Automated solid-phase synthesis of β-mannuronic acid alginates

Article abstract of DOI:10.1002/anie.201108744

A big step towards routine: The title synthesis provides mannuronic acid alginate fragments featuring up to 12 cis-mannosidic linkages, in multi-milligram quantities (see scheme; Bn=benzyl, Lev=levulinoyl). Mannuronic acid building blocks were used in a second-generation carbohydrate synthesizer to secure the stereoselective introduction of the β-mannosidic bonds in a fully automated fashion. Copyright

Full text of DOI:10.1002/anie.201108744

Angewandte
Chemie
DOI: 10.1002/anie.201108744
Carbohydrates
Automated Solid-Phase Synthesis of b-Mannuronic Acid Alginates**
Marthe T. C. Walvoort, Hans van den Elst, Obadiah J. Plante, Lenz Krçck, Peter H. Seeberger,
Herman S. Overkleeft, Gijsbert A. van der Marel,* and Jeroen D. C. Codꢀe*
Carbohydrates are the most structurally diverse biopolymers
found in nature.[1] This variety is mirrored in the plethora of
biological functions that carbohydrates and glycoconjugates
perform.[2] Polysaccharides are prime constituents of the cell
wall of bacteria and fungi. On the outside of the cell they
mediate and modulate the interactions of the cell with its
surroundings.[3] As such, many bacterial and fungal polysac-
charides are involved in immunological processes. Isolation of
these compounds in sufficient amounts and purity to study
their biological role in detail is often an impractical task
because of the micro-heterogeneity of naturally occurring
oligosaccharides. Organic synthesis has the potential to meet
the growing demand for well-defined oligosaccharides,
including (nonnatural) analogues and conjugates.
Compared to the synthesis of the other biopolymers, the
assembly of oligosaccharides is significantly more difficult
because the multiple hydroxy groups of the monomeric
building blocks have to be discriminated during the synthesis.
Perhaps an even bigger obstacle is presented by the union of
two saccharides through the formation of a glycosidic linkage,
which involves the creation of a new stereocenter.[4] As
a direct result, the automated synthesis of oligosaccharide
fragments using solid-phase techniques, which has revolu-
tionized nucleotide and peptide chemistry, is still in its
infancy.[5] In 2001, Seeberger and co-workers reported on
the first automated oligosaccharide synthesizer and showed
that oligosaccharide synthesis is amendable to automated
synthesis.[6] Key to the success of this methodology is the
reliable installation of the desired glycosidic bonds in
stereochemically pure form and high chemical yield. The
stereoselective construction of 1,2-trans-glycosidic bonds
nowadays is a routine exercise, both in solution and on the
solid support. However, the formation of 1,2-cis-glycosidic
linkages, especially the b-mannosidic bond, still presents
a significant synthetic challenge.[4] The most powerful solution
for the b-mannoside problem reported to date was developed
by Crich and co-workers,[7] who showed that benzylidene-
acetal-protected mannosides could be used for the formation
of the 1,2-cis-mannosyl linkage. However, transposing the
benzylidene mannosylation chemistry to the solid support was
met with varying success.[8,9]
Herein we report on the automated solid-phase assembly
of mannuronic acid alginate oligomers, featuring up to twelve
1,2-cis-mannosidic linkages. The structures were constructed
using a second-generation automated oligosaccharide synthe-
sizer[10] and the stereoselective formation of the b-mannosidic
linkages was secured through the use of mannuronic acid
donors. The use of the synthesizer allowed us to rapidly access
target structures, without intermediate purification and in
quantities that are not only sufficient to cater for biological
experiments but also to facilitate verification of the structural
1
integrity of the compounds using standard H and 13C NMR
techniques.
Poly-b-(1,4)-mannuronic acid alginate (A, Scheme 1) is
a major component of the cell wall of various algae.[11] It also
represents the exo-polysaccharide of Pseudomonas aerigu-
nosa,[12,13] an opportunistic, nosocomial gram-negative bacte-
Scheme 1. Target structure b-(1,4)-mannuronic acid alginate (A), which
is synthesized using mannuronic acid building blocks B (X=SPh,
OC(=NPh)CF3). Bn=benzyl, Lev=levulinoyl.
[*] Dr. M. T. C. Walvoort,[+] Dr. H. van den Elst,[+]
Prof. Dr. H. S. Overkleeft, Prof. Dr. G. A. van der Marel,
Dr. J. D. C. Codꢀe
rium, which poses a serious health threat to immunocompro-
mised patients. Small mannuronic acid oligomers have been
shown to have Toll-like receptor 2 and 4 mediated immuno-
modulatory activity.[14] To enable the study of the antigenicity
and immunomodulatory effects of mannuronic acid alginates
we recently started a synthetic campaign aimed at their
construction.[15] During these studies we discovered that
mannuronic acid donors are highly b-selective glycosylating
agents. Using solution-phase chemistry we were able to
assemble a mannuronic acid pentamer using a convergent
synthetic strategy, in which a monomeric acceptor was
elongated with two successive disaccharide donors in 51%
and 69%, respectively.[15a] We reasoned that an automated
solid-phase synthesis approach could be more efficient for the
assembly of a library of larger mannuronic acid alginate
Leiden Institute of Chemistry, Leiden University
P.O.Box 9502, 2300 RA Leiden (The Netherlands)
E-mail: marel_g@chem.leidenuniv.nl
jcodee@chem.leidenuniv.nl
Dr. O. J. Plante
Ancora Pharmaceuticals, Medford, MA 02155 (USA)
Dr. L. Krçck, Prof. Dr. P. H. Seeberger
Department of Biomolecular Systems, Max-Planck-Institute of
Colloids and Interfaces, 14476 Potsdam (Germany)
[+] These authors contributed equally.
[**] The authors thank C. Erkelens and F. Lefeber for their assistance
with NMR analyses. This work was supported by the Netherlands
Organization of Scientific Research (NWO, VIDI grant) and as well
as the Max-Planck Society and a Kçrber Prize (to P.H.S.).
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201108744.
Angew. Chem. Int. Ed. 2012, 51, 4393 –4396
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4393

.
Angewandte
Communications
fragments. The success of this approach clearly hinges on the
efficient construction of the b-mannuronic acid bonds, which
have to be introduced in high yield and in a stereoselective
manner to prevent the formation of inseparable (anomeric)
mixtures. In our solution-phase experiments we generally use
thioglycosyl donors in a preactivation glycosylation protocol.
Since preactivation of the donors is not (yet) possible on the
synthesizer and the nature of the linker prohibits the use of
soft electrophiles, required for the activation of thioglyco-
sides, we chose to use N-phenyltrifluoroacetimidate[16] donor
1 as the key building block (Scheme 2). This donor can be
by cross-metathesis with ethylene using Grubbsꢁ first-gener-
ation precatalyst gave a mixture of anomeric diastereomers
(a/b = 1:3). Although the b product was predominantly
formed, the stereoselectivity was clearly insufficient to be
used in the assembly of larger oligomers. We therefore
lowered the reaction temperature of the glycosylation reac-
tion close to the decomposition temperature of the inter-
mediate triflate (ꢀ408C) to assure its intermediacy in the
reaction. This step resulted in the exclusive formation of the
1
b-linked product, as judged by H NMR spectroscopy of the
sample mixture that was cleaved from the resin. For the
removal of the C4-O-levulinoyl ester optimal reaction con-
ditions were found in the use of H2NNH2·HOAc (10 equiv) in
a mixture of pyridine/AcOH (4:1 v/v) at slightly elevated
temperature (408C). Finally the coupling efficiency in terms
of monomer to dimer conversion was optimized. It was found
that glycosylating with two coupling cycles of 1 (5 equiv) and
TMSOTf as a promotor led to a conversion of approximately
80%. Changing to a protocol in which the TfOH was used as
an activator and the coupling cycle was repeated three times
with 3 equivalents of 1, led to a significantly better conversion
(> 95%). Also, the reaction mixtures were collected after
each coupling, which allowed recovery of the unreacted 1 in
approximately 20%. Separate automated synthesis runs were
conducted to generate tetrasaccharide 2, octasaccharide 3,
and dodecasaccharide 4 (Scheme 3, and see the Supporting
Information).
Scheme 2. Investigation into the activation of glycosylating agent
1 using low-temperature NMR spectroscopy. Tf=trifluoromethanesul-
fonyl.
prepared from known thioglycoside intermediates on a multi-
gram scale,[17] and is activated by a catalytic amount of Lewis
or Brønsted acid. The levulinoyl ester was installed as
a temporary protecting group at C4 because this can be
selectively cleaved under near-neutral conditions without
touching the methyl esters or causing epimerization or
b elimination.
The activation of 1 was first examined in a low-temper-
ature NMR experiment (Scheme 2). To this end 1 was
dissolved in [D2]dichloromethane and treated with a slight
excess of TfOH at ꢀ808C. The donor was rapidly consumed
to provide a conformational mixture of the two anomeric a-
triflates 11a and 11b, as we previously established for the
analogous thiophenyl donor (B-SPh; Scheme 1).[18] The
excess of TfOH in this experiment proved to be too acidic
for the intermediate triflate 11, thus resulting in degradation
of the mixture. As a comparison, when the corresponding
thiophenyl donor was activated using ꢀneutralꢁ conditions
(Ph2SO/Tf2O), the same anomeric triflate was produced.
Under these reaction conditions the temperature of decom-
position of triflate 11 was determined to be ꢀ408C. This
relatively low decomposition temperature provides an indi-
cation of the reactivity of the donors at hand, which is of
importance in establishing optimal glycosylation conditions
(see below).
We then moved to the automated solid-phase synthesizer
to investigate the chemistry on solid support. Merrifield
resin[19] was functionalized with a butenediol linker (loading
0.34 mmolgꢀ1), which allows cleavage of the products from
the solid support through cross-metathesis with ethylene.[20]
Before the assembly of the larger oligomers, we started off
with optimizing the coupling and deprotection steps. In a first
attempt, 1 (2 ꢂ 5 equivalents)[21] was coupled to the resin
under the agency of a catalytic amount of TMSOTf (0.2 equiv
with respect to 1) at 08C. Cleavage of a sample from the resin
Scheme 3. Automated solid-phase assembly of mannuronic acid algi-
nates.
In a generalized procedure, the reaction vessel of the
synthesizer was charged with resin (100 mg, 34 mmol) and this
was subjected to the number of coupling/deprotection cycles
as programmed. After the final deprotection step, the resin
was collected, the products were released from the resin by
4394 www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 4393 –4396

Angewandte
Chemie
metathesis, and the crude reaction mix-
ture was analyzed by LC/MS and NMR
spectroscopy. As can be estimated from
the ELSD trace of the LC chromato-
gram,[17] the crude reaction mixture of
the tetramer synthesis contained about
90% of the desired product 2, next to
a
minor amount of the deletion
sequence trisaccharide, thus indicating
a high coupling efficiency. Importantly,
the NMR spectra of the crude cleavage
mixture showed that the coupling reac-
tions had proceeded with excellent ste-
reoselectivity. The relatively high chem-
ical shifts of the anomeric signals in the
13C APT spectrum are indicative of b-
mannosidic linkages. Furthermore, the
heteronuclear one-bond C1-H1 coupling
constants (JC1-H1 ꢁ 156–158 Hz) unam-
biguously ascertain the installation of
the 1,2-cis-linkages. The construction of
longer fragments proceeded equally
well. The automated solid-phase syn-
thesis of octamer 3 led to a crude
reaction mixture containing approxi-
mately 55% of the desired product,
which equals about 90% efficiency per
coupling cycle. Dodecamer 4 made up
approximately 40% of the crude reac-
tion mixture obtained after 12 repetitive
coupling/deprotection cycles as indi-
cated by LC/MS (Figure 1A), thus rep-
resenting about 90% coupling effi-
ciency. Interestingly, the 1H and
13C APT spectra of the crude reaction
Figure 1. A) LC/MS chromatogram of the crude reaction mixture containing dodecasaccharide 4
(gradient 70!95%); B) LC/MS chromatogram of the saponified mixture containing dodeca-
saccharide 7 (gradient 70!95%); C) LC/MS chromatogram of purified product 7 (gradient
50!90%); D) 1H NMR spectrum of saponified product 7 after HPLC purification; E) 13C APT
NMR spectrum of pure product 7; F) HMBC-GATED NMR spectrum of the pure product 7.
mixtures obtained from compound 3 and 4 are remarkably
similar to the spectrum obtained for the reaction mixture
from 2, and only differ in the intensity of the signals belonging
to the internal mannuronic acid residue.[17] This data indicates
that the structures of the oligomers are very regular, and that
the glycosidic bonds have been introduced with excellent
b stereoselectivity.
Preparative HPLC purification of the oligomers was
accomplished after saponification of the methyl esters
(KOH, THF/H2O, 1:1, v/v), because at this stage an improved
baseline separation of the peaks corresponding to the target
products in the HPLC trace was observed (Figure 1B ). In this
way tetramer 5 was obtained in 24 mg, octamer 6 in 20 mg,
and dodecamer 7 in 17 mg (NMR analysis of product 7 is
shown in Figures 1D–F). The amounts of mannuronates
isolated correspond to overall yields of 47% for tetraman-
nuronate 5 (8 on-resin steps), 16% for octamannuronate 6 (16
on-resin steps), and 11% for dodecamannuronate 7 (24 on-
resin steps).[22] These numbers approach a yield of 90% per
chemical step. Final deprotection of the partially protected
oligomers was accomplished by hydrogenolysis over Pd/C in
THF/H2O/tBuOH (1:1:1, v/v/v) to provide the target tetramer
8, octamer 9, and dodecamer 10 in excellent yields and multi-
milligram quantities.
In conclusion, we have described the automated synthesis
of mannuronic acid alginates featuring up to twelve 1,2-cis-
mannosidic bonds, using a second-generation oligosaccharide
synthesizer. It has been shown that the synthesizer is capable
of rapidly delivering oligosaccharides of a length difficult to
obtain by solution-phase techniques. Importantly, the multi-
milligram quantities of the compounds are not only sufficient
for biological experiments but also enable the full structural
characterization of the compounds. Together with the recent
advances in the stereoselective construction of 1,2-cis-gluco-
sidic and 1,2-cis-galactosidic linkages,[23] this represents an
important step forwards towards routine automated solid-
phase oligosaccharides assembly. Our growing insight into the
mechanisms of glycosylation reactions in combination with
the future availability of commercial synthesizers and ever
more powerful purification techniques will open up new
avenues in glycochemistry and glycobiology.
Received: December 12, 2011
Published online: February 14, 2012
Keywords: carbohydrates · glycosylation · protecting groups ·
.
solid-phase synthesis · synthesis design
Angew. Chem. Int. Ed. 2012, 51, 4393 –4396
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
4395

.
Angewandte
Communications
[13] a) V. L. Campodꢆnico, N. J. Llosa, L. V. Bentancor, T. Maira-
Litran, G. B. Pier, Infect. Immun. 2011, 79, 3455 – 3464; b) D. M.
Ramsey, D. J. Wozniak, Mol. Microbiol. 2005, 56, 309 – 322.
[14] M. Iwamoto, M. Kurachi, T. Nakashima, D. Kim, K. Yamaguchi,
T. Oda, Y. Iwamoto, T. Muramatsu, FEBS Lett. 2005, 579, 4423 –
4429.
[15] a) J. D. C. Codꢃe, L. J. van den Bos, A.-R. de Jong, J. Dinkelaar,
G. Lodder, H. S. Overkleeft, G. A. van der Marel, J. Org. Chem.
2009, 74, 38 – 47; b) L. J. van den Bos, J. Dinkelaar, H. S. Over-
kleeft, G. A. van der Marel, J. Am. Chem. Soc. 2006, 128, 13066 –
13067.
[16] a) Ref. [4b]; b) B. Yu, H. Tao, Tetrahedron Lett. 2001, 42, 2405 –
2407.
[17] See the Supporting Information.
[18] M. T. C. Walvoort, G. Lodder, J. Mazurek, H. S. Overkleeft,
J. D. C. Codꢃe, G. A. van der Marel, J. Am. Chem. Soc. 2009,
131, 12080 – 12081.
[19] R. B. Merrifield, J. Am. Chem. Soc. 1963, 85, 2149 – 2154.
[20] R. B. Andrade, O. J. Plante, L. M. Melean, P. H. Seeberger, Org.
Lett. 1999, 1, 1811 – 1814.
[21] Because of the high overall yield of an automated solid-phase
synthesis campaign, the synthesis can be executed at a signifi-
cantly smaller scale than the analogous solution-phase synthesis.
As a result the amount of building block required is not as big as
often assumed. Depending on the efficiency of both the solution-
phase and solid-phase processes, the latter can in fact be more
“building block economical” than the former.
[22] During the purification of the oligomers the deletion sequences
were also obtained in good purity (see the Supporting Informa-
tion).
[1] a) M. L. Sinnott in Carbohydrate Chemistry and Biochemistry,
The Royal Society of Chemistry, Cambridge, 2007; b) Compre-
hensive Glycoscience, Vol. 1 (Ed: J. P. Kamerling), Elsevier,
Oxford, 2007; c) T. J. Boltje, T. Buskas, G.-J. Boons, Nat. Chem.
2009, 1, 611 – 622; d) P. H. Seeberger, D. B. Werz, Nature 2007,
446, 1046 – 1051.
[2] A. Varki, Glycobiology 1993, 3, 97 – 130.
[3] K. Ohtsubo, J. D. Marth, Cell 2006, 126, 855 – 867.
[4] a) L. K. Mydock, A. V. Demchenko, Org. Biomol. Chem. 2010,
8, 497 – 510; b) X. M. Zhu, R. R. Schmidt, Angew. Chem. 2009,
121, 1932 – 1967; Angew. Chem. Int. Ed. 2009, 48, 1900 – 1934.
[5] a) J. M. J. Frꢃchet, C. Schuerch, J. Am. Chem. Soc. 1971, 93, 492 –
496; b) C.-H. Hsu, S.-C. Hung, C.-Y. Wu, C.-H. Wong, Angew.
Chem. 2011, 123, 12076 – 12129; Angew. Chem. Int. Ed. 2011, 50,
11872 – 11923.
[6] O. J. Plante, E. R. Palmacci, P. H. Seeberger, Science 2001, 291,
1523 – 1527.
[7] D. Crich, Acc. Chem. Res. 2010, 43, 1144 – 1153.
[8] J. D. C. Codꢃe, L. Krçck, B. Castagner, P. H. Seeberger, Chem.
Eur. J. 2008, 14, 3987 – 3994.
[9] D. Crich, M. Smith, J. Am. Chem. Soc. 2002, 124, 8867 – 8869.
[10] a) W. Christ, L. Krçck, O. J. Plante, B. Castagner, P. H.
Seeberger (Ancora Pharmaceuticals), WO 2010/011828 A1,
2010; b) P. H. Seeberger, L. Krçck, D. Esposito, C.-C. Wang, P.
Bindschꢄdler, B. Castagner, submitted for publication.
[11] W. A. P. Black, W. J. Cornhill, E. T. Dewar, J. Sci. Food Agric.
1952, 3, 542 – 550.
[12] a) A. Linker, R. S. Jones, J. Biol. Chem. 1966, 241, 3845 – 3851;
b) S. T. Moe, K. I. Draget, G. Sjꢅk-Bræk, O. Smidsrød in Food
Polysaccharides and Their Applications (Ed.: A. M. Stephen),
Marcel Dekker, New York, 1995, pp. 245 – 286; c) I. D. Hay,
Z. U. Rehman, A. Ghafoor, B. H. A. Rehm, J. Chem. Technol.
Biotechnol. 2010, 85, 752 – 759; d) M. J. Franklin, D. E. Nivens,
J. T. Weadge, P. L. Howell, Front. Microbiol. 2011, 2, 167.
[23] a) T. J. Boltje, J.-H. Kim, J. Park, G.-J. Boons, Nat. Chem. 2010, 2,
552 – 557; b) D. B. Werz, B. Castagner, P. H. Seeberger, J. Am.
Chem. Soc. 2007, 129, 2770 – 2771.
4396 www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 4393 –4396