Enantioselective synthesis of 2-deoxy- and 2,3-dideoxyhexoses

Article abstract of DOI:10.1021/ol025844x

The enantioselective syntheses of C-6 O-TBS- and N-Cbz-protected 2-deoxy- and 2,3-dideoxysugars have been achieved in 6-8 steps from furfural. A combination of chemo-, regio-, and diastereoselective oxidation and reduction reactions produced deoxysugars w

Full text of DOI:10.1021/ol025844x

ORGANIC
LETTERS
2
002
Vol. 4, No. 10
771-1774
Enantioselective Synthesis of 2-Deoxy-
and 2,3-Dideoxyhexoses
1
Michael H. Haukaas and George A. O’Doherty*
Department of Chemistry, UniVersity of Minnesota, Minneapolis, Minnesota 55455
odoherty@chem.umn.edu
Received March 8, 2002
ABSTRACT
The enantioselective syntheses of C-6 O-TBS- and N-Cbz-protected 2-deoxy- and 2,3-dideoxysugars have been achieved in 6−8 steps from
furfural. A combination of chemo-, regio-, and diastereoselective oxidation and reduction reactions produced deoxysugars with various C-6
substitution. A key development of this route was the use of o-nitrobenzenesulfonylhydrazide (NBSH) as a diimide precursor. These overall
procedures allow for the synthesis of eight deoxysugars in either enantiomeric form.
Since their identification in the 1950s the aminoglycosides
have been an important class of antibiotics in the fight against
infections. Aminoglycosides consist of a large class of
Arbekacin and dibekacin are deoxysugar analogues of
kanamycin that were developed to overcome the problems
of enzymatic resistance. Removing the hydroxyl groups at
1
3
mono- and bisglycosidated diaminocyclitols such as kana-
mycins A-C and others (Figure 1). The increased incidence
C-3 and C-4 of kanamycin produces an antibiotic that still
binds to the ribosome but is no longer inactivated by the
enzymes that resistant bacteria have evolved to counter them.
Unfortunately, even these new antibiotics are succumbing
2
3
6
b
to bacterial resistance. Our approach to finding new
(
1) Goodman-Gilman, A.; Goodman, L. S.; Rall, T. W.; Murad, F.
Goodman and Gilman’s The Pharmacological Basis of Therapeutics;
MacMillan Publishing Co.: New York, 1985.
(2) For the first synthesis of Kanamycin A, see: Nakajima, M.;
Hasegawa, A.; Kurihara, N.; Shibata, H.; Ueno, T.; Nishumura, D.
Tetrahedron Lett. 1968, 9, 623.
(3) Hooper, I. R. The Naturally Occurring Aminoglycoside Antibiotics.
In Aminoglycoside Antibiotics; Hooper, I. R., Umezawa, H., Eds.; Springer-
Verlag: NewYork, 1982.
(
4) Brickner, S. J. Chem. Ind. 1997, 131.
(5) Miller, G. H.; Sabatelli, F. J.; Hare, R. S.; Glupczynski, Y.; Mackey,
P.; Shlaes, D.; Shimizu, K.; Shaw, K. J.; Bauernfeind, A.; Schweighart, S.;
Shannon, K.; Patzer, J.; Molinari, G.; Schito, G. C.; Gomezlus, R.;
Gomezlus, S.; Ferreira, H.; Sousa, J. C.; Vaz, M. J. M.; Collatz, E.; Bismuth,
R.; Lambert, T.; Courvalin, P.; Minozzi, C.; Klugman, K. Clin. Infect. Dis.
Figure 1.
1
997, 24, S46.
(6) For instance, worldwide Streptococcus pneumoniae is responsible
for 1.2 million deaths per year in children under the age of 5; see: (a)
Bruyn G. A. W.; Zegers, B. J. M.; van Furth, R. Clin. Infect. Dis. 1992,
of bacterial resistance has created a great worldwide need
for new antibacterial compounds to fight multiple-antibiotic-
resistant pathogens.4 The synthesis and evaluation of new
aminoglycoside analogues will greatly increase our under-
standing of how to control bacterial infections and bacterial
antibiotic resistance.7
14, 251. (b) Janoff, E. N.; Rubins, J. B. Microb. Drug Resist. 1997, 3, 215.
Even more alarming is that Mycobacterium tuberculosis is estimated to
cause 3 million deaths per year; see: (c) Swartz, M. N. Proc. Natl. Acad.
Sci. U.S.A. 1994, 91, 2420.
-6
(7) Umezawa, H.; Kondo, S. Mechanisms of Resistance to Aminogly-
coside Antibiotics. In Aminoglycoside Antibiotics; Hooper, I. R., Umezawa,
H., Eds.; Springer-Verlag: NewYork, 1982.
1
0.1021/ol025844x CCC: $22.00 © 2002 American Chemical Society
Published on Web 04/23/2002

antibiotics is to mimic nature by synthesizing novel deoxyami-
nosugars then to couple them to an aminocyclitol core. These
new aminoglycoside analogues may be the next generation
of nonresistant antibiotics.
Scheme 2
Structure activity relationship analysis of the carbohydrate
portions of the aminoglycosides in terms of antibiotic
resistance reveals two common structural motifs that impart
prolonged bioactivity: the presence of both deoxyamino-
sugars and deoxysugars (Figure 1). The replacement of a
hydroxyl group with an amino group installs an improved
hydrogen-bond acceptor. In contrast, the removal of a
hydroxyl group imparts in vitro stability by lessening the
abilities of naturally occurring glycosidase enzymes to
degrade the structure.3 This approach can be used by
synthetic chemists to improve binding and to overcome the
accompanying resistance that all antibiotics eventually incur.
For the past 5 years our group has been developing
methods for the enantioselective synthesis of carbohydrates
with the goal of accessing uniquely functionalized aminogly-
coside antibiotics (Scheme 1). Previous structure activity
Recently, our group has successfully employed asymmetric
catalysis for the synthesis of D- and L-sugars and iminosugars
from the achiral vinylfuran via chiral furans 4 and 6 (Scheme
9
-11
2).
Four hexoses of the type 5 were prepared from
monoprotected diol 4 in 4-6 steps, as well as the analogous
C-6 aminosugars 7 from furan 6. Similarly this route should
allow for access to unnatural C-2/C-3 dideoxysugars. We
envisioned that C-2 and C-2/C-3 deoxysugars could be
derived from the same intermediates 9 and 10 (Scheme 3)
Scheme 1
Scheme 3
that were used for the syntheses of 5 and 7. Reported herein
are our successful efforts at the conversion of intermediates
studies of the aminoglycoside antibiotics have used semi-
synthesis techniques to strategically remove hydroxyl groups
from the existing aminoglycosides, followed by the addition
9
and 10 into both C-2 deoxysugars and C-2/C-3 dideoxy-
sugars. This approach allows for access to either enantiomer
of N-Cbz-protected and O-TBS-protected C-2 deoxy-allo-
and deoxy-galactosugars and C-2/C-3 dideoxy-gluco- and
galacto-sugars.
8
of functional groups of interest. A complementary approach
is to start with a drastically simplified structure and to
sequentially increase its stereochemical and functional
complexity (e.g., 2 to 1). To implement this strategy, a
flexible route to D- and L-sugars that allows for the synthesis
of various stereoisomers and deoxyanalogues was developed.
Our earlier work has shown that the C-6 O-TBS-protected
and C-6 N-Cbz-protected pyran allylic alcohols 9 and 10
were diastereoselectively produced from furfural in 47% and
9
Having successfully addressed the synthesis of sugars,
1
1
1
4% yields, respectively (Scheme 3). The asymmetry of 9
10
11
iminosugars, and aminosugars (Scheme 2), our next goal
was to apply this methodology to the catalytic asymmetric
synthesis of deoxysugars.
was effectively introduced by a Sharpless dihydroxylation
1
2
of vinylfuran (>90% ee). In contrast, the asymmetry of
0 was ineffectively introduced in terms of enantio- and
1
regioselectivity by the Sharpless aminohydroxylation; how-
(8) (a) Umezawa, S.; Tsuchiya, T. Total Synthesis and Chemical
Modification of the Aminoglycoside Antibiotics. In Aminoglycoside Anti-
biotics; Hooper, I. R., Umezawa, H., Eds.; Springer-Verlag: NewYork,
ever, very high enantioexcess (>98% ee) was achieved with
the Noyori hydrogenation of a furyl ketone.
13,14
1982. (b) Okuda, T.; Ito, Y. Biosynthesis and Mutasynthesis of Aminogly-
coside Antibiotics. In Aminoglycoside Antibiotics; Hooper, I. R., Umezawa,
H., Eds.; Springer-Verlag: NewYork, 1982.
(12) The enantioselectivity was increased to >98% after recrystallization
of the furan diol bisbenzoate.
(13) For specific conditions, see ref 11 and (a) Fujii, A.; Hashiguchi, S.;
Uematsu, N.; Ikariya, T.; Noyori, R. J. Am. Chem. Soc. 1996, 118, 2521.
For catalyst preparation, see: (b) Haack, K.-J.; Hashiguchi, S.; Fujii, A.;
Ikariya, T.; Noyori, R. Angew. Chem., Int. Ed. Engl. 1997, 36, 285.
(
9) (a) Harris, J. M.; Keranen, M. D.; O’Doherty, G. A. J. Org. Chem.
999, 64, 2982. (b) Harris, J. M.; Keranen, M. D.; Nguyen, H.; Young, V.
G.; O’Doherty, G. A. Carbohydr. Res. 2000, 328, 17.
1
(
(
10) Haukaas, M. H.; O’Doherty, G. A. Org. Lett. 2001, 3, 401.
11) Haukaas, M. H.; O’Doherty, G. A. Org. Lett. 2001, 3, 3899.
1772
Org. Lett., Vol. 4, No. 10, 2002

With the establishment of enantioselective routes to both
allyl alcohols 9 and 10, the use of transition-metal-catalyzed
hydrogenation reactions was investigated for the conversion
of 9 into the dideoxysugar 11 (Scheme 4). Our preference
Scheme 5
Scheme 4
decomposition of o-nitrobenzenesulfonylhydrazide (NBSH)
2
0
in the presence of triethylamine. Our choice of the
o-nitrobenzenesulfinate anion as a leaving group was made
easy by the fact that Myers and Zheng had reported a simple
preparation of the labile NBSH. In fact, exposing 9 to an
excess (2-5 equiv) of a 1-2 molar ratio of NBSH and
for the direct hydrogenation reaction of 9 and 10 was
ameliorated by concerns with competing hydrogenolysis (via
21
1
9
3) of the allylic ester or ether bonds.15 In fact, when pyran
was subjected to typical palladium on carbon hydrogenation
triethylamine in CH
2
Cl
2
gave a good yield (91%) of pivolate
conditions (1 atm of H
2
, 10% Pd/C in MeOH) a significant
20
1
1. Not surprisingly, the C-6 N-Cbz-protected pyran 10
amount of the hydrogenolysis product 12 was found (87%
of a 1:1 ratio of 11 and 12).16 Switching the catalyst from
Pd/C to Rh/alumina resulted in less but still a significant
proportion of hydrogenolysis. Thus when exposing 9 to 1
reacted under both diimide conditions to give very good
yields (86% from potassium azodicarboxylate/AcOH and
8
3
1% from NBSH/Et N) of 14. Both sets of reaction condi-
tions gave clean products without the use of silica gel
chromatography.
atm of H
2 2 3
in the presence of 10% Rh on Al O in EtOAc, a
90% yield of a 3.5:1 ratio of 11 and 12 was isolated.
Although pure fractions of 11 could be isolated by silica gel
column chromatography, a more pragmatic route to 11 was
desired. In addition, the C-6 N-Cbz group of 10 would likely
not survive the hydrogenation conditions.
Scheme 6
A practical solution was realized by the use of a diimide
to reduce the dihydropyan ring. Hence, exposing 9 to an
excess of potassium azodicarboxylate (10-15 equiv) and
acetic acid (20-30 equiv) in MeOH gave a good yield (93%)
of pivolate 11.1 Because of our concerns with the potential
acid sensitivity of 11 in the presence of acetic acid, we
searched for an alternative procedure for the generation of
diimide. Initially, benzenesulfonylhydrazide was examined
as a diimide source, but the high temperature required for
the reaction was undesirable.19
7,18
With the hope that a better leaving group would allow for
a more facile generation of diimide, we turned to the
The C-4 epimeric allylic alcohols 15 and 17 were easily
prepared from the corresponding isomers (9 and 10) via a
Mitsunobu/hydrolysis sequence (p-nitrobenzoic acid/DEAD/
(14) The absolute stereochemistries and enantioexcesses of 4 and 6 were
3 3
PPh , then Et N/MeOH). Utilizing the same diimide reduc-
determined by Mosher’s method; see: (a) Sullivan, G. R.; Dale, J. A.;
Mosher, H. S. J. Org. Chem. 1973, 38, 2143. (b) Yamaguchi, S.; Yasuhara,
F.; Kabuto, K. T. Tetrahedron 1976, 32, 1363.
tion conditions on allylic alcohols 15 and 17 provided
excellent yields of the tetrahydropyrans 16 and 18. The axial
alcohols 16 and 18 were routinely isolated in high yields
(15) These model studies were carried out on the anomeric pivolate
because of the propensity of allylic esters to undergo hydrogenolysis
reactions. Thus a procedure established on this substrate should have the
greatest substrate generality.
from 15 (85% from NBSH/Et
Et N).
3
N) and 17 (86% from NBSH/
3
(16) Product 12 could be derived from an in situ acid-catalyzed
elimination reaction of the anomeric pivolate of 11 followed by hydrogena-
tion of the resulting enolether. This possibility was discounted for the Rh/
Al2O3 reaction, when it was shown that 11 was untouched when re-exposed
to the hydrogenolysis conditions.
(19) For a report on the high-temperature use of other benzenesulfonyl-
hydrazides as diimide precursors see: Cusack, N. J.; Reese, C. B.; Risius,
A. C.; Roozpeikar, B. Tetrahedron 1976, 32, 2157.
(20) To the best of our knowledge this is the first use of NBSH as a
diimide precursor for an alkene reduction, although Hunig had previously
shown that NBSH decomposes to diimide; see: Hunig, S.; Muller, H. R.;
Their, W. Angew. Chem., Intl. Ed. Engl. 1965, 4, 271.
(17) Prepared by the method of Heck: Dieck, H. A.; Heck, R. F. J. Org.
Chem. 1975, 40, 1083. Also see: Pasto, D. J. Organic Reactions; Paquette,
L. A., Ed.; John Wiley & Sons: 1991; Vol. 40, p 91.
(
18) If any starting material was detected by 1H NMR, it was easily
converted to product by re-subjecting the crude product to the reaction
conditions.
(21) For the preparation of NBSH: Myers, A. G.; Zheng, B.; Movassaghi,
M J. Org. Chem. 1997, 62, 7507.
Org. Lett., Vol. 4, No. 10, 2002
1773

Scheme 7
Scheme 8
To gain access to the C-2 deoxysugars we hoped to
diastereoselectively oxidize alkenes 19 and 20 (Scheme 7).
Thus we investigated the use of Myers’ reductive 1,3-
transposition reaction22 to reduce the allylic alcohols 9 and
of 19 and 20. For instance, the use of t-BuOH as solvent
increased the facial selectivity for 24 over 25 from a 1.5:1
ratio to 2:1 with a slightly higher combined yield (83%).
While changing solvents was found to have very little effect
on the stereoselectivities, adding a more hindered ligand had
10 to alkenes 19 and 20. This rearrangement reaction seemed
ideal for an equatorial allylic alcohol because the reductive
rearrangement requires an axial allylic diazene as depicted
23
in 21. Thus exposing 9 to 2 equiv of a 1:1:1 ratio of NBSH,
PPh , and DEAD at -30 °C in N-methylmorpholine gave a
3% yield of alkene 19. The allylic alcohol 10 reacted under
similar conditions (2.5 equiv) to give a 78% yield of alkene
0.
With access to the two desired alkenes 19 and 20, we next
a significant affect. Thus, treating pyran 19 to OsO
in the presence of (DHQD) PHAL gave an improved ratio
3:1) with the galacto-diol 22 being the major isomer at
4
/NMO(aq)
3
2
7
(
>
90% conversion. The relative configurations of 22-25
2
were determined by examining relevant coupling constants
from the H NMR. Particularly diagnostic for the 2-deoxy-
1
explored their reactivity under typical dihydroxylation condi-
tions (Scheme 8). From our previous studies, the steric effects
of the axial anomeric pivaloate was expected to control the
facial selectivity, resulting in the C-2 deoxy galactosugars
galactose stereochemistry in 22 and 24 were the large J
values (∼12.5 Hz) for the diaxial protons at C-2 and C-3.
Conversely, both 23 and 25 displayed large J values (>8
Hz) for the diaxial protons at C-4 and C-5, which is indicative
of the 2-deoxyallose stereochemistry.
In summary, we have developed two practical reaction
sequences for the preparation of four C-2 deoxysugars and
four C-2/C-3 dideoxysugars from pyrans 9 and 10. These
routes are amenable to multigram scale preparation. We hope
these new routes to D- and L-sugars will be beneficial for
the synthesis of unnatural aminoglycosides.
2
“
Cl
2 and 24. Surprisingly, treatment of pyran 19 to modified
Upjohn Conditions” (5 mol % OsO , 50% NMO(aq)/CH
; Scheme 8 condition a), afforded a 1.5:1 ratio of C-2-
4
2
-
2
deoxysugars 22/23. The galacto-diol 22 was the predominant
isomer, isolated in a 52% yield along with 35% of the
diastereomeric allo-diol 23. Nearly identical stereoselectivi-
2 2
ties were observed for the C-6 N-Cbz series 20 in CH Cl .
Thus exposing 20 to the same conditions provided a 1.1:1
ratio of C-2-deoxysugars 24/25 with the galacto-diol 24 as
the isomer (47% yield) and the diastereomeric allo-diol 25
Acknowledgment. We thank the Arnold and Mabel
Beckman Foundation and the National Institute of General
Medical Sciences (1R01 GM63150-01A1) for their generous
support of our program.
(44%) as the minor.
We have previously found a significant solvent dependence
of the facial selectivity of the dihydroxylation for substrates
with C-6 N-Cbz substitution.11 However, only small varia-
Supporting Information Available: Complete experi-
1
13
mental procedures and spectral data (MP, IR, H NMR,
C
tions in the selectivity were observed for the dihydroxylations
NMR, HRMS, and/or EA) for all new compounds. This
material is available free of charge via the Internet at
http://pubs.acs.org.
(
22) Myers, A. G.; Zheng, B. J. Am. Chem. Soc. 1996, 118, 4492. Myers,
A. G.; Zheng, B. Tetrahedron Lett. 1996, 37, 4841.
23) We have already shown that the allylic alcohols 9 and 10 are ideal
substrates for the Mitsunobu reaction; see refs 9 and 11.
(
OL025844X
1774
Org. Lett., Vol. 4, No. 10, 2002