Total synthesis of 3-deoxy-D-manno-2-octulosonic acid (KDO) and 2-deoxy-β-KDO

Article abstract of DOI:10.1021/ol9905506

(equation presented) Total syntheses of KDO and 2-deoxy-β-KDO are reported. The C₂-symmetric dienediol 4 was desymmetrized by conversion to its corresponding 1,4-dioxanone 5. Ireland-Claisen rearrangement of 5 provided the 6-vinyldihydropyran-2-carboxylate template 6. Double-Sharpless asymmetric dihydroxylation gave the tetraol 7a, which was converted to KDO and 2-deoxy-β-KDO using methods similar to those previously reported. This synthetic scheme provides a flexible route to KDO and KDO analogues.

Full text of DOI:10.1021/ol9905506

ORGANIC
LETTERS
1999
Vol. 1, No. 1
71-74
Total Synthesis of
3-Deoxy-D-manno-2-octulosonic Acid
(KDO) and 2-Deoxy-â-KDO
Steven D. Burke* and Geoffrey M. Sametz
Department of Chemistry, UniVersity of WisconsinsMadison, 1101 UniVersity AVenue,
Madison, Wisconsin 53706-1396
burke@chem.wisc.edu
Received April 1, 1999
ABSTRACT
Total syntheses of KDO and 2-deoxy-â-KDO are reported. The C -symmetric dienediol 4 was desymmetrized by conversion to its corresponding
2
1,4-dioxanone 5. Ireland−Claisen rearrangement of 5 provided the 6-vinyldihydropyran-2-carboxylate template 6. Double-Sharpless asymmetric
dihydroxylation gave the tetraol 7a, which was converted to KDO and 2-deoxy-â-KDO using methods similar to those previously reported. This
synthetic scheme provides a flexible route to KDO and KDO analogues.
KDO (3-deoxy-D-manno-2-octulosonic acid, 1) is a key
component of the cell wall lipopolysaccharide (LPS) of
Gram-negative bacteria. KDO residues form the necessary
linkage between the polysaccharide and lipid A regions of
LPS.1 The enzyme CMP-KDO synthetase catalyzes what is
believed to be the rate-limiting step of KDO incorporation
into LPS,2 and inhibitors of this enzyme have attracted
interest as potential antibiotics. The most potent of these to
date is 2-deoxy-â-KDO (2).3 Most of the previous syntheses
of 11,4,5,6 and 24b,7,8 are chemical or enzymatic semisyntheses
starting from advanced carbohydrate precursors, and thus are
not easily applicable to the synthesis of structurally diverse
analogues. We report flexible de noVo syntheses of 1 and 2,
34, 5089. (h) Haudrechy, A.; Sinay¨, P. J. Org. Chem. 1992, 57, 4142. (i)
Giese, B.; Linker, T. Synthesis 1992, 46. (j) Frick, W.; Kru¨lle, T.; Schmidt,
R. R. Liebigs Ann. Chem. 1991, 435. (k) Dondoni, A.; Merino, P. J. Org.
Chem. 1991, 56, 5294. (l) Ramage, R.; MacLeod, A. M.; Rose, G. W.
Tetrahedron 1991, 47, 5625. (m) Boons, G. J. P. H.; van der Klein, P. A.
M.; van der Marel, G. A.; van Boom, J. H. Recl. TraV. Chim. Pays-Bas
1990, 109, 273. (n) Horito, S.; Amano, M.; Hashimoto, H. J. Carbohydr.
Chem. 1989, 8, 681. (o) Esswein, A.; Betz, R.; Schmidt, R. R. HelV. Chim.
Acta 1989, 72, 213. (p) Branchaud, B. P.; Meier, M. S. J. Org. Chem. 1989,
54, 1320. (q) Shirai, R.; Ogura, H. Tetrahedron Lett. 1989, 30, 2263. (r)
Itoh, H.; Kaneko, T.; Tanami, K.; Yoda, K. Bull. Chem. Soc. Jpn. 1988,
61, 3356. (s) Imoto, M.; Kusumoto, S.; Shiba, T. Tetrahedron Lett. 1987,
28, 6235. (t) Collins, P. M.; Overend, W. G.; Shing, T. J. Chem. Soc.,
Chem. Commun. 1981, 1139.
(5) For enzymatic syntheses of KDO and related compounds, see: (a)
Kragl, U.; Go¨dde, A.; Wandrey, C.; Lubin, N.; Auge´, C. J. Chem. Soc.,
Perkin Trans. 1 1994, 119. (b) Sugai, T.; Shen, G.-J.; Ichikawa, Y.; Wong,
C.-H. J. Am. Chem. Soc. 1993, 115, 413.
(6) For de noVo syntheses of KDO and related compounds, see: (a)
Schlessinger, R. H.; Pettus, L. H. J. Org. Chem. 1998, 63, 9089. (b) Hu,
Y.-J.; Huang, X.-D.; Yao, Z.-J.; Wu, Y.-L. J. Org. Chem. 1998, 63, 2456.
(c) Lubineau, A.; Auge´, J.; Lubin, N. Tetrahedron 1993, 49, 4639. (d)
Martin, S. F.; Zinke, P. W. J. Org. Chem. 1991, 56, 6600. (e) Smith, D. B.;
Wang, Z.; Schreiber, S. L. Tetrahedron 1990, 46, 4793. (f) Danishefsky,
S. J.; DeNinno, M. P.; Chen, S. J. Am. Chem. Soc. 1988, 110, 3929.
(1) Unger, F. M. AdV. Carbohydr. Chem. Biochem. 1981, 38, 323.
(2) Ray, P. H.; Benedict, C. D.; Grasmuk, H. J. Bacteriol. 1981, 145,
1273.
(3) (a) Hammond, S. M.; Claesson, A.; Jansson, A. M.; Larsson, L.-G.;
Pring, B. G.; Town, C. M.; Ekstro¨m, B. Nature 1987, 327, 730. (b) Goldman,
R.; Kohlbrenner, W.; Lartey, P.; Pernet, A. Nature 1987, 329, 162.
(4) For chemical syntheses of KDO and related compounds starting from
advanced carbohydrate precursors, see: (a) Jiang, S.; Rycroft, A. D.; Singh,
G.; Wang, X.-Z.; Wu, Y.-L. Tetrahedron Lett. 1998, 39, 3809. (b) Lo´pez-
Herrera, F. J.; Sarabia-Garc´ıa, F. Tetrahedron 1997, 53, 3325. (c) Tsuka-
moto, H.; Takahashi, T. Tetrahedron Lett. 1997, 38, 6415. (d) Barton, D.
H. R.; Jaszberenyi, J. Cs.; Liu, W.; Shinada, T. Tetrahedron 1996, 52, 2717.
(e) Du, S.; Plat, D.; Baasov, T. Tetrahedron Lett. 1996, 37, 3545. (f) Gao,
J.; Ha¨rter, R.; Gordon, D. M.; Whitesides, G. M. J. Org. Chem. 1994, 59,
3714. (g) Coutrot, Ph.; Grison, C.; Tabyaoui, M. Tetrahedron Lett. 1993,
10.1021/ol9905506 CCC: $18.00 © 1999 American Chemical Society
Published on Web 05/17/1999

via a dihydropyran template 6 that may be applied toward
the synthesis of a variety of analogues.
procedure. Because of its acid sensitivity,15 6a was converted
immediately to its tert-butyl ester 6b by a modification of
Jackson’s procedure.16 Although boron trifluoride proved too
harsh a catalyst for the esterification, excellent yields were
obtained by relying on adventitious acid catalysis (although
small amounts of 6a persisted). The methyl ester of 6a was
also prepared, but suffered saponification under the subse-
quent Sharpless asymmetric dihydroxylation (SAD) condi-
tions.
Diene 6b was immediately carried into the attempted
double SAD,17 and after 3 days at 0 °C the diene was cleanly
converted to a mixture of 7a18 and its C(7) epimer 7b in
81% and 4% yields, respectively.19 It was hoped that the
endocyclic olefin could be dihydroxylated from the desired
R-face under SAD conditions. However, in all cases
surveyeds(DHQ)2-AQN, (DHQD)2-AQN, (DHQ)2-PHAL,
(DHQD)2-PHALsdihydroxylation occurred from the un-
desired â-face.17 The same facial selectivity was observed
in the dihydroxylation of 6b with OsO4/NMO, suggesting
that the facial bias imparted by the C(6) substituent overrides
any influence the SAD ligands may have. Qualitatively, use
of the new ligand (DHQ)2-AQN20 gave better diastereo-
selectivity than (DHQ)2-PHAL at C(7).
The starting material for our synthesis, the C2-symmetric
dienediol 4, initially was obtained using the method of Yadav
et al.9 However, the length of the synthesis (six steps from
D-mannitol), inconvenience of scale-up, and overall poor
yield (∼20%) prompted us to design a shorter synthesis. By
using a literature procedure, the dibromide 3 (Scheme 1) can
Scheme 1
Comparing the dihydroxylation of 6b using (DHQ)2-
AQN, (DHQD)2-AQN, and standard OsO4/NMO condi-
tions,21 we believe that the desired dihydroxylation at the
terminal olefin is a mismatched case, as evidenced by
(DHQD)2-AQN being more selective for the formation of
7b than (DHQ)2-AQN is for the formation of 7a, and OsO4/
NMO favoring 7b.
Transesterification of 7a to the methyl ester was followed
by differentiation of the two vicinal diols by selective
formation of the C(7), C(8) acetonide, providing 8. The
unprotected hydroxyls were inverted using a modification
of Auge´’s protocol.6c After treatment of the ditriflate with
n-Bu4NOBz, the major products were monobenzoate alco-
hols, with inversion having occurred at both C(4) and C(5).
easily be prepared from D-mannitol in one pot and 63%
yield.10 Double-reductive elimination using zinc, followed
by methanolysis of the crude material, gave 4 in 81% yield.
The key sequence in the synthesis is the conversion of
the C2-symmetric dienediol 4 to the highly functionalized
tetraol 7a in four steps (Scheme 2). Alkylation of the
stannylene acetal of 4 with tert-butyl bromoacetate proceeded
with concomitant cyclization, providing the desymmetrized
dioxanone 5.11 Addition of TFA after the alkylation step
ensured complete lactonization. Conversion of 5 to the silyl
ketene acetal using a modification of Angle’s method,12
followed by Ireland-Claisen rearrangement13 and hydrolysis
of the intermediate silyl ester,14 gave 6a in a one-pot
(13) (a) Burke, S. D.; Armistead, D. M.; Schoenen, F. J.; Fevig, J. M.
Tetrahedron 1986, 42, 2787. (b) Ireland, R. E.; Mueller, R. H. J. Am. Chem.
Soc. 1972, 94, 5897.
(14) Morton, D. R.; Thompson, J. L. J. Org. Chem. 1978, 43, 2102.
(15) Trace amounts of acid cause rearrangement to the lactone 12,
presumably via ring opening to the pentadienyl cation.
(7) For chemical syntheses of 2-deoxy-â-KDO starting from advanced
carbohydrate precursors, see: (a) Ohrui, H.; Morita, M.; Meguro, H.
Carbohydr. Res. 1992, 224, 319. (b) Boons, G. J. P. H.; van Delft, F. L.;
van der Klein, P. A. M.; van der Marel, G. A.; van Boom, J. H. Tetrahedron
1992, 48, 885. (c) Claesson, A. J. Org. Chem. 1987, 52, 4414.
(8) For a de noVo synthesis of 2-deoxy-â-KDO, see: Craig, D.;
Pennington, M. W.; Warner, P. Tetrahedron Lett. 1995, 36, 5815.
(9) Yadav, J. S.; Mysorekar, S. V.; Pawar, S. M.; Gurjar, M. K. J.
Carbohydr. Chem. 1990, 9, 307.
(16) Armstrong, A.; Brackenridge, I.; Jackson, R. F. W.; Kirk, J. M.
Tetrahedron Lett. 1988, 29, 2483.
(17) Kolb, H. C.; VanNieuwenhze, M. S.; Sharpless, K. B. Chem. ReV.
1994, 94, 2483.
(18) The structure was confirmed by X-ray crystallography.
(19) Removal of the bright yellow (DHQ)2-AQN ligand during workup,
by extraction with 3% H2SO4 saturated with K2SO4,17 was unsuccessful.
Copious amounts of yellow precipitate interfered with product isolation.
Although traces of the ligand did not interfere with subsequent steps, pure
7a could be obtained, if desired, after several chromatography/recrystalli-
zation cycles.
(10) Crombez-Robert, C.; Benazza, M.; Fre´chou, C.; Demailly, G.
Carbohydr. Res. 1997, 303, 359.
(11) David, S.; Thieffry, A.; Veyrie`res, A. J. Chem. Soc., Perkin Trans.
1 1981, 1796.
(12) Angle, S. R.; Breitenbucher, J. G.; Arnaiz, D. O. J. Org. Chem.
1992, 57, 5947.
(20) Becker, H.; Sharpless, K. B. Angew. Chem., Int. Ed. Engl. 1996,
35, 448.
(21) VanRheenen, V.; Kelly, R. C.; Cha, D. Y. Tetrahedron Lett. 1976,
1973.
72
Org. Lett., Vol. 1, No. 1, 1999

Scheme 2
Presumably, after one substitution the intermediate mono-
thermodynamic ratio of 10R:10â according to GC.24 Using
these conditions, we observed only a trace of epimerization
by TLC, with more forcing conditions leading only to
saponification. However, quenching the lithium enolate of
10R with ammonium chloride gave a 60% yield of 10â23
along with 18% of recovered 10R.
benzoate displaces the second triflate to form a bridged
acyloxonium species, as shown in eq 1. Hydrolysis of this
species would produce the mixture of monobenzoate alco-
hols.
The conversions of 10R/â to 1 and 2 are shown in Scheme
3. With 10â as the starting material, hydrolysis of the
acetonides with aqueous acetic acid, saponification of the
methyl ester with aqueous sodium hydroxide, and ion
+
exchange with Sephadex SP-25 (NH4 form) provided
Treatment of the mixture with benzoic anhydride (DMAP,
pyr) gave good yields of dibenzoate 9 after recrystallization.18
Immediate methanolysis of the mixture of monobenzoate
alcohols with NaOMe was examined, but gave complex
reaction mixtures. Methanolysis of the benzoates, treatment
2-deoxy-â-KDO 2 as its ammonium salt. The conversion of
10R to KDO requires oxidation at C-2. Oxidation of the
lithium enolate of 10R with MoO5‚Py‚HMPA (MoOPH)25
proved to be superior to the method previously employed
by Auge´.6c With the use of the same deprotection sequence
as that employed for 10â, hemiacetal 1123,26 was converted
to KDO23,27 and isolated as the ammonium salt. Conveniently,
22
with TMSCHN2 and acetonide formation gave 10R,23 an
intermediate in Auge´’s synthesis of KDO. Since adventitious
saponification occurred to a slight extent during the metha-
nolysis step, TMSCHN2 was used to re-esterify any car-
boxylic acid.
(24) Luthman, K.; Orbe, M.; Wåglund, T.; Claesson, A. J. Org. Chem.
1987, 52, 3777.
(25) (a) Vedejs, E.; Engler, D. A.; Telschow, J. E. J. Org. Chem. 1978,
43, 188. (b) Vedejs, E.; Larsen, S. Org. Synth. 1986, 64, 127.
(26) Hemiacetal 11 was obtained and used as an undetermined mixture
of anomers. However, the pure R-anomer could be obtained by recrystal-
lization as described in ref 4s.
(27) Physical and spectral data also agreed with those of an authentic
sample purchased from Sigma.
Claesson and co-workers report that 10R could be epimer-
ized with sodium methoxide, which gave within 2 min a 1:4
(22) Shioiri, T.; Aoyama, T.; Hashimoto, N. Chem. Pharm. Bull. 1981,
29, 1475.
(23) Spectral data agree with those previously reported.
Org. Lett., Vol. 1, No. 1, 1999
73

Scheme 3
Acknowledgment. Financial support of this research from
the ester 10â recovered from the KDO synthesis could be
diverted to the synthesis of 2-deoxy-â-KDO.
the National Institutes of Health (Grant CA 74394), Pfizer,
and Merck is gratefully acknowledged, as is the support of
the departmental NMR facilities by grants from the NIH
(ISIO RRO 8389-01) and the National Science Foundation
(CHE-9208463). We thank Professor Vedejs’ group for
supplying the MoOPH reagent.
In summary, we have demonstrated the rapid elaboration
of dienediol 4 to a highly functionalized tetrahydropyran
suitable for the syntheses of (+)-KDO and 2-deoxy-â-KDO.
In addition, the diene 6 could serve as a versatile template
for the construction of analogues. Further refinements and
extensions of this synthetic route are being investigated and
will be reported as events merit.
OL9905506
74
Org. Lett., Vol. 1, No. 1, 1999