Formal synthesis of (-)-N-acetylneuraminic acid (Neu5Ac) via desymmetrization by ring-closing metathesis

Article abstract of DOI:10.1016/S0040-4039(01)01912-8

A formal total synthesis of (-)-N-acetylneuraminic acid (Neu5Ac), the most naturally abundant sialic acid, has been accomplished using a rigid 6,8-dioxabicyclo[3.2.1]octane template for stereoselective introduction of all oxygen and nitrogen functionality

Full text of DOI:10.1016/S0040-4039(01)01912-8

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 42 (2001) 8747–8749
Formal synthesis of (−)-N-acetylneuraminic acid (Neu5Ac) via
desymmetrization by ring-closing metathesis
Eric A. Voight, Christian Rein and Steven D. Burke*
Department of Chemistry, University of Wisconsin-Madison, 1101 University Avenue, Madison, WI 53706, USA
Received 3 October 2001; revised 10 October 2001; accepted 11 October 2001
Abstract—A formal total synthesis of (−)-N-acetylneuraminic acid (Neu5Ac), the most naturally abundant sialic acid, has been
accomplished using a rigid 6,8-dioxabicyclo[3.2.1]octane template for stereoselective introduction of all oxygen and nitrogen
functionality. The template was obtained via a novel ketalization/ring-closing metathesis bond construction strategy, taking
advantage of an advanced intermediate in our KDN synthesis to complete the efficient assembly of Neu5Ac. © 2001 Elsevier
Science Ltd. All rights reserved.
We recently reported an efficient route to (+)-3-deoxy-
-glycero- -galacto-2-nonulosonic acid (KDN, Fig. 1)
yielding reaction sequences. The efficient total synthesis
of Neu5Ac therefore remains an important goal.
D
D
exploiting the rigid 6,8-dioxabicyclo[3.2.1]octane ring
system for the rapid stereoselective introduction of all
oxygen functionality.1 To further showcase the utility of
our intermolecular ketalization/intramolecular CꢀC
bond formation strategy for substrate desymmetriza-
Neu5Ac differs from KDN only at C-5 (Fig. 1) in the
presence of an acetamido group in place of the C-5
hydroxyl of KDN. In our retrosynthesis of Neu5Ac
(Fig. 2), azidoalcohol 2, with its four axial substituents
and electron rich 3,4-dimethoxyphenyl carboxyl surro-
gate at the anomeric carbon, was seen as a suitable
precursor to 1.1 Tetraacetate 3,1 a late stage KDN
intermediate, was chosen as a suitable convergency
point, requiring a double inversion at C-5 via an inter-
mediate epoxide to obtain 2. Diene 4 was previously
converted to inverted tetraacetate 3 via bis-dihydroxyla-
tion,1,2 the synthesis of 5-acetamido-3,5-dideoxy-
D
-glyc-
ero-
D-galacto-2-nonulosonic acid [N-acetylneuraminic
acid (Neu5Ac, 1), Fig. 1] has been pursued. Neu5Ac
was first isolated in 19513 and has since become the
most studied member of the many known sialic acids
primarily due to its vital biological role as a terminal
residue in glycoproteins, glycolipids, and oligosaccha-
rides in cellular membranes and nerve tissue.4
OMe
OH
N3
OH
HO
AcHN
HO
The chemical synthesis of Neu5Ac has been carried out
by several groups due to the fact that enzymatic synthe-
ses provide only very limited access to Neu5Ac
analogs.5 Each of these chemical syntheses relies on
four sugar-derived stereocenters or lengthy and low
O
O
CO2H
OMe
Si(iPr)2
O
OH
OH
O
O
Neu5Ac
O
(iPr)2Si
1
2
OMe
OMe
OMe
OAc
OMe
5
O
O
OH
OH
8
OH
OH
2
HO
HO
6
OAc
OAc
O
OAc
O
O
O
CO2H
CO2H
1
HO
HO
AcHN
HO
3
3
4
OH
KDN
OH
MeO
OMe
Neu5Ac
O
1
MeO
MeO
Br
O
O
OH
+
Figure 1. Sialic acids.
OH
7
5
6
* Corresponding author. Tel.: (608) 262-4941; fax: (608) 265-4534;
e-mail: burke@chem.wisc.edu
Figure 2. Retrosynthetic analysis.
0040-4039/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(01)01912-8

8748
E. A. Voight et al. / Tetrahedron Letters 42 (2001) 8747–8749
OMe
OMe
OH
E+
OMe
E+
OMe
OH
OMe
OMe
OAc
OH
OMe
NaOMe
TIPSCl2
OH
5
OMe
O
O
OH
5
9
O
MeOH
99 %
DMF/Im.
90 %
O
OAc
4
8
HO
OAc
O
OAc
OH
O
O
4
OH
9
OH
O
3
11
Si(iPr)2
O
O
OH
O
OMe
OMe
(iPr)2Si
OMe
OMe NaN3,
MgSO4
E+
OH
9
TsCl, Et3N,
8
DMAP, CH2Cl2;
:Nu
5
O
O
OMe
O
DMF
59 %
NaOMe/MeOH
4
O
O
OMe
O
5
OH
O
O
74 %
Si(iPr)2
Si(iPr)2
O
O
O
O
O
(iPr)2Si
(iPr)2Si
9
10
O
O
OMe
(iPr)2Si
H
OMe
O
N3
Amberlite (H+)
MeOH; TBAF,
OAc
O
OMe
(iPr)2Si
10
O
AcO
OMe
OMe
O
N3
O
Figure 3. Chemo- and regioselectivities facilitated by rigid
6,8-dioxabicyclo[3.2.1]octane template.
THF, Ac2O,
DMAP, Et3N
86 %
OH
O
O
(iPr)2Si
OAc
AcO
Si(iPr)2
O
12
2
tion and selective stereochemical inversion at C-4, and
4 was obtained via a ketalization/elimination/ring-clos-
ing metathesis (RCM) sequence,1,2 breaking the C2
symmetry of diene diol 6.
Scheme 2. Synthesis of methyl glycoside 12.
lyst was added slowly to a CH2Cl2 solution of triene 5
at rt, affording 5-aryl-7-vinyl-6,8-dioxabicyclo[3.2.1]-
oct-2-ene 4 in an improved yield (93%). A double
Sharpless asymmetric dihydroxylation gave 8 (Fig. 3),
and selective disulfonylation of the C-4 and C-9
hydroxyls, peracetylation, and C-4 inversion with
cesium acetate provided inverted tetraacetate 3 in 76%
overall yield.1 At this point the KDN and Neu5Ac
syntheses diverge.
Fig. 3 highlights the chemo- and regioselectivities
expected for hydroxyl sulfonylation (8 and 9) and epox-
ide nucleophilic ring opening (10) in the KDN and
Neu5Ac intermediates. In the KDN synthesis, the rigid
6,8-dioxabicyclo[3.2.1]octane template in 8 allowed
complete selectivity for disulfonylation at the C-4 equa-
torial hydroxyl and the C-9 primary alcohol using a
dibutyltin oxide-catalyzed tosylation reaction.1 For the
pursuit of Neu5Ac, the C-5 exo-hydroxyl in 9 was
expected to be more reactive towards sulfonylation
than the C-4 endo-hydroxyl, allowing selective a-epox-
ide formation. Furthermore, epoxide 10 was expected
to suffer nucleophilic attack selectively at C-5 in order
to allow a chair-like transition state via trans-diaxial
ring opening.6 Each of these reactions would be depen-
dent upon the steric and conformational constraints of
the rigid 6,8-dioxabicyclo[3.2.1]octane ring system.
With multigram quantities of 3 in hand, methanolysis
of tetraacetate 3 proceeded readily employing NaOMe
(0.25 equiv.) in MeOH (99%, Scheme 2). Selective
protection of the C-8 and C-9 hydroxyls of the resulting
tetraol 11 was best achieved under basic conditions
using 1,3-dichlorotetraisopropyl disiloxane (1.5 equiv.)
and imidazole (3 equiv.) in DMF, affording the tetra-
isopropyl disiloxane 9 (90%).8 Selective sulfonylation
using p-toluenesulfonyl chloride (2.1 equiv.) and tri-
ethyl amine (4.2 equiv.) with catalytic 4-dimethyl-
aminopyridine gave a C-5 tosylate that could be
displaced by the C-4 hydroxyl using NaOMe (4.2
equiv.) in MeOH to provide epoxide 10 (74%).9,10 After
some experimentation, it was found that epoxide
azidolysis could best be accomplished using sodium
azide (10.6 equiv.) and magnesium sulfate (2.1 equiv.)
in DMF (0.1 M) at 90–95°C, giving azidoalcohol 2
(59%).11 Without magnesium sulfate, only 50% conver-
sion could be achieved after 2 days, giving a 2:1 ratio of
epoxide opening regioisomers. Methanolysis of 2 pro-
ceeded readily employing Amberlite IR-120 (plus) acid
resin,12 providing methyl glycoside 12 after removal of
the TIPS group and peracetylation.
An improved route to 31 for application to the synthesis
of Neu5Ac is shown in Scheme 1. An acid catalyzed
ketalization/selenoxide elimination sequence, starting
with diene diol 6 and ketone 7, afforded RCM precur-
sor 5 in 86% overall yield (Scheme 1). It was found that
the RCM reaction could be performed with only 1%
Grubbs’ ruthenium benzylidene catalyst7 when the cata-
MeO
OMe
1) CSA, PhH
2) NaBH4, EtOH
O
OH
OH
MeO
MeO
Br
O
O
+
o-NO2PhSeCN;
aq. H2O2, THF
86% overall
7
6
5
PCy3
OMe
OAc
To complete the synthesis of Neu5Ac, oxidative cleav-
age of the 3,4-dimethoxyphenyl carboxyl surrogate was
achieved using RuO4, formed in situ by adding a cata-
lytic amount of RuCl3·3H2O (0.05 equiv.) to a 2:2:3
CCl4/CH3CN/H2O mixture containing 12 and NaIO4
(12 equiv.),1 providing azidoNeu5Ac derivative 13
(66%, Scheme 3). Reductive acetylation of the azide in
13 gave 14 (69%),5d which has been previously con-
Cl
Cl
Ph
OMe
OMe
Ru
PCy3
OMe
OAc
3 steps
O
O
4
9
CH2Cl2
93 %
76 %
OAc
O
OAc
O
3
4
Scheme 1. Synthesis of tetraacetate 3.

E. A. Voight et al. / Tetrahedron Letters 42 (2001) 8747–8749
8749
cat. RuCl3•3H2O
NaIO4, RT
CCl4/CH3CN/H2O;
tion; Gottschalk, A., Ed. The structure of sialic acids and
their quantitation; Elsevier: Amsterdam, 1972; p. 403.
For chemical syntheses not yet reviewed, see: (d) Li,
L.-S.; Wu, Y.-L.; Wu, Y. Org. Lett. 2000, 2, 891–894; (e)
Kang, S. H.; Choi, H.-w.; Kim, J. S.; Youn, J.-H. Chem.
Commun. 2000, 227–228; (f) Banwell, M.; Savi, C. D.;
Watson, K. J. Chem. Soc., Perkin Trans. 1 1998, 2251–
2252; (g) Takahashi, T.; Tsukamoto, H.; Kurosaki, M.;
Yamada, H. Synlett 1997, 1065–1066.
OAc
OMe
OMe
AcO
AcO
AcO
OMe
OMe
O
N3
AcO
TMSCHN2
PhH/MeOH, RT
66 %
OAc
12
OAc
H2, Pd/C, MeOH;
Ac2O, Et3N,
O
CO2Me
N3
AcO
OAc
DMAP, CH2Cl2
69 %
6. (a) For a recent example, see: Schmidt, B. J. Chem. Soc.,
Perkin Trans. 1 1999, 2627–2637; (b) For a review, see:
Parker, R. E.; Isaacs, N. S. Chem. Rev. 1959, 59, 737.
7. For recent reviews, see: (a) Fu¨rstner, A. Angew. Chem.,
Int. Ed. 2000, 39, 3012–3043; (b) Schrock, R. R. Tetra-
hedron 1999, 55, 8141–8153; (c) Grubbs, R. H.; Chang, S.
Tetrahedron 1998, 54, 4413–4450; (d) Armstrong, S. K. J.
Chem. Soc., Perkin Trans. 1 1998, 371–388; (e) Schuster,
M.; Blechert, S. Angew. Chem., Int. Ed. Engl. 1997, 36,
2036–2056.
13
OAc
OMe
O
CO2Me
AcHN
AcO
OAc
14
Scheme 3. Completion of the formal synthesis of Neu5Ac.
8. (a) Paulsen, H.; Heitmann, A. C. Liebigs Ann. Chem.
1988, 11, 1061–1071; (b) Zbiral, E.; Schmid, W. Monatsh.
Chem. 1985, 116, 253–262.
9. Takatori, K.; Kajiwara, M. Synlett 1995, 3, 280–282.
10. Surprisingly, attempted epoxidation using Mitsunobu
conditions gave almost exclusively the undesired b-epox-
ide despite close precedent suggesting that 10 should be
the major product: Thomson, R.; von Itzstein, M. Carbo-
hydr. Res. 1995, 274, 29–44.
verted to Neu5Ac 1 using 1N aq. sodium hydroxide
followed by acidification with aq. HCl (70%).13a,d The
1H NMR, TLC Rf, IR, MS, and optical rotation data
for 14 matched those reported in the literature.13,14 The
formal total synthesis of Neu5Ac has thus been
achieved in 9.3% overall yield from diene diol 6
employing our convergent ketalization/ring-closing
metathesis strategy and exploiting the resulting rigid
6,8-dioxabicyclo[3.2.1]octane template. The develop-
ment and application of this strategy to other natural
products and ring systems is currently being explored.
11. Monneret, C.; Risse, S.; Ardouin, P.; Gouyette, A. Eur.
J. Med. Chem. Chim. Ther. 2000, 35, 137–146.
12. (a) Baker, R.; Boyes, H. O.; Broom, M. P.; O’Mahony,
M. J.; Swain, C. J. J. Chem. Soc., Perkin Trans. 1 1987,
1613–1621; (b) Ka˜le´, V. N.; Clive, D. L. J. J. Org. Chem.
1984, 49, 1554–1563.
Acknowledgements
13. (a) Julina, R.; Mu¨ller, I.; Vasella, A. Carbohyd. Res.
1987, 164, 415–432; (b) Okamoto, K.; Kondo, T.; Goto,
T. Bull. Chem. Soc. Jpn. 1987, 60, 631–636; (c) Ogura, H.;
Furuhata, K. Carbohyd. Res. 1986, 158, 37–51; (d) Zbiral,
E.; Brandstetter, H. H. Monatsh. Chem. 1985, 116, 87–98;
(e) Ponpipom, M. M.; Buglianesi, R. L.; Shen, T. Y. Can.
J. Chem. 1980, 58, 214–220.
We thank the NIH [Grant CA74394 (SDB) and CBI
Training Grant 5 T32 GM08505 (EAV)] for generous
support of this research. The NIH (1 S10 RR0 8389-01)
and NSF (CHE-9208463) are acknowledged for their
support of the NMR facilities of the University of
Wisconsin-Madison Department of Chemistry.
1
14. Data for 14: H NMR (CDCl3) l 5.41 (dd, J=4, 2 Hz,
1H), 5.31 (d, J=10 Hz, 1H), 5.25 (ddd, J=11.5, 10.5, 5
Hz, 1H), 5.23 (ddd, J=7.5, 2.5, 2 Hz, 1H), 4.81 (dd,
J=12.5, 2.5 Hz, 1H), 4.13 (ddd, J=10.5, 10.5, 10 Hz,
1H), 4.12 (dd, J=12.5, 7.5 Hz, 1H), 3.93 (dd, J=10.5, 2.5
Hz, 1H), 3.82 (s, 3H), 3.27 (s, 3H), 2.44 (dd, J=13, 5 Hz,
1H), 2.15 (s, 3H), 2.09 (s, 3H), 2.04 (s, 3H), 2.02 (s, 3H),
1.89 (s, 3H), 1.89 (dd, J=13, 11.5 Hz, 1H); 13C NMR
(CDCl3) l 171.0 (C), 170.7 (C), 170.6 (C), 170.2 (C),
170.1 (C), 167.3 (C), 98.9 (C), 72.0 (CH), 71.7 (CH), 68.8
(CH), 68.4 (CH), 62.4 (CH2), 52.7 (CH3), 51.3 (CH3),
49.3 (CH), 37.3 (CH2), 23.2 (CH3), 21.0 (CH3), 20.9
(CH3), 20.8 (CH3×2); IR (thin film) 1745, 1664, 1541,
1371, 1225, 1038 cm−1; [h]D22 −12° (c=0.67, CHCl3); mp
131–133°C; HRMS (ESI) calcd. for C21H31O13NaN (M+
Na+) 528.1693, found 528.1675.
References
1. Burke, S. D.; Voight, E. A. Org. Lett. 2001, 3, 237–240.
2. Burke, S. D.; Mu¨ller, N.; Beaudry, C. M. Org. Lett. 1999,
1, 1827–1829.
3. Gottschalk, A. Nature 1951, 167, 845.
4. Schauer, R.; Kelm, S.; Reuter, G.; Roggentin, P.; Shaw,
L. In Biology of The Sialic Acids; Rosenberg, A., Ed.
Biochemistry and role of sialic acids; Plenum: New York,
1995; p. 7.
5. For reviews, see: (a) von Itzstein, M.; Thompson, R. J.
Top. Curr. Chem. 1997, 186, 119–170; (b) DeNinno, M.
P. Synthesis 1991, 583–593; (c) Tuppy, H.; Gottschalk, A.
In Glycoproteins. Their Composition, Structure and Func-