage of a cyclic carbonate
in the presence of two acetamides was considered risky, the
final steps were rearranged. By performing first the hydrogena-
tion, Pearlman’s catalyst cleanly removed the benzyl and Cbz
groups. However, methanolysis of the esters using NaOMe in
MeOH (Zemple´n conditions) followed by the addition of water49
to cleave the pyruvate ester resulted in significant quantities of
triacetate 32 (along with desired 2). This unexpected product
originated from the migration of the AAT C3 acetate to the C4
amine. It has been suggested that a decrease in solvent polarity
can prevent the migration of a C8 acetate to the C9 position in
a neuraminic acid derivative.50 Thus, by first dissolving the
intermediate that results from hydrogenolysis in THF and then
treating this mixture, in dropwise fashion, with NaOMe in
MeOH/H2O, only a very minor amount of acetyl migration was
observed. Under these conditions, the PS A1 tetrasaccharide
repeating unit (2) was produced in 46% yield over the two steps.
The 1H NMR of PS A1 repeating unit 2 (Figure 2) compares
remarkably well with native PS A1 (1).51 This observation came
as somewhat of a surprise since native PS A1 (1) is thought to
exist in a helical secondary structure, whereas PS A1 fragments
of less than three repeating units are too small to form helices.16
a Reagents and conditions (a) DDQ, MeOH, CH2Cl2, 23 °C, 86%; (b)
TMSOTf, CH2Cl2, -30 °C, 90%; (c) i., TBAF, AcOH, THF, 0 °C; ii.,
F3CC(NPh)Cl, Cs2CO3, CH2Cl2, 23 °C, 82%, two steps; (d) EtSH, TMSOTf,
4 Å MS, CH2Cl2, 0 °C, 96%.
Table 1. Optimization of Tetrasaccharide 22 Formation
entry
building block
conditions
yield, %
1
2
3
4
5
29
5
5
5
5
TMSOTf, 0 °C
<10
26
0
58
0
NIS, AgOTf, 0 °C
MeOTf, 0 °C
DMTST, TTBP, 0 °C
Ph2SO, Tf2O, TTBP, -60 to 0 °C
1
was finally achieved when thioglycoside 5 served as the coupling
partner for 6. Using NIS/AgOTf as the promoter system,
tetrasaccharide 22 was isolated in 26% yield as the R-anomer
(Table 1, entry 2). Our inspiration to use thioglycoside 5 came
from Schmidt,39 who had used a thioglycoside in place of a
trichloroacetimidate for a glycosylation involving a sterically
hindered C2-OH glucose nucleophile. However, succinimide
addition40,41 into the activated trisaccharide was a major side
reaction of the NIS-promoted glycosylation. Thus, different
Such a difference in secondary structure is observed in the H
(42) Fu¨gedi, P.; Garegg, P. J.; Lo¨nn, H.; Norberg, T. Glycoconjugate J.
1987, 4, 97.
(43) Fu¨gedi, P.; Garegg, P. J. Carbohydr. Res. 1986, 149, C9.
(44) Code´e, J. D. C.; Litjens, R. E. J. N.; den Heeten, R.; Overkleeft, H. S.;
van Boom, J. H.; van der Marel, G. A. Org. Lett. 2003, 5, 1519.
(45) Nyffeler, P. T.; Liang, C.-H.; Koeller, K. M.; Wong, C.-H. J. Am.
Chem. Soc. 2002, 124, 10773.
(46) Rosen, T.; Lico, I. M.; Chu, D. T. W. J. Org. Chem. 1988, 53, 1580.
(47) Lohman, G. J. S.; Seeberger, P. H. J. Org. Chem. 2004, 69, 4081.
(48) Pearlman, W. M. Tetrahedron Lett. 1967, 17, 1663.
(49) Agnihotri, G.; Misra, A. K. Tetrahedron Lett. 2006, 47, 8493.
(50) Reinhard, B.; Faillard, H. Liebigs Ann. Chem. 1994, 193.
(51) Reprinted with the publisher’s permission: Kalka-Moll, W. M.;
Tzianabos, A. O.; Wang, Y.; Carey, V. J.; Finberg, R. W.; Onderdonk,
A. B.; Kasper, D. L. J. Immunol. 2000, 164, 719. Copyright 2000,
The American Association of Immunologists, Inc.
(39) Zhu, X.; Yu, B.; Hui, Y.; Schmidt, R. R. Eur. J. Org. Chem. 2004,
965.
(40) Grayson, E. J.; Ward, S. J.; Hall, A. L.; Rendle, P. M.; Gamblin, D. P.;
Batsanov, A. S.; Davis, B. G. J. Org. Chem. 2005, 70, 9740.
(41) Bai, Y.; Lowary, T. L. J. Org. Chem. 2006, 71, 9672.
9
106 J. AM. CHEM. SOC. VOL. 133, NO. 1, 2011

Polysaccharide A1 Repeating Unit Synthesis
A R T I C L E S
Scheme 7. Completion of PS A1 Tetrasaccharide Repeating Unit 2a
a Reagents and conditions (a) AcSH, py., 23 °C, 67%; (b) i., KOH, THF, H2O, 23 °C; ii., H2, Pd(OH)2/C, MeOH, 23 °C, quant. conv.; (c) i., H2, Pd(OH)2/
C, MeOH, 23 °C; ii., KOH, THF, H2O, 23 °C or NaOMe, MeOH, 23 °C, 12 h, then, H2O, ∼50% conv. (plus, ∼50% of 2); (d) i., H2, Pd(OH)2/C, MeOH,
23 °C; ii., THF, then 0.5 M NaOMe in 1:1 MeOH/H2O, 23 °C, 46%, two steps.
congested nature of the PS A1 tetrasaccharide repeating unit
(2), formation of the helical secondary structure of native PS
A1 (1) may have little effect on the conformation of the rigid
subunits.
Conclusion
Reported is the first total synthesis of a deprotected PS A1
repeating unit (2) (see Supporting Information) from N-Cbz-L-
threonine in 20 linear steps with a 1.8% overall yield. This
synthesis highlights the de novo preparation of monosaccharide
building blocks, such as AAT 3, as an important means to fuel
oligosaccharide synthesis. Furthermore, the need to further
develop and understand glycosylation techniques involving
unusual and complex carbohydrate targets was demonstrated.
A cursory examination of target 2 would not suggest that any
immediate steric complications would be encountered during
its construction. Not surprisingly, it was found that the same
steric and electronic issues that drive our understanding of
modern synthetic organic chemistry are also applicable to
carbohydrate synthesis. The methodology developed for the
synthesis of PS A1 repeating unit 2 is currently being used to
develop immunological probes for B. fragilis. We hope that
these probes will help unravel the mechanism and action of
zwitterionic PS A1.
Acknowledgment. The authors gratefully thank the Max Planck
Society for funding and the Alexander von Humboldt Foundation
for a postdoctoral research fellowship (R.P.). We would also like
Figure 2. 1H NMR comparison of native PS A1 (1) and PS A1 repeating
unit 2.
to thank Dr. Bernd Lepenies for insightful discussions and Dr. Mark
Schlegel for help with the HPLC.
NMR spectra of native ZPS Sp1 (which also exists as a right-
handed helix) and the chemically synthesized monomeric and
dimeric Sp1 repeating units.52 However, due to the sterically
Supporting Information Available: Experimental procedures
for the synthesis of PS A1 repeating unit 2, NMR spectral
characterization, and other characterization data. This material
is available free of charge via the Internet at http://pubs.acs.org.
(52) Wu, X.; Cui, L.; Lipinski, T.; Bundle, D. R. Chem.sEur. J. 2010,
16, 3476.
JA1087375
9
J. AM. CHEM. SOC. VOL. 133, NO. 1, 2011 107