A new approach to neuraminic acid analogues via 1,2-oxazines

Article abstract of DOI:10.1021/ol802514m

(Chemical Equation Presented) A new stereoselective and potentially very flexible (C₅ + C₃ + C₁) approach to neuramlnic acid derivatives and analogues has been established using enantiopure nitrones and alkoxyallenes as C<

Full text of DOI:10.1021/ol802514m

ORGANIC
LETTERS
2009
Vol. 11, No. 3
527-530
A New Approach to Neuraminic Acid
Analogues via 1,2-Oxazines
Bettina Bressel and Hans-Ulrich Reissig*
Institut fu¨r Chemie und Biochemie, Freie UniVersita¨t Berlin,
Takustr. 3, 14195 Berlin, Germany
hans.reissig@chemie.fu-berlin.de
Received October 30, 2008
ABSTRACT
A new stereoselective and potentially very flexible (C5 + C3 + C1) approach to neuraminic acid derivatives and analogues has been established
using enantiopure nitrones and alkoxyallenes as C3 and C1 building blocks. Substituent OR2 in position 4 of neuraminic acid analogues is
defined by the alkoxyallene employed for the synthesis of the intermediate 1,2-oxazine. Side chain R1 can be varied by using different precursor
nitrones and introduction of different protection groups R3 at the amino function is also possible.
Neuraminic acid (1) and its derivatives have important
biological functions.1 It has been demonstrated that several
derivatives, in particular synthetic analogues, inhibit neuramini-
dases. This ability can be exploited for the treatment of
influenza infections as it is actually achieved by drugs such
as Relenza (zanamivir, 2)2 or Tamiflu (oseltamivir phosphate,
3, Figure 1).3 This ongoing interest and other applications
make efficient and flexible syntheses of neuraminic acid
analogues a matter of intense research.4
In this paper we present a new and potentially very flexi-
ble route to neuraminic acid analogues. As shown in the
retrosynthetic analysis (Scheme 1), neuraminic acid analogue
A can be obtained from protected R-ketoester B. The
corresponding R-hydroxyester C should be available by chain
elongation from aldehyde D by using lithiated alkoxyallene
(1) Biology of the Sialic Acids; Rosenberg, A., Ed.; Plenum Press: New
York, 1995.
Figure 1. Neuraminic acid 1 and its analogues 2 and 3.
(2) von Itzstein, M.; et al. Nature 1993, 363, 418–423.
(3) Kim, C. U.; Lew, W.; Williams, M. A.; Liu, H.; Zhang, L.;
Swaminathan, S.; Bischofberger, N.; Chen, M. S.; Mendel, D. B.; Tai, C. Y.;
Laver, W. G.; Stevens, R. C. J. Am. Chem. Soc. 1997, 119, 681–690.
(4) For recent reviews, see: (a) Hemeon, I.; Bennet, A. J. Synthesis 2007,
1899–1926. (b) Kiefel, M. J.; von Itzstein, M. Chem. ReV. 2002, 102, 471–
490.
E as a C1 building block.5 A primary amino alcohol which
can be oxidized to the crucial aldehyde D is accessible from
1,2-oxazine F by hydrogenolysis as shown previously by our
group.6 It is necessary to introduce an electron-withdrawing
group R3 at the amino function in order to avoid an
intramolecular cyclization of aldehyde D. The required 1,2-
(5) (a) Hormuth, S.; Reissig, H.-U.; Dorsch, D. Liebigs Ann. Chem. 1994,
121–127. (b) Schade, W.; Reissig, H.-U. J. Prakt. Chem. 1999, 341, 685–
686.
10.1021/ol802514m CCC: $40.75
Published on Web 12/30/2008
2009 American Chemical Society

Scheme 1
.
Retrosynthetic Analysis
Scheme 2. Reduction and Chain Elongation of 1,2-Oxazine 4
occurring derivatives.4,10 The enol ether double bond of 1,2-
oxazine 5 was reduced with hydrogen in the presence of Pd/
Al2O3 in ethyl acetate furnishing 1,2-oxazinane 6 as the major
component of an 82:18 mixture of diastereomers (1H NMR,
crude), which were easily separated by column chromatog-
raphy. Under these conditions, the N-acetylated N-O bond
was stable. Its cleavage was smoothly achieved by samarium
diiodide11 leading to amino alcohol 7. After Swern oxida-
tion12 to the aldehyde, addition of lithiated methoxyallene
and ozonolysis, the carbon-chain elongated R-hydroxyester
8 was isolated as a mixture of diastereomers (ca. 1:1).
The R-hydroxyester 8 could easily be oxidized to R-ke-
toester 9 using Dess-Martin periodinane (Scheme 3).13 In
oxazines F can be synthesized diastereoselectively from
enantiopure nitrones H using alkoxyallene G, now an
essential C3 building block.7 The alkoxy substituent OR2 will
appear in position 4 of the corresponding neuraminic acid
analogue. Different nitrones will allow the synthesis of side
chains varying in length and configuration.
To examine the proposed strategy, we decided to employ
1,2-oxazine 4 as a model compound which is smoothly
available from a L-glyceraldehyde-derived nitrone and lithi-
ated methoxyallene.8 Although most of our synthetic studies
have been performed with the enantiomer of 4, a sound basis
of its reactivity was available.6,8,9 Hydrogenolysis of 4 led
to debenzylation followed by diastereoselective reduction of
the double bond and finally N-O cleavage.6 To avoid
selectivity problems during the introduction of an N-
protecting group to the resulting amino alcohol, the hydro-
genolysis was interrupted before N-O cleavage and the
debenzylated derivatives were acetylated to give 1,2-oxazine
5 and 1,2-oxazinane 6 (Scheme 2). In addition, 28% of
starting material 4 could be reisolated. Longer reaction times
resulted in lower yields of desired products because of N-O
cleavage. The acetyl group was chosen for protection as it
has the required electron-withdrawing properties and N-
acetylneuraminic acids are the most important naturally
Scheme 3. Synthesis of Neuraminic Acid Analogue 11
CDCl3, a slow conversion of this product to the cyclic
derivative 10 was observed, which is probably induced by
traces of acid in the solvent. By mass spectrometry it was
confirmed that no elimination had taken place, suggesting
(6) Pulz, R.; Watanabe, T.; Schade, W.; Reissig, H.-U. Synlett 2000,
983–986.
(7) Helms, M.; Schade, W.; Pulz, R.; Watanabe, T.; Al-Harrasi, A.;
Fisˇera, L.; Hlobilova´, I.; Zahn, G.; Reissig, H.-U. Eur. J. Org. Chem. 2005,
1003–1019.
(10) Angata, T.; Varki, A. Chem. ReV. 2002, 102, 439–469
.
(8) Bressel, B.; Egart, B.; Al-Harrasi, A.; Pulz, R.; Reissig, H.-U.;
Bru¨dgam, I. Eur. J. Org. Chem. 2008, 467–474.
(11) Keck, G. E.; McHardy, S. F.; Wagner, T. T. Tetrahedron Lett. 1995,
36, 7419–7422. (b) Chiara, J. L.; Destabel, C.; Gallego, P.; Marco-Contelles,
J. J. Org. Chem. 1996, 61, 359–360. (c) Young, I. S.; Williams, J. L.; Kerr,
M. A. Org. Lett. 2005, 7, 953–955. (d) Revuelta, J.; Cicchi, S.; Brandi, A.
Tetrahedron Lett. 2004, 45, 8375–8377, and references cited therein. See
also 9c.
(9) (a) Al-Harrasi, A.; Reissig, H.-U. Synlett 2005, 1152–1154. (b) Al-
Harrasi, A.; Reissig, H.-U. Angew. Chem. 2005, 117, 6383–6387; Angew.
Chem. Int. Ed. 2005, 44, 6227-6231. (c) Pulz, R.; Al-Harrasi, A.; Reissig,
H.-U. Org. Lett. 2002, 4, 2353–2355
.
528
Org. Lett., Vol. 11, No. 3, 2009

Scheme 4. Synthesis of Neuraminic Acid Derivatives 22, 23, and N-Acetylneuraminic Acid Methyl Ester 24
that this cyclization should be reversible. The final acetal
cleavage could be achieved under acidic conditions. For a
similar derivative, Banwell et al. employed HCl in metha-
nol.14 We obtained good results using Amberlyst-15 in
methanol at 40 °C. The mixture of the esters 9 and 10 was
converted to the neuraminic acid analogue 11 with a methoxy
group in position 4 and a short side chain in an overall yield
of 3% starting from 1,2-oxazine 4. No attempts to optimize
this sequence have been made.
cleavage. This reactivity was ideal for the N-acetylation
leading directly to 1,2-oxazinane 17 in good yield without
any selectivity problems. Additionally, it was observed that
by decreasing the amount of catalyst in order to slow down
the speed of the reaction for easier monitoring by TLC, an
increasing amount of an 1,2-oxazin-4-one was isolated, which
apparently resulted from the enol formed by O-debenzylation
before reduction of the enol ether double bond. The
dependence of the product ratio on the amount of catalyst
cannot be fully explained yet and will be discussed in more
detail elsewhere.16,17 The N-O bonds of N-acetylated 1,2-
oxazinane derivatives 16 and 17 were cleaved with samarium
diiodide in excellent yields. The conversion of the resulting
aminoalcohols 18 and 19 to the required aldehydes was
achieved by Dess-Martin oxidation which led to better
results than Swern oxidation. After addition of lithiated
methoxyallene and ozonolysis the R-hydroxyesters 20 and
21 were formed in moderate yields over three steps (dr ca.
1:1). Dess-Martin oxidation to the corresponding R-ke-
toesters followed by acetal cleavage provided N-acetyl-
neuraminic acid derivatives 22 and 23 in moderate yields
over two steps. Similar R-ketoesters have been deprotected
using HCl/MeOH18 or HF.19 Here, the acetal cleavage was
performed with Amberlyst-15 in methanol. Derivative 22,
bearing a methoxy group in position 4, was obtained in an
overall yield of 19% over ten steps and derivative 23 with
the benzyloxy group in 11% yield over nine steps. Under
transfer hydrogenation conditions with cyclohexene, benzy-
After this proof of principle, we applied the strategy to
the synthesis of neuraminic acid derivatives with a complete
9-carbon backbone (Scheme 4) starting from the D-arabinose
derived nitrone 12,15 which is easily accessible from D-
mannitol. To introduce different substituents in position 4,
methoxy- and benzyloxyallene were used for the synthesis.
The methoxy-substituted 1,2-oxazine 13 has been prepared
before,7 but the yield could be improved to 89% by slight
variations of the original procedure. Partial hydrogenolysis
and acetylation led to 1,2-oxazine 15 and 1,2-oxazinane 16
in analogy to derivative 4, but gratifyingly the yields were
considerably higher. The hydrogenation of the enol ether
double bond of 15 was smoothly achieved with a diastereo-
selectivity of 93:7 (1H NMR, crude), giving pure 1,2-
oxazinane 16 after column chromatography. Benzyloxy-
substituted precursor 14 was also prepared from nitrone 12
in good yield and excellent diastereoselectivity. We were
wondering which functional group of 14 (enol ether moiety,
O-benzyl, or N-benzyl group) would be reduced first during
hydrogenation, assuming that the N-O bond cleavage would
still be the slowest step. When using 30 mol % of Pd/C in
methanol, the hydrogenolysis of the N-benzyl group and the
reduction of the double bond was faster than O-benzyl
(16) Bressel, B. Ph.D. Dissertation, Freie Universita¨t Berlin, 2008.
(17) As the rate of hydrogenation generally increases with the amount
of starting material and high amounts of catalyst are necessary to achieve
good yields of the desired product 17, up-scaling of this partial hydro-
genolysis is limited. If the reaction is too fast, it is not possible to stop it
before O-debenzylation and N-O cleavage occur.
(12) Omura, K.; Swern, D. Tetrahedron 1978, 34, 1651–1660.
(13) Dess, D. B.; Martin, J. C. J. Org. Chem. 1983, 48, 4155–4156.
(14) Banwell, M.; De Savi, C.; Hockless, D.; Watson, K. Chem.
Commun. 1998, 645–646.
(18) Kang, S. H.; Choi, H.; Kim, J. S.; Youn, J.-H. Chem. Commun.
2000, 227–228.
(19) (a) Liu, K.-G.; Yan, S.; Wu, Y.-L.; Yao, Z.-J. J. Org. Chem. 2002,
67, 6758–6763. (b) Liu, K.-G.; Yan, S.; Wu, Y.-L.; Yao, Z.-H. Org. Lett.
2004, 6, 2269–2272.
(15) Dondoni, A.; Franco, S.; Junquera, F.; Mercha´n, F.; Merino, P.;
Tejero, T. Synth. Commun. 1994, 24, 2537–2550.
Org. Lett., Vol. 11, No. 3, 2009
529

loxy derivative 23 was debenzylated quantitatively to afford
the known20 N-acetylneuraminic acid methyl ester 24.
Applying an N-Boc protecting group instead of an N-acetyl
group to the synthesis shortened the number of required steps
because the N-Boc group was already introduced during the
hydrogenolysis by adding Boc2O to the reaction mixture
(Scheme 5). Amino alcohol 25 was obtained in good yield
methanol both compounds could individually be converted
into the neuraminic acid derivative 29 in yields of 46% and
53%, respectively. The overall yield for N-Boc-protected
derivative 29 is 12% over eight steps.
During conversion of R-ketoester 27 to 29 the reaction
has to be interrupted before completion because as a
byproduct pyrrole 30 is slowly formed. Longer reaction times
or higher temperatures led to a decrease of neuraminic acid
derivative 29 and an increase of pyrrole derivative 30 in up
to 30% yield. The biological activity of 30 may be interesting
since its structure is closely related to a known immuno-
suppressive compound.21
Scheme 5. Synthesis of Neuraminic Acid Derivative 29
In conclusion, a new C5 + C3 + C1 strategy to neuraminic
acid derivatives has successfully been devised using nitrones
and alkoxyallenes as easily available starting materials.
N-Acetyl derivatives 22-24 with different 4-substituents
were stereoselectively prepared in reasonable quantities and
with good overall yields. Similarly, the N-Boc-protected
derivative 29 has been synthesized. The 4-substituent depends
on the alkoxyallene used for the synthesis of the 1,2-oxazine.
The structure of drugs such as zanamivir 2 and oseltamivir
3 reveals that variation of 4-substituents may be particularly
important for new neuraminidase inhibitors. Starting from
other nitrones the methodology developed also allows for
the preparation of neuraminic acid analogues with differing
chain lengths as demonstrated by the preparation of C7-
analogue 11. These results show the flexibility of this new
alkoxyallene-based22 approach to neuraminic acid deriva-
tives. Modifications of substituents and the configuration in
different positions should be easily possible and will be
investigated in due time.
Acknowledgment. Generous support of this work by the
Deutsche Forschungsgemeinschaft, the Fonds der Chemi-
schen Industrie, and the Bayer Schering Pharma AG is most
gratefully acknowledged. We also thank Dr. M. Helms (Freie
Universita¨t Berlin) for preliminary experiments.
(dr ca. 95:5). As the Boc-group has a weaker electron-
withdrawing effect than the acetyl group, the Swern oxidation
had to be carefully executed applying short reaction times
and avoiding acidic workup. Otherwise an intramolecular
cyclization occurs leading to a pyrrole derivative by elimina-
tion of water and methanol. As above, the aldehyde can be
transformed into the R-hydroxyester 26 by addition of
lithiated methoxyallene and ozonolysis in moderate yield
over three steps (dr ca. 1:1). After Dess-Martin oxidation
to R-ketoester 27, the cyclic derivative 28 was isolated as a
side product. By acetal cleavage with Amberlyst-15 in
Supporting Information Available: Experimental pro-
cedures and complete characterization for all new com-
pounds. This material is available free of charge via the
Internet at http://pubs.acs.org.
OL802514M
(21) (a) Schwab, S. R.; Pereira, J. P.; Matloubian, M.; Xu, Y.; Huang,
Y.; Cyster, J. G. Science 2005, 309, 1735–1739. For isolation and
characterization of this compound, see: (b) Kro¨plien, U.; Rosdorfer, J. J.
Org. Chem. 1985, 50, 1131–1133. For the synthesis, see: (c) Sweeny, J. G.;
Ricks, E.; Estrada-Valde´s, M. C.; Iacobucci, G. A. J. Org. Chem. 1985,
50, 1133–1134.
(20) (a) Kuhn, R.; Lutz, P.; MacDonald, D. L. Ber. Dtsch. Ges. 1966,
99, 611–617. (b) Derevitskaya, V. A.; Kalinevich, V. M.; Kochetkov, N. K.
Dokl. Akad. Nauk SSSR 1965, 160, 596–599; Dokl. Chem., Engl. 1965,
160, 88-91.
(22) Recent review on the synthetic versatility of lithiated alkoxyallenes:
Brasholz, M.; Reissig, H.-U.; Zimmer, R. Acc. Chem. Res. 2009, 42, 33-44.
530
Org. Lett., Vol. 11, No. 3, 2009