Synthesis of the methyl glycoside of a branched octasaccharide fragment specific for the Shigella flexneri serotype 2a O-antigen

Article abstract of DOI:10.1016/S0040-4039(02)02055-5

An efficient synthesis of the methyl glycoside of a branched octasaccharide representative of Shigella flexneri serotype 2a O-specific polysaccharide is described. The synthesis is based on the use of the trichloroacetimidate methodology, and involves the

Full text of DOI:10.1016/S0040-4039(02)02055-5

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 43 (2002) 8215–8218
Synthesis of the methyl glycoside of a branched octasaccharide
fragment specific for the Shigella flexneri serotype 2a O-antigen†
Fre´de´ric Be´lot,a Corina Costachel,a,b,‡ Karen Wright,a,§ Armelle Phaliponb and
Laurence A. Mularda,*
aUnite´ de Chimie Organique, URA CNRS 2128, Institut Pasteur, 28 rue du Dr Roux, 75 724 Paris Cedex 15, France
bUnite´ de Pathoge´nie Microbienne Mole´culaire, INSERM U 389, Institut Pasteur, 28 rue du Dr Roux, 75 724 Paris Cedex 15,
France
Received 31 August 2002; revised 17 September 2002; accepted 19 September 2002
Abstract—An efficient synthesis of the methyl glycoside of a branched octasaccharide representative of Shigella flexneri serotype
2a O-specific polysaccharide is described. The synthesis is based on the use of the trichloroacetimidate methodology, and involves
the condensation of a pentasaccharide acceptor and a trisaccharide donor as the key step. © 2002 Elsevier Science Ltd. All rights
reserved.
Shigella flexneri serotype 2a is a common infective
agent in humans that is responsible for the endemic
form of shigellosis or bacillary dysentery.2 Shigellosis is
a priority target as defined by the World Health Orga-
nization since this disease is a major cause of mortality
in developing countries, especially among children and
in the immunocompromised population. Several pro-
grams targetting the eradication of this bacterial infec-
tion are under development. Amongst those, the design
of glycoconjugate vaccines based on the use of the
surface O-specific polysaccharide (O-SP) is of particular
interest. Indeed, analogously to the use of bacterial
capsular polysaccharide conjugates as efficacious vac-
cines against infections caused by Haemophilus influen-
zae type b or Neisseria meningitidis Group C,3–5 it was
hypothesized that protein conjugates of detoxified lipo-
polysaccharides (LPS) may protect humans against
infections caused by non-capsulated bacteria.6,7 There is
evidence that conjugates incorporating oligosaccharide
fragments of the native bacterial polysaccharides may
be immunogenic as well.8 As part of a program aimed
at the design of optimal vaccine conjugates based on
the use of synthetic fragments of the O-SP of S. flexneri
2a, the study of the interaction between the bacterial
surface polysaccharide and homologous protective
monoclonal antibodies is under investigation.
The O-SP of S. flexneri 2a is a heteropolysaccharide
defined by the branched pentasaccharide repeating unit
I.9,10 Besides the known methyl glycoside of the EC
disaccharide,11,12 a set of di- to pentasaccharides repre-
sentative of fragments of S. flexneri 2a O-SP have been
synthesized recently in this laboratory.13–15 The use of
these compounds as molecular probes for mapping at
the molecular level the binding characteristics of a set
of protective antibodies against S. flexneri 2a infection
indicated that access to larger oligosaccharides would
help the characterization of the carbohydrate epitopes.
For this purpose, methodologies allowing a straightfor-
ward access to S. flexneri 2a oligosaccharides of larger
size are under study in the laboratory. As part of this
study, we now report the synthesis of the first octasac-
charide in the series, namely the B(E)CDAB(E)C frag-
ment, which was prepared as its methyl glycoside (1).
Keywords: bacterial oligosaccharides; Shigella flexneri; chemical glycosylation.
* Corresponding author. Tel.: +33 1-40-61-36-77; fax: +33 1-45-68-84-04; e-mail: lmulard@pasteur.fr
† See Ref. 1.
‡ Present address: Unite´ des Aspergillus, Institut Pasteur, 25 rue du Dr. Roux, 75 724 Paris Cedex 15, France.
§ SIRCOB, Universite´ de Versailles, 45 avenue des Etats-Unis, 78 035 Versailles Cedex, France.
0040-4039/02/$ - see front matter © 2002 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(02)02055-5

8216
F. Be´lot et al. / Tetrahedron Letters 43 (2002) 8215–8218
acetal to give the required pentasaccharide acceptor 4
(72%, two steps). The use of the isopropylidene acetal,
despite its rather high lability under acidic conditions
compared to that of the more common benzylidene
protecting group, was a crucial element in the strategy.
In fact, the choice of the former derived from previous
work13 in the series demonstrating the lack of reactivity
under glycosylation conditions of similar acceptors
bearing a benzylidene-protected glucosaminyl residue
and confirmation in our hands of these observations
(not described). Next, the readily available trichloroace-
timidate 5 was selected as an efficient trisaccharide
donor (Scheme 2). Condensation of the known disac-
charide acceptor 615 and rhamnopyranosyl donor 719
gave the fully protected trisaccharide intermediate 8
(95%). A conventional two-step deallylation allowed
the selective removal of the anomeric protecting group
and consequently the smooth conversion of compound
8 into the corresponding hemiacetal 9. Standard activa-
tion of the latter resulted in the trichloroacetimidate
donor 5 (77% from 8).
As the linear construction of the target 1 is not practi-
cal, one of the challenges was to define the building
blocks of interest in the design of an efficient conver-
gent synthesis. Based on previous work in the S.
flexneri 2a series15 as well as in the related S. flexneri Y
series,16,17 a retrosynthetic analysis of 1 showed that the
most advantageous disconnection would be at the CD
linkage. Such a strategy would involve a trisaccharide
donor bearing a 2-O-acyl group at the reducing residue
that could direct the glycosylation towards the desired
stereochemistry and a pentasaccharide acceptor bearing
a free hydroxyl group at position 3 of a potential
N-acetyl glucosaminyl residue.
The fully protected pentasaccharide 2, which was
described recently,14 was chosen as a suitable precursor
to the latter. Previous work in the series,15 supported by
independent observations,18 indicated that the benzoyl
group at position 2C was particularly stable and reacted
poorly under Zemple´n conditions. In addition, mild
transesterification conditions were expected to prevent
the migration of the trichloroacetyl (TCA) group to the
vicinal 3D-OH. Taking advantage of these observations,
compound 2 was, indeed, selectively de-O-acetylated
into the triol 3 (Scheme 1), which was subsequently
blocked at positions 4D and 6D by an isopropylidene
At this stage, the key glycosylation step involving
acceptor 4 and donor 5 was investigated with the aim of
finding standard and reproducible conditions allowing
easy access to the fully protected octasaccharide 10
(Scheme 3). The demonstration in this laboratory of the
efficiency of the BF3·OEt2 complex as an alternative to
the less satisfactory TMSOTf catalyst in the synthesis
of related oligosaccharides15 encouraged the use of the
former in the present work. However, reaction of 4 and
5 in the presence of the BF3·OEt2 complex led to a
major condensation product (82%) lacking the iso-
1
propylidene group as seen from the H and 13C NMR
spectra. This side-reaction was avoided when using
triflic acid as the catalyst. In the latter case, the fully
protected octasaccharide 10 was isolated in a satisfac-
tory yield (85%).20 Acidic hydrolysis of the isopropyl-
idene group gave access to the diol 11 (90%), the
availability of which allowed further investigation of
the BF3·OEt2-mediated condensation of 4 and 5. Com-
parison of the NMR data for 11 and the unknown diol
was in favor of identity for the two octasaccharides.
This assumption was further supported by analysis of
Scheme 1. (a) cat. MeONa, MeOH, 30 min; (b) 2-methoxy-
propene, CSA, DMF (72% from 2).
Scheme 2. (a) cat. TMSOTf, anhydrous Et2O, 4 h, −50°C (95%); (b) i. cat. [Ir(COD){PCH3(C6H5)2}2]+PF6−, THF, rt, 20 h, ii.
HgO, HgCl2, acetone/water, rt, 2 h; (c) CCl3CN, DBU, CH2Cl2, 0°C, 1 h (77% from 8).

F. Be´lot et al. / Tetrahedron Letters 43 (2002) 8215–8218
8217
,
,
Scheme 3. (a) cat. TfOH, 4 A MS, 1,2-DCE, −35“+10°C, 2.5 h (85%); (b) BF3·OEt2, 4 A MS, DCM, −60°C“rt, 24 h (11, 82%);
(c) 50% aq. TFA, CH2Cl2, 0°C, 4 h (90%); (d) Bu3SnH, AIBN, PhCH3, 80°C, 2 h; (e) i. MeONa, MeOH, ii. H2O, 60°C, 16 h,
iii. Ac2O, rt, 16 h; (f) cat. MeONa, MeOH, 55°C, 24 h (80%); (g) i. H2, Pd/C, EtOH/AcOEt/1 M aq. HCl (2 equiv.), 72 h, ii. H2,
Pd/C, Et3N, MeOH, 24 h (77%).
from readily available precursors, a strategy based on a
disconnection at the CD linkage and the use of stan-
dard glycochemistry is appropriate for the synthesis of
complex oligosaccharides in the S. flexneri 2a series.
the acetylation product of the unknown diol. A switch
from 3.3 to 3.7 ppm, respectively, was observed for
both H-6D in the corresponding 1H NMR spectra.
Thus, it is assumed that the isolation of 11 in the
BF3·OEt2-mediated glycosylation is a two-step process
involving first the condensation of 4 and 5, and second
the hydrolysis of the acetal protecting group. It is
expected that a better control of the kinetics of the
reaction would avoid the deblocking step. Conversion
of the N-TCA participating group into the required
N-acetyl group was found to be a rather challenging
task. As observed previously in the series, the use of
radical conditions21,22 yielding a mixture of partially
dechlorinated products failed to give the expected 12.
Attempts to convert 11 into 13 by a two-step process
involving alkaline removal of the TCA group and in
situ N-acetylation following a recently described proce-
dure,23 yielded again an unworkable mixture of com-
pounds. The outcome of the latter reaction was
tentatively blamed on the presence of the benzoyl
groups at position 2C and 2C% which, as shown earlier in
analogous compounds, were expected to be rather resis-
tant to hydrolysis due to steric hindrance. Conse-
Acknowledgements
The authors thank Professor P. J. Sansonetti who is a
scholar of the Howard Hughes Medical Institute for his
key input in the project. The authors are grateful to J.
Ughetto-Monfrin (Unite´ de Chimie Organique, Institut
Pasteur) for recording all the NMR spectra. The
authors thank the CANAM, the Fondation pour la
Recherche Me´dicale, the Bourses Mrs. Frank Howard
Foundation and the Bourses Roux foundation for
financial support.
References
1. Part 9 of the series, Synthesis of ligands related to the
O-SP of Shigella flexneri serotype 2a and Shigella flexneri
serotype 5a. For part 8, see Ref. 15.
quently, diol 11 was engaged in
a
controlled
2. Sansonetti, P. Rev. Infect. Dis. 1991, 13, S285–S292.
3. Mulard, L.; Alonso, J. M.; Fournier, J. M. In La Vac-
cinologie; Annales de l’Institut Pasteur/actualite´s; Cohen,
O-deacylation process to give the heptaol 14 (80%).
The latter was next submitted to an efficient, two-step,
in-house process involving first, hydrogenolysis under
acidic conditions which allowed the removal of the
benzyl groups and second, basic hydrochlorination
which resulted in the conversion of the trichloroac-
etamide intermediate into the target octasaccharide 1
(78%).
´
G. N.; Sansonetti, P. Eds.; Editions scientifiques et me´d-
icales; Elsevier SAS: Paris, 2002; pp. 39–54.
4. Wenger, J. D.; Booy, R.; Health, P. T.; Moxon, R. In
New Generation Vaccines; 2nd ed., revised and expanded
ed.; Levine, M. M., Woodrow, G. C., Kaper, J. B.,
Cobon, G. S., Eds.; Marcel Dekker: New York, 1997; pp.
489–502.
5. MacLennan, J. M.; Shackley, F.; Heath, P. T.; Deeks, J.
J.; Flamank, C.; Herbert, M.; Griffith, H.; Hatzmann, E.;
Goilav, C.; Moxon, E. R. J. Am. Med. Assoc. 2000, 283,
2795–2801.
The structure of all intermediates and of the target 1
1
were unambiguously confirmed by H and 13C NMR
analysis as well as by mass spectrometry and elemental
analysis. For selected data see Ref. 24. The results
described herein support our hypothesis that, starting

8218
F. Be´lot et al. / Tetrahedron Letters 43 (2002) 8215–8218
6. Robbins, J. B.; Chu, C.; Watson, D. C.; Szu, S. C.;
J
2,NH=8.2 Hz, NHD), 5.38 (d, 1H, J1,2<1.0 Hz, J2,3=2.8
Daniels, E. M.; Lowe, C. U.; Schneerson, R. Rev. Infect.
Dis. 1991, 13, S362–365.
7. Robbins, J. B.; Schneerson, R.; Szu, S. C. In New Gener-
ation Vaccines; Levine, M. M.; Woodrow, G. C.; Kaper,
J. B.; Cobon, G. S., Eds.; Marcel Dekker, 1997; pp.
803–815.
Hz, H-2C), 5.27 (d, 1H, J1,2<1.0 Hz, J2,3=2.8 Hz, H-2C),
5.15 (m, 2H, J1,2=1.0 Hz, J2,3=2.8 Hz, H-2B and H-1B),
5.08 (dd, 1H, J2,3=2.8 Hz, J3,4=10.0 Hz, H-3B), 5.01 (d,
1H, J1,2<1.0 Hz, H-1C), 4.96 (d, 1H, J1,2<1.0 Hz, H-1A),
4.94 (d, 1H, J1,2<1.0 Hz, H-1C), 4.90 (d, 1H, J1,2=3.3 Hz,
H-1E), 4.85 (d, 1H, J1,2=3.3 Hz, H-1E), 4.82 (dd, 1H,
8. Pozsgay, V. In Adv. Carbohydr. Chem. Biochem.; Horton,
D., Ed.; Academic Press: San Diego, 2000; Vol. 56, pp.
153–199.
9. Simmons, D. A. R. Bacteriol. Rev. 1971, 35, 117–148.
10. Lindberg, A. A.; Karnell, A.; Weintraub, A. Rev. Infect.
Dis. 1991, 13, S279–S284.
11. Berry, J. M.; Dutton, G. G. S. Can. J. Chem. 1974, 54,
681–683.
12. Lipkind, G. M.; Shashkov, A. S.; Nikolaev, A. V.;
Mamyan, S. S.; Kochetkov, N. K. Bioorg. Khim. 1987,
13, 1081–1092.
J
3,4=J4,5=10.0 Hz, H-4B), 4.80 (d, 1H, J1,2=8.5 Hz,
H-1D), 4.68 (d, 1H, J1,2<1.0 Hz, H-1B), 3.29 (s, 3H,
OCH3), 1.98, 1.78, 1.68 (3s, 9H, OAc), 1.26–0.85 (m,
15H, H-6A, 2 H-6B, 2 H-6C), 1.18, 1.16 (2s, 6H, (CH3)2).
13C: l 168.5, 168.2, 168.1, 164.4, 164.1, 160.3 (6C, C=O),
101.4 (bs, 2C, 2 C-1C), 101.2 (C-1A), 101.0 (C-1D), 99.8
(C(CH3)2), 98.4 (bs, 2C, 2 C-1E), 98.2 (C-1B), 97.4 (C-1B),
91.0 (CCl3), 62.1 (C-6D), 59.3 (C-2D), 55.5 (OCH3), 29.4
(C(CH3)2), 21.3, 21.0, 20.9 (3C, OAc), 19.46 (C(CH3)2),
19.2, 19.1, 18.6, 18.2, 17.5 (5C, C-6A, 2 C-6B, 2 C-6C).
FAB MS for C158H174Cl3NO41 (M, 2849.4), m/z 2872.5
([M+Na]+). Anal. calcd for C158H174Cl3NO41 C, 66.60; H,
6.16; N, 0.49; found: C, 66.53; H, 6.28; N, 0.43.
13. Mulard, L. A.; Costachel, C.; Sansonetti, P. J. J. Carbo-
hydr. Chem. 2000, 19, 849–877.
14. Costachel, C.; Sansonetti, P. J.; Mulard, L. A. J. Carbo-
hydr. Chem. 2000, 19, 1131–1150.
15. Segat, F.; Mulard, L. A. Tetrahedron: Asymmetry 2002,
in press.
16. Pinto, B. M.; Reimer, K. B.; Morissette, D. G.; Bundle,
D. R. J. Chem. Soc. Perkin Trans. 1 1990, 293–299.
17. Pinto, B. M.; Reimer, K. B.; Morissette, D. G.; Bundle,
D. R. J. Org. Chem. 1989, 54, 2650–2656.
18. Varga, Z.; Bajza, I.; Batta, G.; Liptak, A. Tetrahedron
Lett. 2001, 42, 5283–5286.
19. Kitagawa, I.; Baek, N. I.; Ohashi, K.; Sakagami, M.;
Yoshikawa, M.; Shibuya, H. Chem. Pharm. Bull. 1989,
37, 1131–1133.
21. Blatter, G.; Beau, J.-M.; Jacquinet, J.-C. Carbohydr. Res.
1994, 260, 189–202.
22. Be´lot, F.; Jacquinet, J.-C. Carbohydr. Res. 2000, 326,
88–97.
23. Scherman, A. A.; Yudina, O. A.; Minorov, Y. V.;
Sukhova, E. V.; Shashkov, A. S.; Menshov, V. M.;
Nifantiev, N. E. Carbohydr. Res. 2001, 336, 13–46.
24. Selected data for 1: 1H NMR (400 MHz, D2O): l 5.12 (d,
1H, J1,2=3.5 Hz, H-1E), 5.09 (d, 1H, J1,2=3.5 Hz, H-1E),
5.06 (d, 1H, J1,2<1.0 Hz, H-1A), 4.98 (d, 1H, J1,2<1.0 Hz,
H-1B), 4.88 (d, 1H, J1,2<1.0 Hz, H-1C), 4.76 (d, 1H,
J1,2<1.0 Hz, H-1B), 4.65 (d, 1H, J1,2=8.5 Hz, H-1D), 4.58
(d, 1H, J1,2=2.0 Hz, H-1C), 3.32 (s, 3H, OCH3), 2.00 (s,
3H, AcNH), 1.35–1.19 (m, 15H, H-6A, 2 H-6B, 2 H-6C).
13C NMR: l 103.3 (C-1C), 102.6 (C-1D), 101.7 (C-1A),
101.4 (C-1B), 101.1 (C-1B), 100.8 (C-1C), 98.6 (C-1E), 98.4
(C-1E), 61.0, 60.9, 60.8 (3C, C-6D, 2 C-6E), 55.9 (C-2D),
55.3 (OCH3), 22.7 (AcNH), 18.3, 18.0, 17.1, 16.9, 16.8
20. Detailed procedure: A mixture of 5 (150 mg, 0.12 mmol)
,
and 4 (150 mg, 83 mmol), 4 A molecular sieves and dry
1,2-DCE was stirred for 1 h then cooled to −35°C. Triflic
acid (5 ml) was added. The stirred mixture was allowed to
reach 10°C in 2 h 30 min. Et3N was added and the
mixture filtered. After evaporation, chromatography of
the residue afforded 10 (208 mg, 85%). Selected data for
10: 1H NMR (400 MHz, CDCl3): l 1H: l 6.82 (d, 1H,
(5C, C-6A,
2 C-6B, 2 C-6C). HRMS: calcd for
C51H87NO36+Na: 1312.4906. Found: 1312.4902.