A general, iterative, and modular approach toward carbohydrate libraries based on ruthenium-catalyzed oxidative cyclizations

Article abstract of DOI:10.1021/jo801528n

(Chemical Equation Presented) Carbohydrates are an omnipresent class of highly oxygenated natural products. Due to their wide spectra of biological activities, they have been in the center of synthetic organic chemistry for more than 130 years. During the past 50 years non-natural carbohydrates attracted the interest of various chemists in the fields of organic, biological, and medical chemistry. Especially desoxygenated sugars proved to be an important class of compounds. Up to date, most non-natural analogues are synthesized starting from natural, enantiomerically pure carbohydrates in multistep synthesis. In this report, we present a synthetic strategy that allows the selective modular synthesis of natural and non-natural carbohydrates within five synthetic steps starting from readily available starting materials. Due to a sequential introduction of O-or N-functionalities, a regioselective protection of each new functional group is possible. The key step in the carbohydrate synthesis is a RuO₄-catalyzed oxidative eyclization via a pH-dependent dehydrogenation-dihydroxylation-cyclization or an oxidative fragmentation-cyclization, leading to highly substituted new carbohydrates, in which each functional group is orthogonally protected and accessible for further synthetic operations.

Full text of DOI:10.1021/jo801528n

A General, Iterative, and Modular Approach toward Carbohydrate
Libraries Based on Ruthenium-Catalyzed Oxidative Cyclizations
Meike Niggemann, Andreas Jelonek, Nicole Biber, Margarita Wuchrer, and Bernd Plietker*
Institu¨t fu¨r Organische Chemie, Fakulta¨t Chemie, UniVersita¨t Stuttgart, Pfaffenwaldring 55,
D-70569 Stuttgart, Germany
bernd.plietker@oc.uni-stuttgart.de
ReceiVed July 12, 2008
Carbohydrates are an omnipresent class of highly oxygenated natural products. Due to their wide spectra
of biological activities, they have been in the center of synthetic organic chemistry for more than 130
years. During the past 50 years non-natural carbohydrates attracted the interest of various chemists in the
fields of organic, biological, and medical chemistry. Especially desoxygenated sugars proved to be an
important class of compounds. Up to date, most non-natural analogues are synthesized starting from
natural, enantiomerically pure carbohydrates in multistep synthesis. In this report, we present a synthetic
strategy that allows the selective modular synthesis of natural and non-natural carbohydrates within five
synthetic steps starting from readily available starting materials. Due to a sequential introduction of O-
or N-functionalities, a regioselective protection of each new functional group is possible. The key step
in the carbohydrate synthesis is a RuO4-catalyzed oxidative cyclization via a pH-dependent
dehydrogenation-dihydroxylation-cyclization or an oxidative fragmentation-cyclization, leading to
highly substituted new carbohydrates, in which each functional group is orthogonally protected and
accessible for further synthetic operations.
Introduction
ing demand for small defined libraries of bio- or biomimetic
molecules has led to a technological push in organic synthesis,
in which parallelization and automation allows for the efficient
preparation of a defined set of molecules within a short period
of time.5 However, as compared to the other classes of biological
building blocks, a comparable synthesis of substituted carbo-
hydrates in a parallel fashion has yet not been developed.
This fact might be attributed in part to the chemical properties
of carbohydrates, in which the ring opening-ring closing leads
to mixtures of anomers of different ring sizes. Apart from this
structural flexibility, the presence of up to six different hydroxyl
groups with different but defined relative and absolute config-
uration causes problems in subsequent regioselective functional
group transformations. The latter aspect has led to the develop-
Carbohydrates belong to the four most important classes of
biomolecules (i.e., carbohydrates, lipids, nucleic acids, and
proteins).1 They are involved in a variety of biological processes
on a cellular level, such as recognition processes, metabolism,
etc. Due to the advent of new and sensitive analytical technolo-
gies, biological processes are nowadays studied not only on a
cellular but also on a molecular level. These studies have not
only led to a deeper understanding of the way biomolecules
act and interact but also spurred the interest in biomimetic
molecules, such as nucleic acid,2 amino acid,3 and carbohydrate4
analogues and their behavior in biological systems. The increas-
(1) Lehninger, A. L.; Nelson, D. L.; Cox, M. M. Principles of Biochemistry,
5th ed.; Macmillan: New York, 2008.
(2) (a) Keer, J. T.; Birch, L. Essentials of Nucleic Acid Analysis: A Robust
Approach; Royal Society of Cambridge: Cambridge, UK, 2008. (b) Moldave,
K. Progress in Nucleic Acid Research and Molecular Biology; Elsevier: San
Diego, 2005.
(3) Seewald, N., Jakubke, H. D. Peptides. Chemistry and Biology; Wiley-
VCH: Weinheim, Germany, 2002.
(4) Varki, A.; Cummins, R.; Esko, J.; Freeze, H.; Hart, G.; Marth, J. Essentials
of Glycobiology; Cold Spring Harbor Laboratory Press: Cold Spring Harbor,
NY, 1999.
(5) (a) Kubinyi, H.; Mu¨ller, G. Chemogenomics in Drug DiscoVerysA
Medicinal Chemistry PerspectiVe; Wiley-VCH, Weinheim, Germany, 2004. (b)
Spring, D. R. Chem. Soc. ReV. 2005, 34, 472.
7028 J. Org. Chem. 2008, 73, 7028–7036
10.1021/jo801528n CCC: $40.75 2008 American Chemical Society
Published on Web 08/16/2008

Approach toward Carbohydrate Libraries
ment of sophisticated protecting group strategies that allow for
a differentiation between the various hydroxyl groups. However,
the use of protecting groups increases the amount of synthetic
operations and thus decreases the overall efficiency of the
synthesis.
It is in line with these arguments that so far carbohydrate
libraries have mainly been prepared by derivatization of the
hydroxyl groups within a given carbohydrate core.6 In the past
10 years, organometallic or organic catalysts have changed the
classical field of carbohydrate chemistry. Using these methods,
multicomponent couplings between different monosaccharides
are performed, giving rise to different substituted polysaccha-
rides in an efficient manner.7 Although some very efficient de
FIGURE 1. The pH-dependent chemoselectivity switch in Ru-catalyzed
oxidations.
able to show that a slight modification of the reaction conditions
allows for a selective yet totally different product formation.11
Different from most inorganic oxidation catalysts, RuO4 exhibits
a strong pH-dependent stability13 that allows for a switch of
the oxidation state and chemoselectivity by a simple adjustment
of the pH value. Whereas RuO4 is stable up to pH 7 and oxidizes
CdC bonds much faster than, for example, aliphatic primary
alcohols,11f RuO4Q formed between pH 7 and 10 readily dehy-
drogenates primary alcohols in the presence of CdC bonds14
(Figure 1).
novo carbohydrate syntheses are reported in the literature,8,9
a
broadly applicable, automatable, modular, and parallel synthesis
that allows for structural variations within the carbohydrate core
itself has not been developed up to date.
Recently, we started a research project aiming to build up
small, defined libraries of orthogonally protected carbohydrate
and carbohydrate mimic cores that would allow for a variety of
further synthetic manipulations. As a first part of this research
project, we report here on our initial results in the development
of a potentially automatable, iterative approach toward structur-
ally diverse orthogonally protected carbohydrates and analogues
applying novel RuO4-catalyzed oxidative cyclizations as central
key steps.
We envisioned that this unique pH-dependent mechanistic
dichotomy in RuO4-catalyzed olefin oxidation would pave the
way for a new type of carbohydrate synthesis as depicted in
Figure 2. Starting from a common precursor i, two mechanisti-
cally distinct reaction pathways are possible. Under acidic
conditions, the stable RuO4 reacts with olefins in a [3 +
2]-cycloaddition to give cyclic ruthenates. Depending on the
pH value, these intermediates undergo different subsequent
transformations. Whereas at a slightly acidic pH of 4-6 the
electrocyclic fragmentation is fast, addition of Brønsted or Lewis
acids accelerates the hydrolysis at pH < 4, leading to the
preferred formation of syn-diols. If cyclization precursor i is
oxidized under slightly acidic conditions with RuO4, fragmenta-
tion of the CdC bond occurs. The dehydrogenation of primary
aliphatic alcohols is comparably slow. The resulting aldehyde
ii might undergo a ring closure, forming lactols of type iii (path
A, Figure 2).15 However, starting the same sequence under
slightly basic conditions would yield perruthenate instead of
RuO4. This metal oxo species is known for its selective
dehydrogenation of alcohols even in the presence of olefins.
Hence, the dehydrogenation from i to iv should occur under
basic conditions. Acidification of the reaction mixture to pH <
4 would lead to the in situ formation of RuO4 from perruthenate
and sets the stage for a fast dihydroxylation of iv to v. The
cyclization would then lead to lactols of type vi (path B,
Figure 2).16
Results and Discussion
Part 1: Development of Ru-Catalyzed Oxidative Cycliza-
tions. Concept. Since the original contribution by Shing in
1994,10 the selective, nondestructive oxidation of CdC bonds
in the presence of RuO4 gained increasing interest within the
past years.11 This oxidant exhibits a strong dependency of its
oxidizing behavior from external parameters such as pH value,
temperature, solvents, and reoxidants.12 Due to this dependency,
RuO4-catalyzed oxidations are often accompanied by side
reactions. However, if these parameters are equalized in the
correct way, the variety of possible oxidation modes allows for
the preparation of different oxygen-containing products out of
one common starting material. Within the past 4 years, we were
(6) (a) Takuya, K.; Osamu, K. Trends Glycosci. Glycotechnol. 1999, 11, 267.
(b) Schweizer, F.; Hindsgaul, O. Curr. Opin. Chem. Biol. 1999, 3, 291.
(7) (a) Wong, C.-H. Carbohydrate-Based Drug DiscoVery; Wiley-VCH:
Weinheim, Germany, 2003. (b) Zhang, Z.; Ollmann, I. R.; Ye, X.-S.; Wischnat,
R.; Baasov, T.; Wong, C.-H. J. Am. Chem. Soc. 1999, 121, 734. (c) Plante,
O. J.; Palmacci, E. R.; Seeberger, P. H. Science 2001, 291, 1523.
(8) (a) Ko, S. Y.; Lee, A. W. M.; Masamune, S.; Reed, L. A., III; Sharpless,
K. B.; Walker, F. J. Science 1983, 220, 949. (b) Ko, S. Y.; Lee, A. W. M.;
Masamune, S.; Reed, L. A., III; Sharpless, K. B.; Walker, F. J. Tetrahedron
1990, 46, 245. (c) McGarvey, G. J.; Kimura, M.; Oh, T.; Williams, J. M.
J. Carbohydr. Chem. 1984, 3, 125.
(9) For a modular organocatalytic approach, see: (a) Enders, D.; Grondal,
C. Angew. Chem. 2005, 117, 1235; Angew. Chem., Int. Ed. 2005, 44, 1210. (b)
Northrup, A. B.; MacMillan, D. W. C. Science 2004, 305, 1753.
(10) (a) Shing, T. K. M.; Tai, V. W.-F.; Tam, E. K. W. Angew. Chem. 1994,
106, 2408; Angew. Chem., Int. Ed. Engl. 1994, 33, 2312. (b) Shing, T. K. M.;
Tam, E. K. W.; Tai, V. W. F.; Chung, I. H. F.; Jiang, Q. Chem.sEur. J. 1996,
2, 50.
Investigations on the pH-Dependent Reactivity. With
regard to an attempted automatable and parallel synthesis of
cyclization precursors, the use of one general protecting group
pattern is of importance. The functional group FG2 at C-4 in i
is of major importance for the stereoselectivity in the dihy-
droxylation event. Hence, our investigations started with a
(11) Ru-catalyzed dihydroxylation: (a) Plietker, B.; Niggemann, M. Org. Lett.
2003, 5, 3353. (b) Plietker, B.; Niggemann, M.; Pollrich, A. Org. Biomol. Chem.
2004, 2, 1116. (c) Plietker, B.; Niggemann, M. J. Org. Chem. 2005, 70, 2402.
RuO4-catalyzed ketohydroxylation: (d) Plietker, B. J. Org. Chem. 2003, 68, 7123.
(e) Plietker, B. J. Org. Chem. 2004, 69, 8287. (f) Plietker, B. Eur. J. Org. Chem.
2005, 1919. Ru-catalyzed oxidative cyclization of [1,n]-dienes: (g) Roth, S.; Stark,
C. B. W. Angew. Chem. 2006, 118, 6364; Angew. Chem., Int. Ed., 2006, 45,
6218. (h) Roth, S.; Go¨hler, S.; Cheng, H.; Stark, C. B. W. Eur. J. Org. Chem.
2005, 4109.
(13) (a) Griffith, W. P.; Suriaatmaja, M. Can. J. Chem. 2001, 79, 598. (b)
Mills, A.; Holland, C. J. Chem. Res. 1997, 368.
(14) Ley, S. V.; Norman, J.; Griffith, W. P.; Marsden, S. P. Synthesis 1994,
639.
(15) For analogous reactions using ozone, see: (a) Burton, J. S. J. Chem.
Soc. 1965, 3433. (b) Baumberger, F.; Vasella, A. HelV. Chim. Acta 1986, 69,
1205.
(16) (a) For a KMnO4-mediated oxidative cyclization, see: Fu, X.; Cook,
J. M. J. Org. Chem. 1993, 58, 661. (b) For a related oxidative cyclization in
natural product synthesis, see: Fu¨rstner, A.; Wuchrer, M. Chem.sEur. J. 2006,
12, 76.
(12) For a recent review on RuO4 catalysis, see: Plietker, B. Synthesis 2005,
2453.
J. Org. Chem. Vol. 73, No. 18, 2008 7029

Niggemann et al.
FIGURE 2. Oxidative cyclization pathways toward various carbohydrate skeletons.
TABLE 1. Influence of Protecting Groups on the Stereoselectivity in Dihydroxylationsa
entry
olefin
R
diol
anti:synb
yield [%]c
1
2
3
4
5
6
7
8
3
4
5
6
7
8
9
10
TBDMS
TIPS
TBDPS
Bn
Bz
Ac
11
12
13
14
15
16
17
18
82:18
81:19
74:26
82:18
69:31
70:30
76:24
70:30
93
95
91
89
92
97
88
87
Piv
C(O)Oi-Bu
a All reactions were performed on a 2 mmol scale using 1.0 mol % of RuCl3 (as a 0.1 M stock solution in water), 10 mol % of CeCl3, and 1.5 equiv
of NaIO4 in a solvent mixture of CH3CN/H2O (6 mL/1 mL) at 0 °C and stopped after full conversion. b Determined by 1H NMR integration.
c Combined isolated yield.
TABLE 2. Fragmentation versus DihydroxylationsInfluence of
the pH Valuea
screening of various protecting groups and their directing effect
on the stereoselective course in RuO4-catalyzed dihydroxylation
(Table 1).
Among the tested protecting groups, silyl and benzyl ethers
induce the highest degree of stereoselectivity (entries 1 and 4,
Table 1). However, allyl silyl ethers are susceptible toward acid-
assisted cleavage. Hence, the more stable benzyl group was
chosen for the protection of the 4-OH group in the subsequent
chemoselectivity study.
A dramatic change in the reactivity and mechanism depending
on the pH value was observed (Table 2), indicating the correct
adjustment of this parameter to be of utmost importance for
the selective preparation of a single type of carbohydrate mimic.
Knowing the optimum pH range for the fragmentation versus
dihydroxylation, we subsequently set out to develop the oxida-
tive cyclizations.
The Ru-Catalyzed Oxidative Fragmentation-Cycliza-
tion-Acylation. On the basis of Yang’s procedure for RuO4-
catalyzed fragmentations of olefins,17 various parameters in the
oxidation of cyclization precursor 20 (obtained by basic
hydrolysis of 6) were varied. Gratifyingly, a slight modification
of the original protocol in combination with an anhydrous
entry
pH
14:19b
conversion [%]b
1
2
3
4
5
6
5
4
3
2
8:92
12:88
44:56
89:11
96:4
81
80
79
86
91
a All reactions were performed on a 2 mmol scale using 1.0 mol % of
RuCl3 (as a 0.1 M stock solution in water), 10 mol % of CeCl3, and 1.5
equiv of NaIO4 in a solvent mixture of CH3CN/H2O (6 mL/1 mL) at 0
°C and stopped after 10 min. The pH value was adjusted by titration
using a 2 N NaOH solution prior to the addition of starting material.
b Determined by GC integration.
workup and in situ acylation allowed for the isolation of the
desired cyclization product in good yield (Scheme 1).
Knowing that the pH range of the fragmentation already
allowed for the intramolecular cyclization, the more complex
(17) Yang, D.; Zhang, C. J. Org. Chem. 2001, 66, 4814.
7030 J. Org. Chem. Vol. 73, No. 18, 2008

Approach toward Carbohydrate Libraries
SCHEME 1. Oxidative CdC Bond Cleavage-Cyclization
oxidation of alcohol 20 using a combination of RuCl3/Al2O3 in
ethyl acetate at 80 °C to be a convenient alternative for the in
situ formation of aldehyde 22.19 The immobilized Ru catalyst
is liberated upon addition of acetonitrile and the aqueous slurry
of CeCl3/NaIO4. After full conversion, filtration is followed by
acylation (method B, Scheme 3). The desired product 23 was
obtained in significantly improved yield and identical diaste-
reoselectivity.
dehydrogenation-dihydroxylation-cyclization was approached
subsequently.
SCHEME 3. One-Pot Ru-Catalyzed
Dehydrogenation-Dihydroxylation-Cyclization Sequence
The Ru-Catalyzed Sequential Dehydrogenation-Dihydro-
xylation-Cyclization-Acylation. Initial efforts focused on the
development of a stepwise dehydrogenation of 20 to the
corresponding aldehyde 22 and its subsequent dihydroxyla-
tion-cyclization to 24. In order to facilitate the workup, the
crude product solution was not isolated but directly subjected
to a concomitant acylation to give the bisacylated lactols syn-
and anti-25 in a ratio of 9:91 in an overall yield of 81%
starting from 22 (Scheme 2).
SCHEME 2. RuO4-Catalyzed
Dihydroxylation-cyclization-Acylation sequence
Part 2: Development of a Modular, Potentially Autom-
atable Carbohydrate Library Syntheses. Synthetic Concept.
In order to build up a defined library of carbohydrates, we
envisioned a three-dimensional matrix of potential carbohydrate
precursors i, defined by the three major building blocks A, B,
and C, to be the optimum base for the development of an
efficient carbohydrate library synthesis. From a retrosynthetic
point of view, the carbohydrate precursor can be formed via
Having elaborated a stepwise protocol for the dehydrogena-
tion and oxidative cyclization, we turned our interest toward
the development of a catalytic oxidative sequence. Initial
experiments focused on the use of two different reoxidants. On
the basis of literature known procedures, alcohol 20 was
oxidized at pH 10 using NaBrO3 as the oxidizing agent.18 After
full conversion, simple filtration of the suspension and treatment
of the filtrate with aqueous CeCl3/NaIO4 immediately generated
RuO4 and initiated the final dihydroxylation event. A subsequent
filtration-acylation delivered the 2,3-bisdeoxypyranosides 25
in good overall yield with diastereoselectivities comparable to
the ones obtained in the stepwise protocol (method A, Scheme
3). Although this method proved successful, further investiga-
tions led to the development of a more practical method. We
considered the change between two different stoichiometric
oxidizing salts and the need for an intermediate removal of one
oxidant by filtration to be problematic for large-scale applica-
tions. In furtherance of our investigations on the pH-dependent
chemoselectivity switch, we found the literature reported aerobic
FIGURE 3. Modular, three-dimensional approach toward cyclization
precursor i (n(A,B,C) refers to the number of building blocks).
(18) Behr, A.; Eusterwiemann, K. J. Organomet. Chem. 1991, 403, 209.
J. Org. Chem. Vol. 73, No. 18, 2008 7031

Niggemann et al.
SCHEME 4. Building Block A Synthons 26-29 (BOM )
Variation of Building Block A. A set of five different, chiral
pool derived building block A synthons were chosen as initial
starting points for the preparation of a carbohydrate library
(Scheme 4). With regard to the attempted parallel and autom-
atable synthesis of cyclization precursors, a special emphasis
was placed on the use of identical reagents and conditions for
the subsequent transformations. Hence readily accessible lac-
tones 1 and 30-32 were subjected to a reduction-vinylation-
acylation sequence (Scheme 5). The allylic alcohols were
directly transformed into the desired cyclization precursors using
a one-pot benzylation-saponification protocol.20 The cyclization
precursors 20 and 37-39 were obtained in good to excellent
diastereoselectivities and yields. For the preparation of benzyl
ether (2S,3R,4S)-39, a slightly modified synthetic route was
chosen. Due to the inaccessibility of the corresponding
lactone, the literature known acetate 33 was subjected to a
highly selective 1,2-addition using vinyl zinc chloride
followed by the one-pot benzylation-saponification as
described above (Scheme 5).
benzyloxymethoxy)
the combination of building block A, the backbone of the
carbohydrate, and a vinyl metal species (building block B) plus
a possible functional group transformation (building block C).
The variation of these three building blocks against each other
allows for a general access toward a matrix that is defined by
[n(A) × n(B) × n(C)] permutation, as shown in Figure 3.
Moreover, whereas these theoretical arguments account for
the preparation of a structural diverse library of compounds,
the synthetic approach is of practical use only if (i) the number
of synthetic and diversifying processes is low and if (ii) the
transformations can be performed under parallel conditions on
a defined matrix of starting materials. Hence, in order to avoid
unnecessary synthetic operations (e.g., protecting group ma-
nipulations), every heterofunctional group should be introduced
iteratively, allowing for its selective and orthogonal protection
prior to the introduction of further functionality. This operational
simple trick sets the stage for a short synthesis of orthogonally
protected cyclization precursors and carbohydrate mimics that
can selectively be deprotected for further synthetic applications.
Whereas the requirements mentioned above build the base
for an automatable, parallel synthesis, a common “synthetic
point of convergence” needs to be defined. Up to this point,
the synthetic routes toward the various building blocks A might
differ; however, all further preparative manipulations have to
be performed in an identical and hence automatable manner.
The C-4 epimeric series of cyclization precursors is accessible
in a two-step sequence of oxidation-reduction using Li[Al(Ot-
Bu)3H] or NaBH4 as reducing agents. Allylic alcohols
34-36were obtained in moderate to good diastereoselectivi-
ties (Scheme 6). Having nine different cyclization precursors
in hand, we subsequently turned our attention toward the
variation of the remaining building blocks.
Variation of Building Block B. The stereoselective course
in RuO4-catalyzed dihydroxylations of CdC bonds allows for
the direct translation of π-bond geometry into relative config-
uration of two newly introduced C-O bonds. Hence, we were
wondering whether the 1,2-addition of a long chain vinyl metal
species could be used not only for the preparation of new types
of glycolipids but also for the generation of two stereoisomeric
carbohydrates formed solely through the double bond geometry.
Thus, (E)- and (Z)-vinyl zinc species 41 were prepared according
to literature known procedures21 and reacted with lactone 1 in
analogy to the conditions listed in Scheme 5 to furnish the
corresponding (E)- and (Z)-allyl alcohols (E)-42 and (Z)-42 in
good yields (Scheme 7).
SCHEME 5. Preparation of Cyclization Precursors 20 and 37-39
7032 J. Org. Chem. Vol. 73, No. 18, 2008

Approach toward Carbohydrate Libraries
SCHEME 6. Preparation of C-4 Epimeric Cyclization Precursors 35-37
SCHEME 7. Control of π-Bond Geometry in 1,2-AdditionssPreparation of (E)- and (Z)-42
SCHEME 8. Variation of Building Block B via Cross
Metathesis
SCHEME 9. Variation of Building Block C by Nucleophilic
Substitution
Apart from the 1,2-addition that allows for a control of π-bond
geometry, the selective cross metathesis22 between allyl benzyl
ether 20 and various olefins 43-47 allows for the fast
preparation of a defined set of new cyclization precursors
48-52. The E-configured olefins are formed as the sole products
(Scheme 8). A permutation of the various building blocks A
20 and 37-39 with the five olefins 43-47 was not performed
in this study; however, it should be mentioned that the nine
different precursors can be combined with these olefins to give
45 different cyclization precursors with defined stereochemistry.
Taking into account the possibility of controlling the π-bond
geometry in the 1,2-addition using either (E)- or (Z)-42, the
number of precursors increases already to 63 solely by permuta-
tion of the building blocks mentioned in this study.
Variation of Building Block C. The allyl alcohol moiety in
the cyclization precursors prepared so far sets the stage for a
J. Org. Chem. Vol. 73, No. 18, 2008 7033

Niggemann et al.
7034 J. Org. Chem. Vol. 73, No. 18, 2008

Approach toward Carbohydrate Libraries
variety of functional group transformations. Among the various
substitution reactions of hydroxyl groups, we envisioned the
exchange for a halide atom to be of particular interest. The
biological impact of such a functional group appears to be very
promising plus it paves the way for the introduction of new
C-S or C-N bonds. The general synthetic route is shown in
Scheme 9. The permutation of these four structural variations
with the synthetic operations presented above increases the
amount of potential cyclization precursors from 63 to 252.
However, the deprotection of allylic halides 53 and 54 under
various conditions proved to be problematic, and the desired
alcohols 58 and 59 were obtained as instable liquids in low
yield.
preparation of methyl or thioglycosides by variation of the
workup procedure appears to be interesting.23 Thus, as a proof
ofprinciple,allylbenzylether20wasdehydrogenated-dihydroxy-
lated using the standard conditions. The reaction mixture was
separated from inorganic salts by simple filtration, and the filtrate
was treated with an acidic methanol solution or using Fu¨rstner’s
disulfide method.24 The free hydroxyl group was acylated
subsequently, and the desired products were obtained in good
yield as a mixture of anomers (Scheme 10).
SCHEME 10. Variation of the Work-up
ProtocolsPreparation of Alkyl and Thioglycosides
Preparation of a Carbohydrate Library. Having in hand
a set of potential carbohydrate precursors, we subsequently
turned our attention toward the synthesis of a small defined
carbohydrate library (Table 3) by applying the sequential Ru-
catalyzed oxidative fragmentation-cyclization-acylation condi-
tions shown in Scheme 1 to a set of cyclization precursors
possessing monosubstituted olefins, such as 20, 37-39, and
57-60. Very much to our delight, the set of functional and
protecting groups chosen for this study proved to be stable under
the reaction conditions. The desired lactols were obtained in their
acylated form in good yields as a mixture of anomers. Only the
allylic halides 58 and 59 were unstable under the reaction
conditions. These first encouraging results were followed by the
successful application of the dehydrogenation-dihydroxylation-cy-
clization-acylation to the set of cyclization precursors. With the
exception of the allylic halides, a set of 18 different carbohydrates
and carbohydrate mimics were prepared in a highly diastereose-
lective manner. The relative configuration for each stereoisomer
was investigated and established by intense NMR spectroscopic
analysis on the isolated isomerically pure compounds. Several
biologically active, fully protected specific carbohydrates are among
the member of this library (Table 4).
Conclusion
In the present report, a new iterative de novo carbohydrate
synthesis is described. The synthetic concept is based upon a
short, modular, parallel synthesis of various allyl benzyl ethers,
which are transformed into the corresponding orthogonally
protected carbohydrate mimics via two novel Ru-catalyzed
oxidative cyclizations. The overall simplicity of the synthetic
operations plus the availability of the starting materials makes
this approach especially amenable for the preparation of small,
defined carbohydrate libraries in an automatable way. Future
work will concentrate on this aspect with the goal to develop
glycolipid and peptide libraries and to study their biological
activity against various targets.
TABLE 4. Important Known Carbohydrate Structures
entry
number
name
1
2
3
4
5
6
7
8
9
(4S,5R)-61
(4S,5S)-61
L-2-deoxyarabinose
L-2-deoxyribose
L-xylose
D-arabinose
L-2-deoxyglucose
L-2-deoxyallose
L-altrose
D-glucose
L-allose
L-mannose
(3R,4S,5R)-62
(3S,4S,5S)-62
(4S,5R,6S)-67
(4S,5S,6R)-67
(3R,4S,5R,6S)-68
(3R,4S,5S,6R)-68
(3S,4S,5R,6S)-68
(3S,4S,5S,6R)-68
Experimental Section
10
The general procedures for the oxidative fragmentation-cyclization
and for the dehydrogenation-dihydroxylation-cyclization are given
below. The spectroscopic data of all carbohydrates shown above
are listed in the Supporting Information.
General Procedure for the Oxidative Fragmentation-Cycli-
zation. The appropriate alkene (1 mmol) was dissolved in acetone
(2 mL), acetonitrile (2 mL), and water (1.7 mL). A solution of
RuCl3 (0.1 M, 300 µL, 0.03 mmol, 3 mol %) and NaIO4 (428 mg,
2 mmol) was added. After stirring for 20 min, the mixture was
diluted with ethyl acetate (10 mL), filtered, and washed with a
mixture of NaSO3 solution and NaHCO3 solution (1:1, v/v) (10
Increasing the Synthetic UtilitysThe Workup. A simple
but very important way to increase the synthetic utility of our
protocol is the workup procedure. The lactols derived via
oxidative cyclization have been acylated during the workup
procedure. However, with regard to attempted applications in
the synthesis of more complex di- or polysaccharides, the
(19) Yamaguchi, K.; Mizuno, N. Chem.sEur. J. 2003, 9, 4353.
(20) Allylic alcohol 2 was obtained in a racemic way and used as such
throughout this study. However, aysmmetric addition of a vinyl zinc reagent in
the presence of an enantiomerically pure amino alcohol might be used for the
preparation of the isomerically pure alcohol. In the present study, the main focus
was placed on the use of the allyl benzyl ethers in the cyclization event.
(21) (a) Oppolzer, W.; Radinov, R. N. HelV. Chim. Acta 1992, 75, 170. (b)
Negishi, E.-I.; Williams, R. M.; Lew, G.; Yoshida, T. J. Organomet. Chem.
1975, 92, C4. (c) Campbell, J. B.; Molander, G. A. J. Organomet. Chem. 1978,
156, 71. (d) Jeon, S.-J.; Fisher, E. L.; Carroll, P. J.; Walsh, P. J. J. Am. Chem.
Soc. 2006, 128, 9618.
(22) For an excellent and most recent review on the potential of cross
metathesis in natural product synthesis, see: Hoveyda, A.; Zhugralin, A. R. Nature
2007, 450, 243, and references cited therein.
(23) Ogura, H.; Hasegawa, A.; Suami, T. Carbohydrates: Synthetic Methods
and Applications in Medicinal Chemistry; Wiley-VCH: Weinheim, Germany,
1998.
(24) Fu¨rstner, A. Liebigs Ann. Chem. 1993, 11, 1211.
J. Org. Chem. Vol. 73, No. 18, 2008 7035

Niggemann et al.
mL). Phases were separated, and the aqueous layer was extracted
with ethyl acetate (2 × 15 mL). The combined organic layers were
dried over Na2SO4, filtered, and evaporated in vacuum (max 30
°C). The obtained crude product was acylated directly. To a solution
of the substrate dissolved in dry dichloromethane (2 mL) were added
pyridine (815 µL, 10 mmol), acetic anhydride (470 µL, 5 mmol),
and DMAP (catalytic amount). After 12 h, the mixture was diluted
with ether (20 mL) and hydrolyzed by the addition of 2 N HCl.
Phases were separated, and the aqueous layer was extracted with
ether (3 × 20 mL). The combined organic layers were washed with
saturated NaHCO3 solution dried over Na2SO4, filtered, and
evaporated in vacuum. The obtained crude product was purified
via flash chromatography.
gently heated until a bright yellow suspension was formed. The
yellow suspension was cooled to 0 °C, added to the aldehyde
Ru-Al2O3 suspension, and stirred for 5 min. Solid Na2SO4 (5 g)
was added followed by ethyl acetate (10 mL). The solid was filtered
of, the filter cake was washed several times with ethyl acetate, and
the filtrate was concentrated in vacuum. The crude product was
directly acylated as described above.
(2S,3R,4S,5R,6S)-68: Rf 0.31 (3.1 iso-hexanes/ethyl acetate);
1
[R]20 ) -66 (c ) 0.1, CH2Cl2); H NMR (400 MHz, CDCl3) δ
D
7.37-7.31 (m, 5H), 5.90 (d, J ) 7.0 Hz, 1H), 4.86 (d, J ) 11.5
Hz, 1H), 4.84 (ddd, J ) 11.0, 3.0, 3.0 Hz, 1H), 4.71 (dd, J ) 9.6,
7.0 Hz, 1H), 4.68 (d, J ) 11.5 Hz, 1H), 4.32 (s, 1H), 4.13 (dd, J
) 12.5, 11.0 Hz, 1H), 4.09 (d, J ) 9.5 Hz, 1H), 3.56 (d, J ) 12.5
Hz, 1H), 2.12 (s, 3H), 1.97 (s, 3H), 1.46 (s, 3H), 1.43 (s, 3H) ppm;
13C NMR (100 MHz, CDCl3) δ 170.1, 169.7, 138.4, 128.4, 128.2,
128.0, 111.3, 95.3, 76.7, 75.8, 74.5, 74.2, 71.5, 61.1, 26.9, 26.8,
21.4, 20.9 ppm; IR (film) ν 2986 (m), 2936 (m), 1754 (s), 1497
(w), 1455 (w), 1371 (m), 1225 (s), 1156 (m), 1111 (m), 1079 (m),
1058 (m), 1033 (m), 1013 (m) cm-1; m/z (EI) 335 (5), 276 (5),
229 (5), 173 (5), 127 (10), 115 (10), 105 (10), 91 (100), 81(5), 65
(5); HRMS (FAB + HR) calcd for C18H23O6 335.1495, found
(3S,4S,5R)-62: Rf 0.31 (4:1 iso-hexanes/ethyl acetate); [R]20
)
D
-72 (c ) 1.7, CHCl3); 1H NMR (400 MHz, CDCl3) δ 7.38-7.30
(m, 5H), 6.10 (d, J ) 6.7 Hz, 1H), 4.75 (d, J ) 12.6 Hz, 1H), 4.69
(d, J ) 12.6 Hz, 1H), 4.58 (dd, J ) 7.5, 3.0 Hz, 1H), 4.27-4.24
(m, 1H), 3.68 (m, 1H), 3.60 (dd, J ) 6.8, 3.0 Hz, 1H), 2.05 (s,
3H), 1.55 (s, 3H), 1.35 (s, 3H) ppm; 13C NMR (100 MHz, CDCl3)
δ 169.2, 131.2, 128.6, 128.2, 128.1, 107.4, 93.7, 74.1, 73.7, 72.5,
72.2, 64.0, 26.5, 25.6, 25.3 ppm; IR (film) ν 2987 (m), 2934 (m),
1745 (s), 1633 (m), 1455 (w), 1371 (m), 1226 (s), 1166 (w), 1064
(s) 1009 (m) cm-1; m/z (EI) 322 (1), 219 (5), 156 (5), 141 (10), 91
(100), 65 (5); HRMS (ESI + Na+) calcd for C15H19O4 263.1283,
found 263.1280.
335.1476. (2R,3R,4S,5R,6S)-68: Rf 0.31 (3:1 iso-hexanes/ethyl
1
acetate); [R]20 ) -45 (c ) 0.1, CH2Cl2); H NMR (400 MHz,
D
CDCl3) δ 7.37-7.28 (m; 5H), 5.91 (d, J ) 7.0 Hz, 1H), 4.84 (dd,
J ) 3.0 Hz, 1H), 4.83 (d, J ) 11.5 Hz, 1H), 4.73 (d, J ) 11.5 Hz,
1H), 4.60 (dd, J ) 9.5, 7.0 Hz, 1H), 4.28 (dd, J ) 9.5, 1.5 Hz,
1H), 4.09 (d, J ) 14.0 Hz, 1H), 3.95 (s, 1H), 3.77 (d, J ) 14.0 Hz,
General procedure for the dehydrogenation-dihydroxylation-
cyclization. Step 1: Preparation of Ru-Al2O3 Catalyst. Al2O3
(2.00 g, 19.6 mmol) was stirred with an aqueous solution of RuCl3
(2 mL, 0.1 M) for 30 min at ambient temperature. The brown
aqueous phase became lighter, and the powder turned dark gray.
The powder was filtered, washed with water (15 mL), and dried in
vacuum. The obtained powder was added into water (30 mL), and
the pH value of the solution was slowly adjusted to 13 by addition
of an aqueous solution of NaOH (1.0 M). The resulting slurry was
stirred for 24 h at ambient temperature. The color of the powder
turned from dark gray to dark green-gray. The solid was filtered
off, washed with water (20 mL), and dried in vacuum to afford
Ru-Al2O3 (1.9 g).
Step 2: Sequential Dehydrogenation-Dihydroxylation-Cy-
clization. A suspension of Ru-Al2O3 (250 mg) in ethyl acetate
(10 mL) was stirred for 5 min at ambient temperature. Alcohol
(0.5 mmol) was added, molecular oxygen was passed through the
suspension, and the resulting mixture was stirred at 85 °C for 72 h.
After 24 h Ru-Al2O3 (250 mg) was added, and heating was
continued for further 50 h (100 mg). After cooling to ambient
temperature, acetonitrile (10 mL) was added and the slurry was
stirred for 20 min. An aqueous solution of H2SO4 (100 µL, 2 N)
was added while stirring for 20 min. The suspension was cooled
to 0 °C. Meanwhile, NaIO4 (161 mg, 0.75 mmol) and CeCl3 ·7H2O
(18.6 mg. 0.05 mmol, 10 mol %) were stirred in 3 mL of H2O and
1H), 2.12 (s, 3H), 2.11 (s, 3H), 1.48 (s, 3H), 1.45 (s, 3H) ppm; 13
C
NMR (100 MHz, CDCl3) δ 170.8, 169.7, 133.0, 128.5, 128.1, 128.0,
111.1, 94.9, 76.5, 74.9, 74.8, 74.1, 70.8, 61.2, 27.0, 21.4, 21.0 ppm;
IR (film) ν 2986 (m), 2935 (m), 1754 (s), 1632 (w), 1496 (w),
1455 (m), 1371 (m), 1226 (s), 1153 (m), 1088 (m), 1061 (m), 1011
(m) cm-1; m/z (EI) 334 (5), 276 (5), 229 (5), 187 (5), 173 (5), 127
(5), 115 (5), 105 (5), 91 (100), 82 (5), 65 (5); HRMS (FAB + HR)
calcd for C18H23O6 335.1495, found 335.1478.
Acknowledgment. This paper is dedicated to Prof. Volker
Ja¨ger on the occasion of his 65th birthday. The authors thank
the Fonds der Chemischen Industrie (Liebig fellowship for B.P.,
Kekule fellowship for M.W.), the Deutsche Forschungsgemein-
schaft (Emmy-Noether fellowship for B.P., Ph.D. grant for M.N.
(PL 300/4-1 and 300/4-2) and A.J. (PL 300/5-1)), and the Dr.-
Otto-Ro¨hm-Geda¨chtnisstiftung for generous financial support.
Supporting Information Available: Experimental proce-
dures for the preparation and spectral data for all reported
1
compounds plus copies of H NMR spectra. This material is
available free of charge via the Internet at http://pubs.acs.org.
JO801528N
7036 J. Org. Chem. Vol. 73, No. 18, 2008