Synthesis of a six-carbon sialic acid using an indium-mediated coupling.

Article abstract of DOI:10.1021/ol990320r

[reaction: see text] The synthesis of a six-carbon truncated sialic acid is described. A key step in the synthesis was an indium-mediated allyl addition to a serine-derived aldehyde. Careful choice of protecting groups was found to be necessary in order to prevent unwanted side reactions throughout the sequence. The truncated sialic acid was obtained in a form suitable for activation as a glycosyl donor.

Full text of DOI:10.1021/ol990320r

ORGANIC
LETTERS
2000
Vol. 2, No. 14
2003-2005
Synthesis of a Six-Carbon Sialic Acid
Using an Indium-Mediated Coupling
Mark D. Chappell and Randall L. Halcomb*
Department of Chemistry and Biochemistry, UniVersity of Colorado,
Boulder, Colorado 80309-0215
halcomb@colorado.edu
Received October 11, 1999
ABSTRACT
The synthesis of a six-carbon truncated sialic acid is described. A key step in the synthesis was an indium-mediated allyl addition to a
serine-derived aldehyde. Careful choice of protecting groups was found to be necessary in order to prevent unwanted side reactions throughout
the sequence. The truncated sialic acid was obtained in a form suitable for activation as a glycosyl donor.
Sialic acids are a class of monosaccharides that are found at
the termini of oligosaccharides in many mammalian cellular
systems.1 Sialic acids are represented by the prototypical
congener N-acetylneuraminic acid (NeuAc, Scheme 1).
unnatural derivatives of this residue.2 Specifically, there was
a need for the truncated six-carbon sialic acid 1, especially
in a form that would be a suitable glycosyl donor or an
immediate precursor thereof. Investigations of oligosaccha-
rides which contain 1 could provide information concerning
the role of the three-carbon glycerol side chain of sialic acids
in biochemical and cellular processes, such as those that are
mediated by carbohydrate-protein recognition events.
Sialic acids are generally prepared from derivatives of
N-acetylmannosamine (ManNAc) by an aldol condensation
with pyruvate which is catalyzed by NeuAc aldolase.3
However, compound 1 could not be prepared by this
Scheme 1
Scheme 2
During the course of research in this laboratory into the
biochemical roles of oligosaccharides that contain sialic acids,
there arose a need for synthetic access to natural and
(1) (a) Rademacher, T. W.; Parekh, R. B.; Dwek, R. A. Annu. ReV.
Biochem. 1988, 57, 785. (b) Varki, A. Glycobiology 1993, 3, 97. (c) Dwek,
R. A. Chem. ReV. 1996, 96, 683.
(2) (a) Chappell, M. D.; Halcomb, R. L. J. Am. Chem. Soc. 1997, 119,
3393. (b) Chappell, M. D.; Halcomb, R. L. Tetrahedron 1997, 53, 11109.
10.1021/ol990320r CCC: $19.00 © 2000 American Chemical Society
Published on Web 06/09/2000

Scheme 3
enzymatic method because aldehyde precursors under five
organometallic reagents.7 An ozonolysis of compound 4
would then afford the R-keto ester 5. Upon removal of its
blocking group, the primary alcohol was anticipated to
spontaneously cyclize onto the resulting ketone to afford the
hemiketal 6.
carbons in length, such as N-acetylserinal needed here
(Scheme 1), are generally poor substrates for Neu5Ac
aldolase.4 Consequently, a nonenzymatic approach was
developed for the construction of compound 1.5
The aldehyde starting material 2 was ultimately derived
from D-serine. The synthesis sequence began with protection
of serine derivative 78 as a p-methoxybenzyl ether to afford
8 in 81% yield (Scheme 3). Reduction of the methyl ester
with LiBH4 smoothly provided alcohol 9 (88%), which was
subsequently oxidized with the Dess-Martin periodinane9
to give aldehyde 10 (93%). Exposure of aldehyde 10 to the
reagent derived from methyl (bromomethyl)acrylate 1110 and
indium led to the production of a mixture of diastereomeric
adducts 12 and 13 in excellent yield. The relative configura-
tions of 12 and 13 were determined by conversion to 16 and
the corresponding epimer at the benzoyloxy-bearing carbon,
respectively, and subsequent analysis of coupling constants
in the NMR spectra. There was no evidence for racemization
of 10 during the course of the reaction.11 The initial trials of
this reaction, which were conducted using THF as solvent,
afforded compounds 12 and 13 in essentially equal quantities.
The poor diastereoselectivity was somewhat unexpected
because it had been reported that with related 2-aminoalde-
hydes involving carbamate protecting groups, the desired syn
In designing a synthesis plan (Scheme 2), it was recog-
nized that the basic components required for the hypothetical
enzyme-mediated aldol condensation shown in Scheme 1
might be useful for a nonenzymatic route. It was then
surmised that the protected target compound, represented by
6 in Scheme 2, could be derived from building blocks such
as 2 and 3. A key step upon which the synthesis would hinge
is the addition of a nucleophile, represented by 3, to the
serine-derived aldehyde 2 to provide 4. In actuality, an
indium-mediated nucleophilic addition (M ) InLn) was
chosen to accomplish this transformation, primarily because
this method had previously been used to synthesize full
length sialic acids.6 These indium-mediated reactions are
known to be compatible with a variety of functional groups,
including those which are not orthogonal to more reactive
(3) (a) Wong, C.-H.; Halcomb, R. L.; Ichikawa, Y.; Kajimoto, T. Angew.
Chem., Int. Ed. Engl. 1995, 34, 412. (b) Wong, C.-H.; Halcomb, R. L.;
Ichikawa, Y.; Kajimoto, T. Angew. Chem., Int. Ed. Engl. 1995, 34, 521.
(c) Wong, C.-H.; Whitesides, G. M. Enzymes in Synthetic Organic
Chemistry; Elsevier: New York, 1994.
(4) Kim, M.-J.; Hennen, W. J.; Sweers, H. M.; Wong, C.-H. J. Am. Chem.
Soc. 1988, 110, 6481.
(5) A synthesis of a similar compound, which utilized D-gluconolactone
as the starting material, was previously reported: Vlahov, I. R.; Vlahova,
P. I.; Schmidt, R. R. Tetrahedron Lett. 1991, 32, 7025. For recent synthetic
studies, see: Burke, S. D.; Sametz, S. M. Org. Lett. 1999, 1, 71 and
references therein.
Scheme 4
(6) (a) Gordon, D. M.; Whitesides, G. M. J. Org. Chem. 1993, 58, 7939.
(b) Chan, T. H.; Lee, M. C. J. Org. Chem. 1995, 60, 4228. (c) Choi, S. K.;
Lee, S.; Whitesides, G. M. J. Org. Chem. 1996, 61, 8739. (d) Chan, T. H.;
Xin, Y.-C.; von Itzstein, M. J. Org. Chem. 1997, 62, 3500.
(7) (a) Cintas, P. Synlett 1995, 1087. (b) Li, C. J.; Chan, T. H. Organic
Reactions in Aqueous Media; John Wiley & Sons: New York, 1997. (c)
Paquette, L. A.; Rothhaar, R. R.; Isaac, M.; Rogers, L. M.; Rogers, R. D.
J. Org. Chem. 1998, 63, 5463. (d) Chan, T. H.; Lu, W. Tetrahedron Lett.
1998, 39, 8605.
(8) Hassall, C. H.; Thomas, J. O. J. Chem. Soc. 1968, 1495.
(9) Dess, D. B.; Martin, J. J. Org. Chem. 1983, 48, 4155.
(10) Anzeveno, P. B.; Campbell, J. A. Synth. Commun. 1986, 16, 387.
(11) Conversion of 12 and 13 into the corresponding R-(+)- and S-(-
)-Mosher esters showed that no significant aldehyde racemization occurred
before or during the indium-mediated addition. Dale, J. A.; Mosher, H. S.
J. Am. Chem. Soc. 1973, 95, 512.
(12) Paquette, L. A.; Mitzel, T. M.; Isaac, M. B.; Crasto, C. F.; Schomer,
W. W. J. Org. Chem. 1997, 62, 4293.
2004
Org. Lett., Vol. 2, No. 14, 2000

Scheme 5
isomer was significantly more predominant.12 It was also
mixture of R and â anomers, as determined by NMR
spectroscopy and by subsequent transformations. There was
no evidence of the production of any compounds derived
from reductive amination pathways during the latter opera-
tion. It should be noted that synthesis pathways in which
the acetamide was in place from the beginning of the
sequence were not fruitful but were plagued with low yields
and side reactions.
In summary, an efficient method has been developed for
the synthesis of a six-carbon truncated sialic acid analogue.
A key step in this synthesis was the indium-mediated
coupling of an allyl nucleophile with a protected serine
aldehyde, which allowed the presence of the somewhat
reactive methyl ester. Despite the limited stereoselectivity
of the allyl addition, a Mitsunobu inversion of the minor
diastereomer made it feasible to transform both C-4 epimers
into a common intermediate in the synthesis sequence. In
addition, the mild deprotection chemistry and in situ acylation
of the amino group should allow this sequence to be
amenable to the construction of similar compounds contain-
ing a variety of C-5 amides.
reported that changes in the ionic strength of the reaction
solvent can provide slight increases in the syn:anti diaster-
eomeric ratio in such systems.7,12 Consequently, variations
on the reaction were attempted using 1 M NH4Cl as the
solvent or cosolvent. Reactions conducted in 1 M aqueous
NH4Cl had significantly shortened reaction times but did not
exhibit substantially different stereoselectivities. Reactions
conducted in solvents composed of combinations of THF
and aqueous NH4Cl in various ratios had outcomes similar
to those in THF:H2O mixtures. Ultimately, it was found that
reactions performed in 1:1 THF:H2O provided the highest
combination of diastereomeric ratios (1.4:1) and isolated
yields (88%).
Fortunately, the syn and anti diastereomers 12 and 13 were
easily separable by silica gel chromatography, and both
compounds could both be converted to a common intermedi-
ate through a single synthesis operation (Scheme 4). There-
fore, no additional attempts to optimize the diastereoselec-
tivity beyond those discussed above were undertaken.
Protection of compound 12 as its benzoate ester was
accomplished by acylation with benzoic anhydride in the
presence of 4-(dimethylamino)pyridine (DMAP). Compound
14 was straightforwardly obtained in 71% yield. Minor
diastereomer 13 was also converted to intermediate 14 in
64% yield by implementing a Mitsunobu inversion of the
alcohol stereocenter, with benzoate being employed as the
nucleophile. It should be noted that di-tert-butyl azodicar-
boxylate (DTAD) was chosen over the more traditional
diethyl azodicarboxylate because it provided higher yields.
Deprotection of the p-methoxybenzyl ether was performed
by treatment of 14 with DDQ to afford alcohol 15, with no
evidence of benzoate migration (Scheme 5). Ozonolysis
followed by a reductive workup provided an R-ketoester
intermediate, which spontaneously cyclized to provide
hemiketal 16. Finally, the benzyl carbamate was converted
to an acetamide by hydrogenolysis in the presence of acetic
anhydride. The product 17 existed as an inseparable 10:1
Acknowledgment. This research was supported by the
NIH (GM55852). This work was greatly facilitated by a 500
MHz NMR spectrometer that was purchased partly with
funds from an NSF Shared Instrumentation Grant (CHE-
9523034). M.D.C. thanks the University of Colorado Gradu-
ate School for a Dean’s Small Grant Award. R.L.H. also
thanks the National Science Foundation (Career Award), the
Camille and Henry Dreyfus Foundation (Camille Drefus
Teacher-Scholar Award), Pfizer, and Novartis for support.
Supporting Information Available: Experimental pro-
cedures for the preparation of and spectral data for com-
pounds 8-10 and 12-14 (by both methods) and 15-17.
This material is available free of charge via the Internet at
http://pubs.acs.org.
OL990320R
Org. Lett., Vol. 2, No. 14, 2000
2005