Practical Synthesis of 1,6-Anhydro-2,4-dideoxy-β-D-glycerohexopyranos-3-ulose from Levoglucosan

Full text of DOI:10.1021/jo991719w

2588
J . Org. Chem. 2000, 65, 2588-2590
Sch em e 1
P r a ctica l Syn th esis of
1,6-An h yd r o-2,4-d id eoxy-â-D-glycer o-
h exop yr a n os-3-u lose fr om Levoglu cosa n
Kevin M. Belyk,* William R. Leonard, J r.,*
Dean R. Bender, and David L. Hughes
Department of Process Research, Merck Research
Laboratories, Merck & Co., Inc. P.O. Box 2000,
Rahway, New J ersey 07065
Received November 3, 1999
The potential utility of 1,6-anhydro-2,4-dideoxyhexo-
pyranoses (e.g., 3 and 4, Scheme 1) as chirons in the
synthesis of important biologically active compounds has
increased interest in developing methods for their prepa-
ration. In particular, a number of methods have been
described in accounts of the synthesis of precursors to
the lactonic portion of mevinic acids, which are potent
inhibitors of the enzyme HMG-CoA reductase.1 The 1,6-
anhydro-2,4-dideoxyhexopyranoses have also been em-
ployed in the synthesis of a wide variety of other
biologically important compounds including indanomy-
cin,2 and the spiroacetal portions of the avermectins and
milbemycins.3 Moreover, the 1,6-anhydro-2,4-dideoxy-
hexopyanose ring system is found in pheromones such
as the brevicomins4 and the multistriatins,5 as well as
the aroma constituents of beer.6
Sch em e 2
We required enantiomerically pure 4 as a key fragment
in the synthesis of the potent 5-lipoxygenase inhibitor 1
(Scheme 1).7
Despite the wide-spread reports of their use, few
efficient methods exist for the large-scale preparation of
1,6-anhydro-2,4-dideoxyhexopyranoses.8 Cˇ erny´ and co-
workers described the first synthesis of ketone 4.9
Preparation entailed tosylation of levoglucosan (2) at C-2
and C-4 to give 5a followed by oxidation at C-3 to give
ketone 6a (cf., Scheme 2). Reduction of the tosylate
groups with Raney nickel afforded 4 in low overall yield
from 2 (ca. 24%). Kelly and Roberts later reported an
improved preparation of 4 whereby the oxidation and
reduction steps of the original synthesis were reversed.10
However, the reduction of 5a with LiBHEt3 resulted in
a 5:1 mixture of 3 and its 3,4-dideoxy isomer that
required chromatographic separation after selective oxi-
dation of 3 to ketone 4. Subsequently, David and co-
workers demonstrated that complete regioselectivity for
the reduction of 5a could be obtained by nucleophilic
displacement of the tosylate groups with thiophenol
followed by reductive cleavage of the phenylsulfide groups
with a large excess of Raney Ni.11 In our hands, however,
we found that the Raney Ni reduction of the intermediate
bis(phenyl sulfide) was not reproducible and gave low
yields despite a wide variety of conditions and types of
Raney Ni. Only when an excess of Bu3SnH or ((CH3)3-
Si)3SiH was used in the phenyl sulfide reduction could
acceptable yields be obtained. Other groups have also
reported radical chain dideoxygenation of derivatives of
2 to prepare 3,12 which can readily be oxidized to give
(1) Boquel, P.; Cazalet, C.; Chapleur, Y.; Samreth, S.; Bellamy, F.
Tetrahedron Lett. 1992, 33, 1997 and references therein.
(2) Edwards, M.; Ley, S.; Lister, S.; Palmer, B.; Williams, D. J . Org.
Chem. 1984, 49, 3503.
(3) (a) Fox, C.; Hiner, R.; Warrier, U.; White, J . Tetrahedron Lett.
1988, 29, 2923 (b) Baker, R.; Boyes, J .; Broom, M.; O′Mahony, M.;
Swain, C. J . Chem. Soc., Perkin Trans. 1 1987, 1613.
(4) (a) Mori, K. Tetrahedron 1974, 30, 4223. (b) Silverstein, R.;
Brownlee, R.; Bellas, T.; Wood, D.; Browne, L. Science 1968, 159, 889.
(5) (a) Sum, P. E.; Weiler, L. Can. J . Chem. 1978, 56, 2700. (b)
Pearce, G.; Gore, W.; Silverstein, R.; Peacock, J .; Cuthbert, R.; Lanier,
G.; Simeone, J . J . Chem. Ecol. 1975, 1, 115.
(6) Tressl, R.; Friese, L.; Fendesack, F.; Koppler, H. J . Agric. Food
Chem. 1978, 26, 1422.
(7) (a) Le´ger, S.; Omeara, J .; Wang, Z. Synlett 1994, 829. (b) Girard,
Y.; Fortin, R.; Delorme, D.; Dube, D.; Hamel, P.; Lepine, C.; Ducharme,
Y. European Patent 579304, 1994.
(8) For reviews, see: (a) Witczak, Z. J ., Ed. Levoglucosenone and
Levoglucosans: Chemistry and Applications; ATL Press Inc. Science
Publishers: Mount Prospect, New York, 1994. (b) Cˇ erny´, M.; Stanek,
J ., J r. Adv. Carbohydr. Chem. Biochem. 1977, 34, 23.
(9) (a) Pecka, J .; Stanek, J . J r.; Cˇ erny´, M. Collect. Czech. Chem.
Commun. 1974, 39, 1192. (b) Cˇ erny´, M.; Paca´k, J .; Stanek, J . J r.
Carbohydr. Res. 1970, 15, 379. (c) Cˇ erny´, M.; Kalvoda, L.; Paca´k, J .
Collect. Czech. Chem. Commun. 1968, 33, 1143.
(11) David, C.; Gesson, J . P.; J acquesy, J . C. Tetrahedron Lett. 1989,
30, 6015.
(12) Barton, D. H. R.; J ang, D.; J aszberenyi, J . Tetrahedron Lett.
1992, 33, 6629. See also ref 1.
(10) Kelly, A. G.; Roberts, J . S. Carbohydr. Res. 1979, 77, 231.
10.1021/jo991719w CCC: $19.00 © 2000 American Chemical Society
Published on Web 03/16/2000

Notes
J . Org. Chem., Vol. 65, No. 8, 2000 2589
the desired ketone 4.13 In general, the use of undesirable
reagents and the need for chromatographic purification
of 3 made the use of radical reduction routes unattractive.
obtained in high yield (95%).18 Compared to the original
chromic anhydride oxidation of 5a ,9c this oxidation
method provides 6a in at least 40% higher yield without
need for further purification. The ketone was isolated as
a mixture of the free ketone and its hydrate;19 however,
this mixture was directly employed in the last step.
The final step required reductive cleavage of the
tosylate groups to give ketone 4. We found that reduction
of the carbonyl activated tosylate groups could be ac-
complished in the presence of zinc and a proton source.
Initial studies using acetic acid as the proton source
produced small amounts of desired ketone 4, along with
intermediate monotosylates 7a /7b, and the secondary
product enone 9.20 Conversion of 6a to 4 was improved
by use of aq NH4Cl as the proton source; however,
secondary formation of 9 was rapid.21 Substitution of solid
NH4OAc as the proton source reduced enone formation
to <1%, but incomplete reduction was observed. The use
of zinc, previously activated by an aq HCl wash, proved
to be the important factor in obtaining complete conver-
sion. Thus, reduction of 6a was accomplished with
activated zinc in THF using solid NH4OAc as the proton
source.
Upon completion, the reaction mixture was filtered and
the filtrate containing 4 was neutralized with solid K2-
CO3. Removal of the solids by filtration provided 4 as a
solution in THF. Concentration of this solution and
dilution with MTBE/hexanes resulted in crystallization
of the unreduced tritosylate 5d . Concentration of the
mother liquors afforded 4 in excellent yield (93%) and
purity (96 area % by GC).22
In summary, we have developed a practical synthesis
of 1,6-anhydro-2,4-dideoxy-â-D-glycero-hexopyranos-3-
ulose (4) from commercially available 1,6-anhydro-D-
glucose (2). This chromatography-free process has been
successfully demonstrated on a 23 g product scale in 75%
yield.
The three-step approach via the deoxygenation of the
derivatized ketone 6a proved to be most amenable to
large-scale preparation after several major improvements
were made. Most notable of these were the use of NaBrO3
with RuCl3 for the alcohol oxidation at C-3 and the
reductive cleavage of the R,R′-ditosylate ketone using Zn/
NH4OAc. To the best of our knowledge, this is the first
example of a one-pot R,R′-substituted dideoxygenation of
a ketone by a metal or metal salt other than Raney Ni.
Herein we report an efficient three-step process for the
preparation of ketone 4 in 75% yield from commercially
available14 levoglucosan (2). The process is amenable to
multi-kilogram scale.
Resu lts a n d Discu ssion
The synthesis of 1,6-anhydro-2,4-dideoxy-â-D-glycero-
hexopyranos-3-ulose (4) is illustrated in Scheme 1. The
original procedure15 for tosylation of levoglucosan (2) was
modified to produce at least a 50% higher yield of 5a and
to simplify the purification of ketone 4. Use of only a
slight excess of TsCl (220 mol %) in dry pyridine at a
low temperature (-5 °C) afforded the ditosylate 5a in
excellent yield (85% by HPLC wt % vs std) after extrac-
tive workup with EtOAc, aq citric acid, and water. Under
the optimized conditions, the tritosylate 5d was obtained
as the major impurity (ca. 14%) which was easily
removed by crystallization in the final step, while only
minor amounts of the monotosylates 5b and 5c were
present (ca. 0.5%). The desired selectivity for the C-2 and
C-4 positions is most likely due to steric hindrance at
C-3 from the acetal bridge. Purification of 5a was not
necessary before proceeding.
The next step in the synthesis called for oxidation of
5a to ketone 6a . Oxidation with RuCl3/NaOCl in CH3-
CN and AcOH16 gave 6a in only 15% yield. The low yield
under these conditions was attributed to incomplete
conversion. Complete oxidation of 5a was obtained only
in the presence of a large excess of NaOCl or NaIO4;
however, epimerization at C-2 and C-4 was observed
under these conditions. The best conditions were found
with the substitution of NaBrO3 as oxidant. This finding
has previously been observed for RuCl3-catalyzed oxida-
tions of alcohols, yet there are few examples in the
literature which utilize NaBrO3 despite its advantages
as an efficient, inexpensive oxidant.17 We found that
complete conversion of 5a to ketone 6a could be ac-
complished in a solvent system of CH3CN/AcOH/H2O
using NaBrO3 (65 mol %) in the presence of 1 mol %
RuCl3. After extractive workup, the ditosyl ketone 6a was
Exp er im en ta l Section
Gen er a l Meth od s. 1H and 13C NMR spectra were recorded
in CDCl3 at a frequency of 250.1 and 62.9 MHz, respectively.
Removal of residual solvents was monitored by 1H NMR. Gas
chromatography was conducted on an HP-5 column (30 m × 0.32
mm). HPLC analysis was carried out on a Zorbax Rx-C8 column
(4.6 × 25 mm) with detection at 220 nm using CH3CN and H2O
as eluents. Reactions can be monitored by silica gel TLC using
4:1 EtOAc/hexanes. Elemental analyses were performed by
Schwarzkopf Microanalytical Laboratory, Woodside, NY.
Zinc activation for the preparation of 4 was accomplished by
the following procedure:23 Zinc dust (400 g, 325 mesh) was
mechanically stirred for 1 h at 20 °C with 1.5 L 1.5% aq HCl.
The aqueous layer was decanted, and the solids were washed
with 2 × 1.5 L of THF. After filtration the solid was dried at
145 °C under vacuum overnight prior to use.
1,6-An h yd r o-2,4-d i-O-p-tolylsu fon yl-â-D-glu cop yr a n ose
(5a ). Over a 2 h period, TsCl (129 g, 678 mmol) was added to a
(13) Cai, D.; Hughes, D. L.; Verhoeven, T. R. Tetrahedron Lett. 1996,
37, 2357.
(18) Monotosylate byproducts 5b and 5c from the tosylation step
were also oxidized to give 6b and 6c. As expected, tritosylate 5d was
not affected under the reaction conditions.
(19) Typically 60% hydrate exists. The hydrate can be converted to
the ketone by azeotropic or molecular sieve drying.
(20) Cˇ erny´ et al. observed reductive cleavage of at least one of the
tosylate groups using zinc and acetic acid (see footnote 9c), but they
apparently did not develop this approach.
(14) Levoglucosan (1,6-anhydro-â-D-glucose) is commercially avail-
able from Aldrich, P.O. Box 2060, Milwaukee, WI 53201 or Schweiz-
erhall, Inc., 3001 Hadley Rd, South Plainfield, NJ 07080. For lab-scale
preparations of 1,6-anhydro-D-glucose, see Zottola, M.; Alonso, R.; Vite,
G.; Fraser-Reid, B. J . Org. Chem. 1989, 54, 6123 and references
therein.
(15) Cˇ erny´, M.; Gut, V.; Paca´k, J . Collect. Czech. Chem. Commun.
(21) See Experimental Section for the preparation of enone 9.
(22) Ketone 4 contained monodeoxy reduction adducts 8a and 8b
arising from 6b and 6c (ca. 1% by GC) and tritosylate 5d (ca. 2% by
GC).
1961, 26, 2542.
(16) These conditions were reported for the oxidation of 3. See ref
13.
(17) Kanemoto, S.; Tomioka, H.; Oshima, K.; Nozaki, H. Bull. Chem.
Soc. J pn. 1986, 59, 105 and references therein.
(23) In situ activation of zinc with 1,2-dibromoethane lead to
increased amounts of enone 9.

2590 J . Org. Chem., Vol. 65, No. 8, 2000
Notes
solution of levoglucosan (2) (50 g, 308 mmol) in anhydrous
pyridine (250 mL) at -10 °C in four portions, keeping the
temperature below -5 °C. The reaction was aged for 12.5 h at
-5 °C. Water (10 mL) was added to the white suspension, and
the reaction was aged 0.5 h. Ethyl acetate (1 L) was added, and
the mixture was washed with 23 wt % aq citric acid (4 × 375
mL) and water (375 mL). The organic layer was concentrated
and flushed with CH3CN to remove EtOAc. The ditosylate 5a
was obtained in 85% isolated yield (123 g) along with tritosylate
5d in 14% isolated yield (26 g) as a solution in CH3CN.
Separation of 5a and 5d was not required before proceeding.24
Ditosylate 5a : 1H NMR spectrum was in agreement with
published data.25 13C NMR δ 21.64 (2C), 65.93, 69.34, 74.61,
77.40, 78.68, 99.67, 127.81, 127.87, 130.02 (2C), 132.68, 132.89,
145.41, 145.46. Tritosylate 5d : 1H NMR δ 2.46 (3H, s), 2.47 (6H,
s), 3.69 (1H, dd, J ) 8.2, 5.7 Hz), 4.04 (1H, d, J ) 8.4 Hz), 4.15
(1H, br d, J ) 1.2 Hz), 4.51 (1H, br s), 4.57-4.62 (2H, om), 5.25
(1H, br s), 7.35 (6H, m), 7.68 (4H, m), 7.78 (2H, m). 13C NMR δ
21.63, 21.66, 21.69, 64.81, 71.93, 73.60, 73.73, 73.80, 98.51,
127.87, 127.92, 128.11, 129.98, 130.01, 130.05, 131.77, 132.36,
132.67, 145.44, 145.50, 145.81.
1,6-An h yd r o-2,4-d i-O-p-tolylsu fon yl-â-D-r ibo-h exop yr a -
n os-3-u lose (6a ). To a solution of 5a (100 g, 213 mmol) in CH3-
CN (300 mL) were added glacial acetic acid (62 mL) and RuCl3‚
3H2O (440 mg, 2.1 mmol) at 0 °C. To the resulting dark brown
mixture was added NaBrO3 (21 g, 138 mmol) as a solution in
water (98 mL) over 2.5 h, keeping the temperature below 8 °C.
The resulting black mixture was aged 1.5 h at 0 °C. 2-Propanol
(1.0 mL) was added at 0 °C and aged 15 min. The reaction was
poured into a mixture of isopropyl acetate (850 mL) and 15% aq
Na2S2O3 (100 mL). The resulting mixture was stirred for 15 min,
and the layers were separated. The organic layer was washed
with water (200 mL), sat. sodium bicarbonate (2 × 540 mL), and
water (200 mL). The organic layer was concentrated and flushed
with THF. Ditosyl ketone 6a was obtained in 95% yield (90 g)
as a solution in THF and required no further purification. 1H
NMR δ 2.46 (6H, s), 3.83 (1H, dd, J ) 8.5, 5.0 Hz), 3.90 (1H, dd,
J ) 8.5, 1.2 Hz), 4.48 (1H, d, J ) 1.0 Hz), 4.60 (1H, d, J ) 1.0
Hz), 4.94 (1H, ddd, J ) 5.0, ∼1.2, ∼1.2 Hz), 5.65 (1H, d, J ) 1.4
Hz), 7.34 (4H, m), 7.76 (4H, m). 13C NMR δ 21.65 (2C), 66.55,
76.43, 77.15, 79.08, 101.17, 128.01, 128.04, 129.93 (2C), 132.27
(2C), 145.61, 145.67, 190.49.
1,6-An h yd r o-2,4-d id eoxy-â-D-glycer o-h exop yr a n os-3-u l-
ose (4). Crushed NH4OAc (370 g, 4.80 mol) was added to a
suspension of activated zinc (314 g, 4.80 mol) in THF (1.3 L)
with mechanical stirring at 20 °C. The suspension was aged for
45 min and then cooled to 0 °C. The crude ditosyl ketone 6a (90
g, 0.19 mol) in THF (400 mL) was added over 2.5 h with stirring,
keeping the temperature below 5 °C. The reaction was warmed
to 20 °C and aged for 16 h. The suspension was filtered, and
the salts were washed with THF (1.0 L). Powdered (325 mesh)
potassium carbonate (93 g, 0.67 mol) was added to the combined
organic filtrates with stirring and aged at 20 °C for 22 h. The
salts were filtered and washed with THF (300 mL). The
combined filtrates were concentrated, and 4 was obtained as a
solution in THF. The unreduced tritosylate 5d was removed from
4 by precipitation from 1:2:7 MTBE:THF:hexanes. Filtration of
5d and concentration of the mother liquors afforded ketone 4 in
93% isolated yield (23 g, 96 area % pure by GC) as a yellow oil:
bp 90 °C/3 mmHg; [R]20 -103° (c 0.7, CHCl3) [lit.9a [R]D -103°
D
1
(c 0.8, CHCl3)]; H NMR δ 2.41 (1H, d, J ) 16.8 Hz), 2.52 (2H,
om), 2.74 (1H, ddd, J ) 16.8, 4.8, 1.4 Hz), 3.77 (1H, m), 3.82
(1H, m), 4.80 (1H, dd, J ) ∼4.8, ∼4.8 Hz), 5.74 (1H, br s); 13C
NMR δ 46.80, 48.51, 69.49, 72.11, 100.43, 204.35. Anal. Calcd
for C6H8O3: C, 56.24; H, 6.29. Found: C, 56.23; H, 6.43.
En on e (9). Ketone (4) is added to TFA at room temperature
and aged 5 h. The enone is then isolated as an unstable light
brown oil by prep TLC (4:1) EtOAc:hexanes.1H NMR δ 2.38 (1H,
br s), 2.40 (1H, ddd, J ) 16.8, 3.6, 1.3 Hz), 2.78 (1H, dd, J )
16.8, 14.3 Hz), 3.80 (1H, dd, J ) 12.3, 5.4 Hz), 3.93 (1H, dd, J )
12.3, 3.1 Hz), 4.54 (1H, dddd, J ) 14.3, 5.4, 3.3, 3.3 Hz), 5.45
(1H, dd, J ) 6.0, 1.3 Hz), 7.40 (1H, d, J ) 6.0 Hz).
(24) If desired, the ditosylate 5a can be separated from the tri-
tosylate 5d by chromatography on RP-C18 adsorbent (Delta-Pak C18,
15 µm 100 Å). The ditosylate 5a is isolated by elution with 1: 1 CH3-
CN/water.
(25) Van Rijsbergen, R.; Anteunis, M. J . O.; De Bruyn, A. Bull. Soc.
Chim. Belg. 1982, 91, 297.
J O991719W