Short de novo synthesis of fully functionalized uronic acid monosaccharides

Article abstract of DOI:10.1002/anie.200502742

(Chemical Equation Presented) Short and Sweet: A convergent route leads to orthogonally protected D-glucuronic and L-iduronic acid thioglycoside building blocks, commonly used in heparin oligosaccharide assembly (see scheme). Rapid access to substantial q

Full text of DOI:10.1002/anie.200502742

Angewandte
Chemie
A typical monosaccharide building block used in oligo-
saccharide assembly is equipped with different protecting
groups that mask the hydroxy and amine functions and an
anomeric leaving group that can be activated to induce the
formation of a glycosidic linkage. These differentially pro-
tected and functionalized monosaccharides have traditionally
been accessed from naturally occurring sugar starting materi-
als through a series of protection-deprotection maneuvers.
Such a process establishes the desired protecting group
pattern and typically requires 6–20 steps depending on the
sugar, the protecting group pattern, and the anomeric leaving
group.[4]
To avoid these lengthy procedures for producing building
blocks for the assembly of larger structures, efforts have been
directed at the synthesis of orthogonally protected hexose
monosaccharides from noncarbohydrate precursors.[5] The
landmark syntheses of hexoses by Masamune, Sharpless, and
co-workers[6] were enabled by the ability of new synthetic
methods to introduce specific stereogenic centers selectively.
The aldol reaction has also proved to be a particularly useful
tool in the creation of hexoses from simple precursors.
Mukaiyama and co-workers[7] reported a stereoselective
synthesis of pentoses and hexoses through the use of silyl
enol ethers in aldol condensations to furnish partially
protected pentoses and hexoses. Unfortunately, these funda-
mental explorations were not investigated further than the
initial proof of concept. Recently, proline-catalyzed aldol
reactions were used to fashion aldo- and ketohexose pre-
cursors for Mukaiyama-type aldol reactions, which were then
used to synthesize partially protected glucose, mannose, and
allose monosaccharides.[8] The yields and selectivities
reported for these transformations, although excellent, were
highly dependent on the specific protecting groups that were
chosen.
Oligosaccharide assembly requires monosaccharide build-
ing blocks that contain orthogonal protecting-group patterns
and a readily activated anomeric leaving group. Herein we
report a convergent route to orthogonally protected d-
glucuronic and l-iduronic acid thioglycoside building blocks,
which are commonly used in the assembly of heparin
oligosaccharides.[9] This approach relies on the selective
Mukaiyama-type aldol reaction[10] that unifies a silyl enol
ether and a thioacetal-containing aldehyde (derived from the
chiral pool).
Retrosynthetic analysis of the uronic acids A (Scheme 1)
revealed that the fully protected uronic acid thioglycosides
could be obtained through cyclization of the linear hexoses B.
The open-chain hexoses can be formed, in turn, through a
Mukaiyama-type aldol reaction of an appropriately protected
ketene acetal C with a thioacetal-containing aldehyde D.
Synthesis of the key thioacetal aldehydes 8 and 9
commenced with readily available l-arabinose (1;
Scheme 2). The aldehyde was converted into the correspond-
ing thioacetal by reaction with ethanethiol in the presence of
hydrochloric acid.[11] Subsequent protection of the 4,5-diol
with 2,2-dimethoxypropane furnished the crystalline aceto-
nide 2.[12] At this stage, the protecting-group patterns for the
C2 and C3 hydroxy groups were selected. The reaction of 2
with NaH, in the presence of excess benzyl bromide, afforded
Carbohydrate Synthesis
DOI: 10.1002/anie.200502742
Short De Novo Synthesis of Fully Functionalized
Uronic Acid Monosaccharides**
Mattie S. M. Timmer, Alexander Adibekian, and
Peter H. Seeberger*
Dedicated to Professor Albert Eschenmoser
on the occasion of his 80th birthday
The biological importance of carbohydrates in a host of
fundamental cellular processes has dramatically increased the
demand for pure, synthetically derived oligosaccharides.[1]
Over the past 100 years, chemical methods for the assembly
of all classes of carbohydrates have been developed. This has
resulted in sophisticated glycosylation reactions that allow the
formation of even the most difficult glycosidic linkages with a
high degree of selectivity.[2] More recently, automated meth-
ods for the rapid combination of monosaccharide building
blocks on a solid phase were reported.[3] The time required to
assemble an oligosaccharide has been reduced more than 100-
fold, and as a result, the need for substantial quantities of and
rapid access to fully functionalized building blocks has
increased.
[*] Dr. M. S. M. Timmer, A. Adibekian, Prof. Dr. P. H. Seeberger
Laboratory for Organic Chemistry
Swiss Federal Institute ofTechnology (ETH) Zürich
ETH Hönggerberg HCI F315
Wolfgang-Pauli-Strasse 10, 8093 Zürich (Switzerland)
Fax: (+41)44-633-1235
E-mail: seeberger@org.chem.ethz.ch
[**] This research was supported by the ETH Zürich, a Niels–Stensen
Postdoctoral Fellowship (to M.T.) and a KekulØ Fellowship from the
Fonds dre Chemischen Industrie (to A.A.). We thank Professor D.
Seebach for helpful discussions.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2005, 44, 7605 –7607
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
7605

Communications
BF3·Et2O mediated aldol reaction[13] of aldehyde 8 with silyl
enol ether 10[14] gave a 1:1:1 mixture of three products as was
determined by analysis of the 1H NMR spectrum of the crude
product mixture. The three individual aldol products were
isolated by column chromatography and identified after
conversion into the corresponding thiopyranosides. The free
hydroxy groups of 11, 14, and 17 were first transformed into 9-
fluorenylmethyl carbonates followed by cleavage of the silyl
ether in the presence of pyridine·HF to obtain the alcohols 12,
15, and 18, respectively, in good overall yields. Although
initial attempts to effect cyclization under various acidic
conditions proved to be challenging, NIS-promoted activation
of the thioacetal led to the formation of
Scheme 1. Retrosynthetic analysis ofuronic acid thioglycosides.
PG=protecting group, TMS=trimethylsilyl.
pyranose carbohydrates 13, 16, and 19,
respectively, in quantitative yields.
At this stage, the absolute configurations
of the pyranoses were established by the
1H NMR coupling patterns of the carbohy-
drate ring protons. For compound 13, the
large coupling constants (J = 8–10 Hz)
between 2-H, 3-H, 4-H, and 5-H provided
the evidence to assign the gluco configura-
tion. The ido configuration of 16 was
assigned based on the characteristically
small coupling constants (J = 1–3 Hz) for
protons 2-H to 5-H. This configuration was
further corroborated by the W coupling,
4J2,4 = 1.0 Hz, which is typical of a 2,4-diaxial-
substituted pyranose. The coupling pattern
for the less common altruronic acid 19 was in
full accordance with the reported coupling
constants for altroses.[15] Notably, no trace of
the galacto uronic acid, the fourth possible
isomer, was found. This absence can be
Scheme 2. Synthesis of differentially protected thioacetal aldehydes from l-arabinose:
a) EtSH, conc. aq. HCl, 10 min, 77%; b) 2,2-dimethoxypropane, pyridinium p-toluenesulfo-
nate (cat.), acetone, 1.5 h, 81%; c) BnBr, TBAI (cat.), NaH, DMF, 08C, 4 h; d) nBu2SnO,
toluene, Dean–Stark trap; then BnBr, CsF, TBAI (cat.), DMF; e) Ac2O, pyridine, 46% (two
steps); f) AcOH/H2O (1:1, v/v), 508C, 1 h, 6: 62% (two steps: c and f), 7: 92%; g) NaIO4,
H2O/THF, 08C, 15 min, 8: 82%, 9: 80%. Bn=benzyl, DMF=N,N-dimethylformamide,
TBAI=tetrabutylammonium iodide.
dibenzylarabinoside 3, whereas formation of a tin ketal of the
diol 2 and subsequent monobenzylation gave 4. After
acetylation of the remaining hydroxy group in 4, selectively
protected 5 was isolated in 46% overall yield together with
45% of the 2-O-benzyl-3-O-acetyl regioisomer. Deprotection
of the acetal groups in 3 and 5 was followed by sodium
periodate mediated cleavage
rationalized through the assumption of a nonchelating,
ꢀ
open-chain transition state that allows only the three C C
bond formations observed and not the sterically demanding
Si, Si attack that would lead to the galacto isomer.
Once the absolute configurations of the aldol products
had been determined, the selectivity of the aldol reaction was
of the resulting vicinal diols
to form the desired alde-
hydes 8 and 9 in very good
overall yields. Notably, these
two key intermediates are
derived from cheap, com-
mercially available starting
materials
through straightforward and
high-yielding transforma-
(l-arabinose)
tions that are readily scala-
ble.
After this efficient route
for the synthesis of thioace-
tal aldehydes was estab-
lished, the aldol reaction/
cyclization sequence was
explored (Scheme 3). The
Scheme 3. Synthesis of l-glucuronic acid, l-iduronic acid, and l-altruronic acid building blocks: a) Method A:
BF3·Et2O, CH2Cl2, 08C, 15 min, 93% (11/14/17=1:1:1); Method B: MgBr2·Et2O, toluene, ꢀ78!ꢀ308C, 1 h,
quant. (only 11); b) 1. FmocCl, pyridine, 2 h; 2. HF·pyridine, THF, 16 h, 12: 83%, 15: 89%, 18: 84% (two
steps); c) NIS, CH2Cl2, 15 min, quant. (13, 16, and 19). Fmoc=9-fluorenylmethoxycarbonyl, NIS=N-iodo-
succinimide.
7606 www.angewandte.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2005, 44, 7605 –7607

Angewandte
Chemie
explored further. It has been shown previously that stereo-
selectivities can be significantly improved by the use of metal
Lewis acids in a chelation-controlled Mukaiyama-type aldol
reaction.[13] We envisaged that Felkin–Anh addition of a silyl
enol ether to a aldehyde, in an open transition state, should
Keywords: aldol reaction · carbohydrates · thioglycosides ·
uronic acids
.
[1] a) A. Varki, Glycobiology 1993, 3, 97 – 130; b) R. A. Dwek,
Chem. Rev. 1996, 96, 683 – 720; c) R. Roy, Drug DiscoveryToday
2004, 1, 327 – 336; D. M. Ratner, E. W. Adams, B. R. Su, J.
OꢀKeefe, M. Mrksich, P. H. Seeberger, ChemBioChem 2004, 5,
379 – 382; d) D. H. Dube, C. R. Bertozzi, Nat. Rev. Drug
Discovery 2005, 4, 477 – 488; e) D. B. Werz, P. H. Seeberger,
Chem. Eur. J. 2005, 11, 3194 – 3206.
preferentially furnish
a glucuronic acid. Indeed, the
MgBr2·Et2O-promoted aldol reaction of the aldehyde 8 and
the silyl enol ether 10 (Method B, Scheme 3) afforded
glucuronic acid 11 in quantitative yield as a single diaster-
eomer.
[2] For recent reviews on oligosaccharide synthesis, see: a) S.
Hanessian, Preparative Carbohydrate Chemistry, Marcel
Dekker, New York, 1997; b) Carbohydrates in Chemistry and
Biology, Vol. 1 (Eds.: B. Ernst, G. W. Hart, P. Sinaþ), Wiley-
VCH, Weinheim, 2000; c) B. Davies, J. Chem. Soc. Perkin Trans.
1 2000, 2137 – 2160; d) P. J. Garegg, Adv. Carbohydr. Chem.
Biochem. 2004, 59, 69 – 113.
The BF3·Et2O-mediated aldol reaction between the selec-
tively protected aldehyde 9 and ketene acetal 10 was also
examined (Scheme 4). Surprisingly, analysis of the crude
[3] a) T. Kanemitsu, O. Kanie, Trends Glycosci. Glycotechnol. 1999,
11, 267 – 276; b) P. H. Seeberger, W. C. Haase, Chem. Rev. 2000,
100, 4349 – 4394; c) O. J. Plante, E. R. Palmacci, P. H. Seeberger,
Science 2001, 291, 1523 – 1527; d) P. Sears, C.-H. Wong, Science
2001, 291, 2344 – 2350; e) O. J. Plante, E. R. Palmacci, R. B.
Andrade, P. H. Seeberger, J. Am. Chem. Soc. 2001, 123, 9545 –
9554; f) P. H. Seeberger, Solid Support Oligosaccharide Syn-
thesis and Combinatorial Carbohydrate Libraries, Wiley, New
York, 2001; g) T. Kanemitsu, O. Kanie, Comb. Chem. High
Throughput Screening 2002, 5, 339 – 360; h) O. J. Plante, P. H.
Seeberger, Curr. Opin. Drug DiscoveryDev. 2003, 6, 521 – 525.
[4] a) T. Ogawa, Chem. Soc. Rev. 1994, 23, 397 – 407; b) S. J.
Danishefsky, M. T. Bilodeau, Angew. Chem. 1996, 108, 1380 –
1419; Angew. Chem. Int. Ed. Engl. 1996, 35, 1482 – 1522; c) T.
Hudlicky, D. A. Entwistle, K. K. Pitzer, A. J. Thorpe, Chem. Rev.
1996, 96, 1195 – 1220.
Scheme 4. Synthesis ofselectively protected d-glucuronic and l-id-
uronic acid building blocks. a) Method A: BF3·Et2O, CH2Cl2, 08C,
15 min, 95% (20/23=3:2); Method B: MgBr2·Et2O, toluene, ꢀ78!
ꢀ308C, 1 h, 98% (only 20); b) 1. FmocCl, pyridine, 2 h; 2. HF·pyridine,
THF, 16 h, 21: 79%, 24: 76% (two steps); c) NIS, CH2Cl2, 15 min,
quant. (22 and 25), (a/b=1:1).
[5] For recent reviews on de novo synthesis of carbohydrates, see:
a) R. R. Schmidt, Pure Appl. Chem. 1987, 59, 415 – 424; b) A.
Kirschning, M. Jesberger, K.-U. Schoning, Synthesis 2001, 507 –
540.
[6] S. Y. Ko, A. W. M. Lee, S. Masamune, L. A. Reed, K. B.
Sharpless, F. J. Walker, Science 1983, 220, 949.
[7] a) T. Mukaiyama, I. Shiina, S. Kobayashi, Chem. Lett. 1990, 12,
2201 – 2204; b) T. Mukaiyama, H. Anan, I. Shiina, S. Kobayashi,
Bull. Soc. Chim. Fr. 1993, 130, 388 – 394.
[8] a) A. B. Northrup, D. W. C. MacMillan, Science 2004, 305, 1752 –
1755; b) D. Enders, C. Grondal, Angew. Chem. 2005, 117, 1235 –
1238; Angew. Chem. Int. Ed. 2005, 44, 1210 – 1212.
[9] For a recent review, see: C. Noti, P. H. Seeberger, Chem. Biol.
2005, 12, 731 – 756.
[10] a) T. Mukaiyama, K. Banno, K. Narasaka, J. Am. Chem. Soc.
1974, 96, 7503 – 7509; b) T. Mukaiyama, Org. React. 1982, 28,
203 – 331.
[11] E. Fischer, Ber. Dtsch. Chem. Ges. 1894, 27, 677.
[12] F. L. van Delft, A. R. P. M. Valentijn, G. A. van der Marel, J. H.
van Boom, J. Carbohydr. Chem. 1999, 18, 165 – 190.
[13] D. A. Evans, M. J. Dart, J. L. Duffy, M. G. Yang, J. Am. Chem.
Soc. 1996, 118, 4322 – 4343.
[14] K. Hattori, H. Yamamoto, J. Org. Chem. 1993, 58, 5301 – 5303.
[15] C. Altona, C. A. G. Haasnoot, Org. Magn. Reson. 1980, 13, 417 –
429.
reaction mixture by 1H NMR spectroscopy showed a 3:2 ratio
of two diastereomers, and only traces of a third. Separation of
the individual isomers followed by protecting-group manip-
ulations and NIS-mediated cyclization led to the isolation of
the pyranoses 22 and 25. Through comparison of the 1H NM R
coupling constants with those of the pyranosides 13 and 16,
thioglycoside 22 was identified as d-glucuronic acid and 25 as
l-iduronic acid. Finally, the BF3·Et2O-mediated aldol reaction
between 9 and 10 (analogous to the reaction of 8 and 10) led
to the formation of the gluco-configured product as the only
detectable diastereomer (Method B, Scheme 4).
In conclusion, a highly convergent route to orthogonally
protected d-glucuronic and l-iduronic acid thioglycoside
building blocks has been developed. Rapid access to sub-
stantial quantities of monosaccharides that contain practical
protecting group patterns and a readily activatable anomeric
leaving group will greatly facilitate oligosaccharide assembly
by using automated methods. The construction of heparin
analogues and the development of efficient routes to further
carbohydrate building blocks are currently under investiga-
tion.
Received: August 4, 2005
Published online: November 3, 2005
Angew. Chem. Int. Ed. 2005, 44, 7605 –7607
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 7607