De novo asymmetric synthesis of the anthrax tetrasaccharide by a palladium-catalyzed glycosylation reaction

Article abstract of DOI:10.1002/anie.200701354

(Chemical Equation Presented) Choose your poison: In the de novo asymmetric synthesis of the anthrax tetrasaccharide 1 (25 steps from acetylfuran), the absolute configurations of three L sugars and one D sugar were set by the Noyori reduction of acetylfuran with the S,S and the R,R reagent, respectively. The configuration of the remaining carbohydrate unit was set by highly diastereoselective palladium-catalyzed glycosylation and post-glycosylation transformations.

Full text of DOI:10.1002/anie.200701354

Communications
DOI: 10.1002/anie.200701354
Tetrasaccharide Synthesis
De Novo Asymmetric Synthesis of the Anthrax Tetrasaccharide by a
Palladium-Catalyzed Glycosylation Reaction**
Haibing Guo and George A. OꢀDoherty*
Anthrax is a zoonotic disease caused by the spore-forming
bacterium Bacillus anthracis.[1] Bacillus anthracis belongs to
the family Bacillaceae, which consists of a diverse group of
bacteria, all of which form endospores.[2] Bacillus anthracis is
the most important member of this genus and is considered to
be a potent agent for biological warfare. Its protective
polypeptide capsule consists of poly-d-glutamic acid, which
inhibits phagocytosis.[3] Recently, a tetrasaccharide made up
of three l-rhamnose sugars and a rare sugar, d-anthrose, was
isolated from the capsule (Scheme 1).[4] The uniqueness of the
describe our successful application of this methodology for
the de novo asymmetric synthesis of 1.[8]
In our retrosynthetic analysis we envisioned 1 as being
prepared from tetrasaccharide 4, which in turn could be
prepared by a glycosylation of trisaccharide 5 with phosphate
6 (Scheme 2). At the outset, we hoped to use our de novo
approach to prepare both of these fragments (5 and 6) from
the achiral acetylfuran 7, which is significantly cheaper than
either l-rhamnose (2) or d-fucose (3).[9]
Scheme 1. Anthrax tetrasaccharide 1.
Scheme 2. Retrosynthesis of anthrax tetrasaccharide 1. Bn=benzyl,
LevOH=levulinic acid
anthrose sugar and the resistance of carbohydrate structures
to evolutionary change make the anthrax tetrasaccharide 1 an
interesting target for anthrax detection and vaccine develop-
ment.[5] Thus, synthetic access to this tetrasaccharide is
desired.
Recently, two carbohydrate-based approaches to the
anthrax tetrasaccharide and one to a related trisaccharide
have been reported.[6] In these routes the stereochemistry is
derived from the known but less common sugar l-rhamnose
(2) and the rare d-fucose (3). In contrast to these traditional
approaches, we have been investigating de novo asymmetric
approaches to mono, di-, and oligosaccharides.[7] Herein we
Our synthesis of the anthrose portion of the tetrasacchar-
ide commenced with the Noyori reduction of the acetylfuran
7 to install the d stereochemistry (Scheme 3). Subsequent
Achmatowicz rearrangement (NBS/H2O) and diastereoselec-
tive (a/b = 3:1) Boc protection ((Boc)2O/DMAP) provided
pyranone 8 in 60% overall yield.[10] Exposure of the pyranone
8 and p-methoxybenzyl alcohol to our palladium glycosyla-
tion conditions (0.5% Pd0/1% PPh3) produced PMB-pyra-
none 9 in excellent yield (96%) as a single diastereomer.
Luche reduction (NaBH4/CeCl3) of pyranone 9 followed by
methyl carbonate formation (ClCO2CH3/DMAP) produced
allylic carbonate 10 in 92% yield for the two steps.[11] The
methyl carbonate group of 10 was regio- and stereoselectively
replaced with an azide group by a Pd allylation (TMSN3,
[{(allyl)PdCl}2], dppb) to afford allylic azide 11 (93%).[12]
Dihydroxylation of 11 under Upjohn conditions (OsO4/
NMO) installed the manno stereochemistry, and regioselec-
tive protection (BnBr/Bu2SnO) provided benzyl ether 12
(97%).[13] Finally the axial hydroxy group at C2 in 12 was
converted to give gluco stereochemistry by an SN2 displace-
ment. Thus, alcohol 12 was treated with triflic anhydride and
inverted to give the equatorial alcohol 13 with NaNO2
(56%).[14] Acylation of 13 (LevOH/DCC/DMAP) and
[*] H. Guo, Prof. G. A. O’Doherty
Department of Chemistry
West Virginia University
217 Clark Hall, P.O. 6045
Morgantown, WV 26506 (USA)
Fax: (+1)304-293-4904
E-mail: george.odoherty@mail.wvu.edu
Homepage: http://www.as.wvu.edu/~odoherty/site/
[**] We thank the NIH (GM63150) and NSF (CHE-0415469) for their
generous support of our research program. Funding for a 600-MHz
NMR spectrometer by the NSF-EPSCoR (grant 0314742) is also
gratefully acknowledged.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
5206
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 5206 –5208

Angewandte
Chemie
Scheme 3. Synthesis of anthrose monosaccharide 6 and 15. Noyori
(R,R)=(R)-Ru(h6-mesitylene)-(R,R)-N-(4-toluenesulfonyl)-1,2-
diphenylethylenediamine, NBS=1-bromo-2,5-pyrollidinedione,
PMBOH=p-methoxybenzyl alcohol, (Boc)2O=di-tert-butyl dicarbon-
ate, TMSN3 =trimethylsilyl azide, dppb=1,4-bis(diphenylphosphino)-
butane, NMO=N-methyl morpholine-N-oxide, Tf2O=trifluoro-
methylsulfonic anhydride, DCC=N,N’-dicyclohexylcarbodiimide,
DMAP=4-dimethylaminopyridine, DDQ=2,3-dichloro-5,6-dicyano-1,4-
benzoquinone, TCA=trichloroacetic imidate.
Scheme 5. Synthesis of trisaccharide 23.
removal of the PMB group provided the anomeric alcohol 14
(89%). Finally two anthrose precursors were prepared from
14, phosphate 6 (89%) and imidate 15 (83%).
With the d-anthrose monosaccharide in hand, we turned
to the synthesis of the tris-l-rhamno trisaccharide 5, which
required an l-rhamno sugar with an unprotected C2 hydroxy
group (17, Scheme 4). Analogously, the l-pyranone ent-8 was
yield.[17] In an analogous fashion the final rhamno sugar in 23
was installed at the C3 hydroxy group of 20 (Pd glycosylation
(20 + ent-8!21), Luche reduction and dihydroxylation, 79%
yield of 22) and by orthoester chemistry the C2/C4 hydroxy
groups of 22 were selectively acylated (MeC(OMe)3/TsOH;
Ac2O; AcOH/H2O, 97%) to form diacetate 23 in good overall
yield (77% for four steps).[17]
Unfortunately, our attempts at glycosylation of 23 with
either the phosphate 6 or imidate 15 failed. Instead only
hydrolysis of the spiroketal protecting group was observed.
Thus, we decided to prepare the more acid-stable trisacchar-
ide 5, which could be prepared in four steps from 23
(Scheme 6). Acylation of 23 (LevOH/DCC/DMAP), fol-
lowed by removal of the spiroketal (TFA/H2O) provided
diol 24. The two hydroxy groups were acetylated (Ac2O/Py)
and the levulinate group was selectively deprotected
(H2NNH3OAc) to produce trisaccharide 5 (91% from 23).
Our return to the final glycosylation step with the acid-
stable rhamno trisaccharide 5 was more successful leading to
the synthesis of the anthrax tetrasaccharide 1 (Scheme 7). In
contrast to 23, exposure of 5 to either imidate 15 or phosphate
6 and catalytic amounts of TMSOTf produced tetrasaccharide
4 in good yields (85% for 15 and 90% for 6).[18] The anthrose
Scheme 4. Pd0-catalyzed glycosylation synthesis of 17. p-TsOH=p-
toluenesulfonic acid.
prepared in three steps from acetylfuran 7 by simply switching
to the (S,S)-Noyori catalyst. By using our Pd-glycosylation
procedure (BnOH, 0.25% Pd0/0.5% Ph3P), we protected the
anomeric position of pyranone ent-8 as a benzyl ether (90%
yield). A post-glycosylation Luche reduction and dihydrox-
ylation installed the rhamno triol 16 (75%, overall yield).[15]
Finally the equatorial hydroxy groups of 16 at C3 and C4 were
selectively protected using the Ley spiroketal procedure
yielding 17 (66% from ent-8) which has a free axial hydroxy
group at C2 for subsequent glycosylation.[16]
Palladium-catalyzed glycosylation of the axial hydroxy
group at C2 in 17 with pyranone ent-8 provided the pyranone
18 in 86% yield (Scheme 5). Once again a Luche reduction
and Upjohn dihydroxylation diastereoselectively produced
the rhamno triol 19 (86%, two steps). The triol 19 was treated
with trimethyl orthoacetate and catalytic p-toluenesulfonic
acid to form a cyclic orthoester intermediate, which sub-
sequently underwent acetylation at C4 and regioselective
hydrolytic opening to afford the 2,4-diacetate 20 in 97%
Scheme 6. Synthesis of trisaccharide 5. TFA=trifluoroacetic acid,
Py=pyridine.
Angew. Chem. Int. Ed. 2007, 46, 5206 –5208
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5207

Communications
[4] J. M. Daubenspeck, H. Zeng, P. Chen, S. Dong, C. T. Steichen,
N. R. Krishna, D. G. Pritchard, C. L. Turnbough, Jr., J. Biol.
Chem. 2004, 279, 30945 – 30953.
[5] M. Tamborrini, D. B. Werz, J. Frey, G. Pluschke, P. H. Seeberger,
Angew. Chem. 2006, 118, 6731 – 6732; Angew. Chem. Int. Ed.
2006, 45, 6581 – 6582.
[6] There have been two syntheses of the anthrax tetrasaccharide
with an anomeric protecting group at the reducing end. a) D. B.
Werz, P. H. Seeberger, Angew. Chem. 2005, 117, 6474 – 6476;
Angew. Chem. Int. Ed. 2005, 44, 6315 – 6318; b) R. Adamo, R.
Saksena, P. Kovac, Carbohydr. Res. 2005, 340, 2579 – 2582; c) R.
Saksena, R. Adamo, P. Kovac, Bioorg. Med. Chem. Lett. 2006,
16, 615 – 617. Of the two syntheses that in Ref. [6a] is the shorter.
For the related trisaccharide, see: d) A. S. Mehta, E. Saile, W.
Zhong, T. Buskas, R. Carlson, E. Kannenberg, Y. Reed, C. P.
Quinn, G.-J. Boons, Chem. Eur. J. 2006, 12, 9136 – 9149.
[7] a) R. S. Babu, M. Zhou, G. A. OꢀDoherty, J. Am. Chem. Soc.
2004, 126, 3428 – 3429; b) M. Zhou, G. A. OꢀDoherty, Org. Lett.
2006, 8, 4339 – 4342.
[8] While there are many uses of the term “de novo” in carbohy-
drate chemistry, we use the term de novo asymmetric synthesis
to refer to the use of catalysis for the asymmetric synthesis of
carbohydrates from achiral compounds (e.g., this de novo route
derives its asymmetry from enantiodivergent Noyori reductions
of acetylfuran 7).
Scheme 7. Completion of the synthesis of anthrax tetrasaccharide 1.
TMSOTf=trimethylsilyl trifluoromethanesulfonate, HBTU=O-benzo-
triazole-N,N,N’,N’-tetramethyluronium-hexafluorophosphate
[9] Sigma-Aldrich sells acetylfuran 7 for $0.09gꢀ1, l-rhamnose (2)
for $5gꢀ1 and d-fucose (3) for $71gꢀ1, see: www.sigmaaldrich.-
com.
methyl ether was installed in 25 by selective levulinate
hydrolysis (H2NNH3OAc, 96%) and silver(I) oxide promoted
methylation (Ag2O in MeI, 94%).[19] A one-pot global
deprotection of the acetate in 25 along with azide reduction
afforded a primary amine (PEt3/LiOH/H2O, 95%), which was
selectively coupled with 3-hydroxy-3-methylbutanoic acid
(HBTU/Et3N, 90%) to give amide 26. Finally the natural
product 1 was prepared by hydrogenolysis of both benzyl
groups (H2, Pd/C) in good yield (96%).[20]
In summary, a de novo asymmetric synthesis of the natural
product anthrax tetrasaccharide 1 has been developed in 25
steps (longest linear, 39 total steps) 13% overall yield from
achiral acetylfuran 7. This highly stereocontrolled route
provides sufficient quantities of 1 for further studies. While
this route is longer than the Seeberger approach in terms of
longest linear sequence (20 steps and 7% overall yield from
d-fucose (3)), it is shorter in terms of total steps (41 total
steps). Thus we demonstrate the practicality of de novo
approaches for oligosaccharide synthesis.[8] Further applica-
tion of this approach to the preparation of an anthrax vaccine
and detection device is ongoing.
[10] a) M. Li, J. G. Scott, G. A. OꢀDoherty, Tetrahedron Lett. 2004, 45,
1005 – 1009; b) R. S. Babu, G. A. OꢀDoherty, J. Carbohydr.
Chem. 2005, 24, 169 – 177; c) H. Guo, G. A. OꢀDoherty, Org.
Lett. 2005, 7, 3921 – 3924.
[11] a) J. L. Luche, J. Am. Chem. Soc. 1978, 100, 2226 – 2227; b) M. H.
Haukaas, G. A. OꢀDoherty, Org. Lett. 2001, 3, 401 – 404.
[12] a) R. N. de Oliveira, L. Cottier, D. Sinou, R. M. Srivastava,
Tetrahedron 2005, 61, 8271 – 8281; b) H. Guo, G. A. OꢀDoherty,
Org. Lett. 2006, 8, 1609 – 1612.
[13] V. VanRheenen, R. C. Kelly, D. Y. Cha, Tetrahedron Lett. 1976,
17, 1973 – 1976.
[14] a) R. Saksena, R. Adamo, P. Kovac, Carbohydr. Res. 2005, 340,
1591 – 1600; b) D. W. C. Jones, R. J. Nash, E. A. Bell, J. M.
Williams, Tetrahedron Lett. 1985, 26, 3125 – 3126.
[15] While in principle 16 can be prepared by Fischer glycosylation,
in practice for moderately expensive sugars, like l-rhamnose, a
four-step protocol is preferred: 1) peracylation, 2) selective
hydrolysis of the anomeric position with HBr, 3) Ag- or Hg-
promoted glycosylation of the anomeric bromide, and 4) LiOH
global ester hydrolysis]; see Ref. [6].
[16] a) S. V. Ley, M. Woods, A. Zanotti-Gerosa, Synthesis 1992, 52 –
54; b) S. V. Ley, R. Leslie, P. D. Tiffin, M. Woods, Tetrahedron
Lett. 1992, 33, 4767 – 4770.
[17] See Ref. [7b] and a) J. F. King, A. D. Allbutt, Can. J. Chem. 1970,
48, 1754 – 1769; b) Y. E. Tsvetkov, A. S. Shashkov, Y. A. Knirel,
U. Zahringer, Carbohydr. Res. 2001, 335, 221 – 243.
[18] While both reactions worked well, we had a preference for the
phosphate glycosylation (5 + 6), see: a) O. J. Plante, E. R.
Palmacci, R. B. Andrade, P. H. Seeberger, J. Am. Chem. Soc.
2001, 123, 9545 – 9554; b) R. R. Schmidt, W. Kinzy, Adv.
Carbohydr. Chem. Biochem. 1994, 50, 21 – 123.
Received: March 28, 2007
Published online: May 30, 2007
Keywords: anthrax tetrasaccharide · carbohydrates ·
.
glycosylation · natural products · palladium
[1] a) P. S. Brachman, A. F. Kaufmann in Bacterial Infections of
Humans (Eds.: A. S. Evans, P. S. Brachman), Plenum Medical
Book Company, New York, 1998, pp. 95 – 111.
[19] Significantly lower yields were observed when cosolvents were
used. Similarly low yields were also observed by Boons et al., see
Ref. [6d].
[20] Since the spectral data for synthetic 1 matches that of the
isolated material, this constitutes the first synthesis of 1; see
Ref. [6].
[2] a) M. Mock, A. Fouet, Annu. Rev. Microbiol. 2001, 55, 647 – 671;
b) P. Sylvestre, E. Couture-Tosi, M. Mock, Mol. Microbiol. 2002,
45, 169 – 178.
[3] a) W. L. Nicholson, N. Munakata, G. Horneck, H. J. Melosh, P.
Setlow, Microbiol. Mol. Biol. Rev. 2000, 64, 548 – 572; b) M.
Moayeri, S. H. Leppla, Curr. Opin. Microbiol. 2004, 7, 19 – 24.
5208 www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 5206 –5208