Synthesis of 4,6-dideoxyfuranoses through the regioselective and diastereoselective oxyfunctionalization of a dimethylphenylsilyl-substituted chiral homoallylic alcohol

Article abstract of DOI:10.1021/jo010549w

The 4,6-dideoxyfuranoses 10a and 10b have been synthesized by starting from the readily available E-5-dimethylphenylsilyl-2-hexene-4-ol (1)and employing successively three versatile oxyfunctionalization methods, namely photooxygenation, metal-catalyzed epoxidation, and oxidative desilylation. Photooxygenation of the hydroxy vinylsilane 1 and subsequent triphenylphosphine reduction of the hydroperoxides 3 afford the like-4a and unlike-4b diols, which have been converted separately to the tetrahydrofurans (2S*,3R*,5R*)-7a and (2S*,3R*,5S*)-7b by a combination of diastereoselective epoxidation and regioselective intramolecular epoxide-ring opening. In the epoxidation reaction, catalyzed by Ti(OiPr)₄ or VO(acac)₂, only one diastereomer (dr > 95:5) of the epoxide 5 is obtained. Further intramolecular opening of the epoxide ring in erythro-5 occurs regioselectively at the C-α position and diastereoselectively under inversion of the configuration of the silyl-substituted stereogenic center to generate only one diastereomer of the tetrasubstituted tetrahydrofurans 7. Oxidative desilylation of the latter gave the hitherto unknown 4,6-dideoxyfuranoses 10a and 10b. The use of the optically active E-5-dimethylphenylsilyl-2-hexene-4-ol (1) as starting material, which is readily available through lipase-catalyzed kinetic resolution, leads to the D- and L-4,6-dideoxysorbofuranoses 10a and D- and L-4,6-dideoxyfructofuranoses 10b in up to 98% enantiomeric excess.

Full text of DOI:10.1021/jo010549w

J . Org. Chem. 2001, 66, 7365-7371
7365
Syn th esis of 4,6-Did eoxyfu r a n oses th r ou gh th e Regioselective a n d
Dia ster eoselective Oxyfu n ction a liza tion of a
Dim eth ylp h en ylsilyl-Su bstitu ted Ch ir a l Hom oa llylic Alcoh ol
Waldemar Adam,* Chantu R. Saha-Mo¨ller, and Katharina S. Schmid
Institut fu¨r Organische Chemie, Universita¨t Wu¨rzburg, Am Hubland, D-97074 Wu¨rzburg, Germany
adam@chemie.uni-wuerzburg.de
Received May 30, 2001
The 4,6-dideoxyfuranoses 10a and 10b have been synthesized by starting from the readily available
E-5-dimethylphenylsilyl-2-hexene-4-ol (1) and employing successively three versatile oxyfunction-
alization methods, namely photooxygenation, metal-catalyzed epoxidation, and oxidative desily-
lation. Photooxygenation of the hydroxy vinylsilane 1 and subsequent triphenylphosphine reduction
of the hydroperoxides 3 afford the like-4a and unlike-4b diols, which have been converted separately
to the tetrahydrofurans (2S*,3R*,5R*)-7a and (2S*,3R*,5S*)-7b by a combination of diastereose-
lective epoxidation and regioselective intramolecular epoxide-ring opening. In the epoxidation
reaction, catalyzed by Ti(OiPr)4 or VO(acac)2, only one diastereomer (dr >95:5) of the epoxide 5 is
obtained. Further intramolecular opening of the epoxide ring in erythro-5 occurs regioselectively
at the C-R position and diastereoselectively under inversion of the configuration of the silyl-
substituted stereogenic center to generate only one diastereomer of the tetrasubstituted tetrahy-
drofurans 7. Oxidative desilylation of the latter gave the hitherto unknown 4,6-dideoxyfuranoses
10a and 10b. The use of the optically active E-5-dimethylphenylsilyl-2-hexene-4-ol (1) as starting
material, which is readily available through lipase-catalyzed kinetic resolution, leads to the D- and
L-4,6-dideoxysorbofuranoses 10a and D- and L-4,6-dideoxyfructofuranoses 10b in up to 98%
enantiomeric excess.
In tr od u ction
In past years, we have intensively studied the selective
oxyfunctionalization of chiral allylic alcohols to the
corresponding epoxy diols by photooxygenation7,8 and
direct titanium-catalyzed epoxidation.9 Alternatively, the
intermediary allylic hydroperoxides of the photooxygen-
ation were reduced by triphenylphosphine and subse-
quently the resulting alkene diols were epoxidized under
Ti(OiPr)4 catalysis with â-hydroxy hydroperoxides as
oxygen donors.10 Here we report that by employing this
methodology, the hitherto unknown 4,6-dideoxyfuranoses
10 may be prepared stereoselectively from the dimeth-
ylphenylsilyl-substituted chiral homoallylic alcohol 1
(Scheme 1). Most gratifying, also the optically active 4,6-
dideoxysorbofuranose 10a and 4,6-dideoxyfructofuranose
10b derivatives have been now made available for the
first time by starting from the enantiomerically enriched
hydroxy vinylsilane 1.
While monodeoxy sugars are present in a wide variety
of natural products1-3 relatively little is known about
dideoxy sugars, in particular, the 4,6-dideoxyhexoses.
Such dideoxy sugars, e.g., the 4,6-dideoxyhexose chalcose
(see structures), are found as structural units in bioactive
compounds.4 The biosynthetic pathway of 4,6-dideoxy-
pyranoses has been studied,4 and even a few derivatives
have been synthesized by chemical methods.5,6 In con-
trast, 4,6-dideoxyfuranoses (see structures) are unknown
to date in nature, nor have derivatives been synthesized.
Resu lts
The silyl-substituted homoallylic alcohol 1 was syn-
thesized by hydrosilylation, analogous to the method of
Oshima et al. (Scheme 2).11 The active silylating reagent,
* Fax: 049(0)931/8884756. Internet: http://www-organik.chemie.uni-
wuerzburg.de.
(1) Williams, N.; Wander, J . In The Carbohydrates, Chemistry and
Biochemistry; Pigman, W., Horton, D., Eds.; Academic Press: New
York, 1980; Vol. 1B, pp 761-799.
(7) Prein, M.; Adam, W. Angew. Chem., Int. Ed. Engl. 1996, 35, 477-
494.
(8) Stratakis, M.; Orfanopoulos, M. Tetrahedron 2000, 56, 1595-
1615.
(9) (a) Martin, V. S.; Woodard, S. S.; Katsuki, T.; Yamada, Y.; Ikeda,
M.; Sharpless, K. B. J . Am. Chem. Soc. 1981, 103, 6237-6240. (b)
Adam, W.; Wirth, T. Acc. Chem. Res. 1999, 32, 703-710. (c) Adam,
W.; Richter, M. J . Acc. Chem. Res. 1994, 27, 57-62.
(10) Adam, W.; Peters, K.; Renz, M. J . Org. Chem. 1997, 62, 3183-
3189.
(2) Ichikawa, Y.; Lin, Y. C.; Dumas, D. P.; Shen, G. J .; Garcia-
J unceda, E.; Williams, M. A.; Bayer, R.; Ketcham, C.; Walker, L. E.;
Paulson, J . C.; Wong, C. H. J . Am. Chem. Soc. 1992, 114, 9283-9298.
(3) Borman, S. Chem., Eng. News 1992, Dec 7, 25.
(4) Thorson, J . S.; Lo, S. F.; Liu, H.-w.; Hutchinson, C. R. J . Am.
Chem. Soc. 1993, 115, 6993-6994.
(5) Danishefsky, S.; Kerwin, J . F., J r. J . Org. Chem. 1982, 47, 1597-
1598.
(6) Yamamoto, Y.; Komatsu, T.; Maruyama, K. J . Orgmet. Chem
1985, 285, 31-42.
(11) Wakamatsu, K.; Nonaka, T.; Okuda, Y.; Tu¨ckmantel, W.;
Oshima, K.; Utimoto, K.; Nozaki, H. Tetrahedron 1986, 42, 4427-4436.
10.1021/jo010549w CCC: $20.00 © 2001 American Chemical Society
Published on Web 09/29/2001

7366 J . Org. Chem., Vol. 66, No. 22, 2001
Adam et al.
Sch em e 1. Syn th etic Rou te to th e
4,6-Did eoxyfu r a n oses 10
gen source instead of tert-butyl hydroperoxide was neces-
sary to sustain the titanium catalytic cycle10) under
titanium catalysis gave the tetrahydrofurans 7 as major
products, but no epoxides 5 nor tetrahydropyrans 8 were
obtained (Table 1, entries 1 and 6). Evidently, the
erythro-5 epoxide was immediately converted to the
tetrahydrofuran 7 during the reaction. The direct epoxi-
dation of the hydroperoxides 2, catalyzed by Ti(OiPr)4,
is not desirable because much (ca. 30%) reduction to the
diols 4 is observed.
The vanadium-catalyzed oxidation14 of the like-diol 4a
with tert-butyl hydroperoxide gave essentially exclusively
the tetrahydrofuran 7a in 78% yield of isolated material
(entry 2), and only traces of the enone 6. In contrast,
under the same reaction conditions, the oxidation of the
unlike-diol 4b led to the (RR*,1R*,3S*)-5b epoxide as
major product (entry 7), which was isolated in 42% yield
when neutral conditions were carefully maintained. Also
a significant (30%) amount of tetrahydrofuran 7b was
formed. The latter was obtained as the exclusive product,
when the vanadium-catalyzed reaction was conducted at
45 °C for 10 h (entry 8). Thereby, the tetrahydrofuran
7b was isolated in 82% yield after silica gel chromatog-
raphy.
The oxidation of the diols 4a and 4b with methyltri-
oxorhenium (MTO) and UHP (urea/hydrogen peroxide
adduct) led to the tetrahydrofuran 7 and tetrahydropyran
8 products in a ratio of 65:35, together with small (4-
6%) amounts of enone 6 (entries 3 and 9). NMR monitor-
ing of the oxidation showed that initially a mixture of
the diastereomeric epoxides was observed (65:35 in favor
of the erythro-5 epoxide), which was slowly converted to
a mixture of tetrahydrofuran 7 and tetrahydropyran 8.
The oxidation of the like-diol 4a with mCPBA15 (entry
4) gave the tetrahydrofuran 7a as major product (69%)
and minor amounts (23%) of tetrahydropyran 8a , as well
as a small quantity of a mixture of the diastereomeric
epoxides 5a (8%). NMR monitoring of the reaction
progress showed only a moderate diastereoselectivity (ca.
80:20) for the erythro-5a and threo-5a epoxides. In
contrast, the oxidation of the unlike-diol 4b by mCPBA
gave a mixture of the epoxides erythro-5b and threo-5b
as major products (71%) in a erythro/ threo ratio of 96:4
(entry 10). At the beginning of the reaction, a diastere-
omeric ratio of 81:19 was determined by NMR spectros-
copy for the epoxides 5b. Also, significant amounts of the
tetrahydrofuran 7b (16%) and tetrahydropyran 8b (13%)
were isolated, as well as traces of the enone 6.
(PhMe2Si)3ZnLi, attacks the triple bond from the less
crowded side cis-selectively to give the E-configured
5-dimethylphenylsilyl-4-hexen-2-ol (1) in high regiose-
lectivity and in a 66% yield of isolated material. Only
8% of the regioisomeric E-4-dimethylphenylsilyl-4-hexen-
2-ol was formed.
Photooxygenation of the homoallylic alcohol 1 with
tetraphenylporphyrin (TPP) as sensitizer (Scheme 2) gave
the diastereomeric hydroperoxy alcohols 2 in high yield
(90%). The silyl group at the double bond in the homoal-
lylic alcohol 1 directs the H abstraction regioselectively
to the gem allylic position in the singlet-oxygen ene
reaction (gem effect).7,12 Only traces (<5%) of the regio-
isomeric hydroperoxide 2′ were produced, which im-
mediately decomposed to the enone 3 and dimethylphen-
ylsilanol. The diastereomeric hydroperoxy alcohols 2a
(like) and 2b (unlike) were obtained as a 50:50 mixture,
the lack of diastereoselectivity in this photooxygenation
is due to the absence of 1,3-allylic strain.9b,13
The diastereomers 2a and 2b could not be separated
by silica gel chromatography; nevertheless, after reduc-
tion by triphenylphosphine to the diastereomeric diols
4a (like) and 4b (unlike), these were isolated individually
by silica gel chromatography in a total yield of 91%
(Scheme 2). For convenience, the photooxygenation of the
homoallylic alcohol 1 and reduction of the hydroperoxide
product 2 may be conducted in an one-pot process to
afford the diol 4 in 90% overall yield; one tedious
purification step is thereby obviated.
Various oxidants were tested for the epoxidation of the
diols 4a and 4b. Besides the expected epoxides 5, some
enone 6 and also the tetrahydrofuran 7 and tetrahydro-
pyran 8 products (intramolecular nucleophilic attack on
the epoxide ring) were obtained (Scheme 3). The product
composition depended on the oxidant and the reaction
conditions; the results are summarized in Table 1.
The oxidation of the like-diol 4a and the unlike-diol
4b with 2,3-dimethyl-3-hydroperoxy-2-butanol (this oxy-
Ethane carboperoxoic acid, which was used to provide
mild and nonacidic oxidation conditions (after oxygen
transfer the resulting carbonic acid decomposes into the
neutral CO2 and ethanol),16 epoxidized the diols 4 in low
diastereoselectivity (dr ca. 60:40, entries 5 and 11). In
the oxidation of the like-diol 4a , 52% of the diastereo-
meric epoxides 5a and a total of 48% of the tetrahydro-
furan 7a (40%) and tetrahydropyran 8a (8%) were
obtained (entry 5). Moreover, a small amount of the enone
6 (2%) was detected. After two weeks at room tempera-
ture (20 °C), the crude reaction mixture had transformed
quantitatively to the cyclic products 7a and 8a in a ratio
(14) Itoh, T.; J itsukawa, K.; Kaneda, K.; Teranishi, S. J . Am. Chem.
Soc. 1979, 101, 159-169.
(12) Adam, W.; Saha-Mo¨ller, C. R.; Schmid, K. S. J . Org. Chem.
2000, 65, 1431-1433.
(13) Adam, W.; Saha-Mo¨ller, C. R.; Schambony, S. B.; Schmid, K.
S.; Wirth, T. Photochem. Photobiol. 1999, 70, 476-483.
(15) Adam, W.; Griesbeck, A. G.; Wang, X. Liebigs Ann. Chem. 1992,
193-197.
(16) Ru¨sch gen. Klaas, M.; Warwel, S. Synth. Commun. 1998, 28,
251-260.

Synthesis of 4,6-Dideoxyfuranoses
J . Org. Chem., Vol. 66, No. 22, 2001 7367
Sch em e 2. Syn th esis of th e Dia ster eom er ic Diols l-4a a n d u -4b
Sch em e 3. Ep oxid a tion of th e Diols 4 by Va r iou s Oxid a n ts
of 62:38. However, the oxidation of the unlike-diol 4b by
A number of suitable methods for the preparation of
trisubstituted tetrahydrofurans have been reported, which
include intramolecular nucleophilic attack of hydroxy
groups on epoxides21 and electrophilic22 as well as radi-
cal23 cyclizations. Protodesilylation of our tetrasubsti-
tuted tetrahydrofurans 7 would provide a convenient and
efficient entry to trisubstituted tetrahydrofurans. Un-
fortunately, when the desilylation of the tetrasubstituted
tetrahydrofuran 7 with CsF was tried, only Peterson-type
elimination products were obtained instead of the ex-
pected trisubstituted derivative. Also the conditions of
the homo-Brook rearrangement24 [KOtBu in DMSO/H2O
(10:1)] led to base-catalyzed elimination of dimethylphen-
ylsilanol (cf. Supporting Information). Thus, this poten-
tially attractive synthetic methodology was abandoned.
The nearly enantiomerically pure 4,6-dideoxy sugars
10 were obtained from the optically active hydroxy
vinylsilane 1. For this purpose, the racemic material was
submitted to the enzyme-catalyzed kinetic resolution by
acetylation.25 The BSL lipase (Burkholderia sp., CHIRA-
ethane carboperoxoic acid gave exclusively the diaster-
eomeric epoxides 5b (dr 64:36, entry 11). These were
transformed into a mixture (64:36) of the tetrahydrofuran
7b and the tetrahydropyran 8b on standing of the crude
product mixture at room temperature (20 °C) for two
weeks.
With dimethyldioxirane (DMD) as oxidant, allylic
oxidation to the enone 6 (54%) was observed instead of
epoxidation (<15%). This is not surprising, since DMD
is sensitive to steric effects, as previously observed for
similar vinylsilanes.17
For the Fleming oxidation (oxidative desilylation)18 of
the dimethylphenylsilyl to the hydroxy group by treat-
ment with Br2-AcOOH under retention of the configu-
ration, the alcohol functionalities in the tetrahydrofurans
7 were protected by acetylation. Under standard condi-
tions,19 the diacetates 9 were synthesized in 80-90%
yield (Scheme 4). The oxidation by KBr-AcOOH gave
both anomers of the 4,6-dideoxyfuranoses 10a (R:â ) 55:
45, 57% yield) and 10b (R:â ) 70:30, 62% yield). The
amount of R anomer may be enriched by crystallization
(e.g., for 6-deoxysorbose, R:â ) 95:5).20
(20) Rao, S. T.; Swaminathan, P.; Sundaralingam, M. Carbohydr.
Res. 1981, 89, 151-154.
(21) Fukuyama, T.; Vranesic, B.; Negri, D. P.; Kishi, Y. Tetrahedron
Lett. 1978, 31, 2741-2744.
(22) (a) Tonn, C. E.; Palazon, J . M.; Ruiz-Perez, C.; Rodriguez, M.
L.; Martin, V. S. Tetrahedron Lett. 1988, 29, 3149-3152. (b) Polt, R.;
Wijayaratne, T. Tetrahedron Lett. 1991, 32, 4831-4834.
(23) Hartung, J . Eur. J . Org. Chem. 2001, 619-632.
(24) Hudrlik, P. F.; Hudrlik, A. M.; Kulkarni, A. K. J . Am. Chem.
Soc. 1982, 104, 6809-6811.
(17) Adam, W.; Prechtl, F.; Richter, M. J .; Smerz, A. K. Tetrahedron
Lett. 1995, 36, 4991-4994.
(18) Fleming, I.; Henning, R.; Parker, D. C.; Plaut, H. E.; Sanderson,
P. E. J . J . Chem. Soc., Perkin Trans. 1 1995, 317-337.
(19) Adam, W.; Saha-Mo¨ller, C. R.; Schmid, K. S. Tetrahedron:
Asymmetry 1999, 315-322.

7368 J . Org. Chem., Vol. 66, No. 22, 2001
Ta ble 1. Ep oxid a tion of th e Dih yd r oxy Vin ylsila n es 4
Adam et al.
product ratio [%]b
furan 7
entry
1
4
oxidantc
time[h]
48
convna [%]
81
mba [%]
60
epoxide 5d
pyran 8
l-4a
Ti(OiPr)4,
ROOHf
e
>95
-
2
l-4a
VO(acac)2,
tBuOOH
27
94
79
e
>95
-
3
4
5
6
l-4a
l-4a
l-4a
u-4b
MTO/UHP
mCPBA
EtOC(O)O2H
Ti(OiPr)4,
ROOHf
36
24
60
20
61
>95
>95
>95
47
65
>95
40
g
66
69
40
34
23
08
-
8 (80:20)
52 (55:45)h
e
>95
7
8
u-4b
u-4b
VO(acac)2,
tBuOOH
VO(acac)2,
tBuOOHi
MTO/UHP
mCPBA
12
10
93
60
82
70 (>95:05)
30
-
-
>95
e
>95
9
10
11
u-4b
u-4b
u-4b
36
24
40
71
>95
75
43
79
89
j
64
16
-
36
13
-
71 (96:04)k
>95 (64:36)l
EtOC(O)O2H
a
Conversion and mass balance (mb) were determined by NMR spectroscopy after silica gel chromatography. bProduct ratios after
silica gel chromatography, determined by NMR spectroscopy (error (5% of the stated values), except entries 5 and 11, which were analyzed
as crude reaction mixtures. cOxidations were conducted in CHCl3 or CH2Cl2 at 20 °C, except entries 5 and 11, for which diethyl carbonate
was used as solvent at 40 °C. dThe erythro:threo ratio is given in parentheses. eOn monitoring the reaction mixture by 1H NMR spectroscopy,
the intermediary erythro-5 epoxide (dr >95:5) was observed, which was totally converted to the tetrahydrofuran 7 at the end of the
reaction. f2,3-Dimethyl-3-hydroperoxy-2-butanol was used as oxygen donor. gOn monitoring the reaction mixture by 1H NMR spectroscopy,
the intermediary epoxides erythro-5a :threo-5a (dr ) 65:35) were observed. hAfter two weeks at 20 °C, the crude reaction mixture had
i
j
transformed quantitatively to 7a and 8a (62:38). The reaction mixture was heated to 45 °C. On monitoring the reaction mixture by 1H
NMR spectroscopy, the intermediary epoxides erythro-5b:threo-5b (dr ) 64:36) were observed. kThe dr value of the epoxide 5b (erythro:
threo) was 81:19 at the beginning, but changed as the reaction progressed. After two weeks at 20 °C, the crude reaction mixture had
l
transformed quantitatively to 7b and 8b (64:36).
Sch em e 4. Acetyla tion of th e Tetr a h yd r ofu r a n s 7 a n d Oxid a tive Desilyla tion of th e Resu ltin g Aceta tes 9
to th e 4,6-Did eoxyfu r a n oses 10
Sch em e 5. Lip a se-Ca ta lyzed Kin etic Resolu tion of
th e Hyd r oxy Vin ylsila n e 1, Its Desilyla tion to th e
Hom oa llylic Alcoh ol 12, a n d th e Hyd r olysis of th e
Aceta te (R)-11 to th e Hyd r oxy Vin ylsila n e (R)-1
ZYME L-1, Boehringer Mannheim) was chosen for the
reactions at the semipreparative scale (4.5 mmol), in
which at 52% conversion the nearly enantiomerically
pure derivatives (S)-1 (>98% ee) and (R)-11 (90% ee) were
obtained (Scheme 5). The mixture of the alcohol (S)-1 and
the acetate (R)-11 was readily separated quantitatively
by silica gel chromatography.
The S configuration of the hydroxy vinylsilane (+)-1
was determined by chemical correlation. Thus, desilyla-
tion of (+)-1 with toluenesulfonic acid in acetonitrile/
water (6:1) gave the known homoallylic alcohol (S)-12
(Scheme 5, 60% yield).19 The configuration in this enan-
tioselection is in accordance with the established empiri-
cal rule proposed by Kazlauskas et al.26 for the lipase-
catalyzed kinetic resolution of secondary alcohols. To
obtain the (R)-1 alcohol, the acetate (R)-11 was hydro-
lyzed quantitatively by K2CO3 in methanol (90% ee,
Scheme 5). As reference samples for the chiral HPLC
analysis, it was necessary to prepare the racemic acetate
(25) Wong, C. H.; Whitesides, G. M. Enzymes in Synthetic Organic
Chemistry; Pergamon Press: Oxford/New York/Tokyo, 1994; pp 41-
130.
(26) Kazlauskas, R. J .; Weissfloch, A. N. E.; Rappaport, A. T.; Cuccia,
L. A. J . Org. Chem. 1991, 56, 2656-2665.
11 by acetylation of the alcohol 1 with acetic anhydride
under basic conditions (92% yield).19

Synthesis of 4,6-Dideoxyfuranoses
J . Org. Chem., Vol. 66, No. 22, 2001 7369
F igu r e 1. Determination of the relative configuration for the tetrahydrofurans 7 by NOESY experiments.
Sch em e 6. Syn th esis of th e Op tica lly Active
4,6-Did eoxysor bofu r a n oses 10a a n d
4,6-Did eoxyfr u ctofu r a n oses 10b fr om th e
En a n tiom er ica lly En r ich ed Hyd r oxy Vin ylsila n es
(R)-1 a n d (S)-1
arrows b indicate a trans arrangement between the silyl
group at C-2 and the neighboring C-3 hydroxy group.
Thus, the silyl-substituted C-2 carbon atom possesses the
2S* configuration. Since the intramolecular epoxide-ring
opening (SN2 process) of 5a to the tetrahydrofuran 7a
occurs with inversion of the configuration at the silyl-
substituted R epoxide reaction center, the latter stereo-
genic carbon atom must have the RR* relative configu-
ration. From these data we infer the (RR*,1R*,3R*)
configuration for the erythro-5a epoxide. In the same way
(Figure 1, right), from the NOE effects of the tetrahy-
drofuran (2S*,3R*,5S*)-7b, an unlike configuration for
the diol (2S*,4R*)-4b and a (RR*,1R*,3S*) configuration
for the erythro-5b epoxide may be concluded.
The relative configuration of the tetrahydropyran 8a
was found to be (2R*,4R*,5S*), as shown in Figure 2
(left). By proceeding as in the case of the tetrahydrofuran
7a , from the NOE effects coded by the solid arrows a , a
like configuration may be assigned for the two stereogenic
centers of the diol (2R*,4R*)-4a . The NOE effects coded
by the dashed arrows b indicate a cis arrangement of the
silyl group at C-5 and the neighboring C-4 hydroxy group
and a 5S* configuration for the silyl-substituted C-5
carbon atom. The corresponding threo-5a epoxide pos-
sesses an RS configuration, because the R-carbon stereo-
genic center is not changed during this epoxide-opening
reaction (â attack). In the same way, from the NOE
effects of the tetrahydropyran (2S*,4R*,5S*)-8b an unlike
configuration for the diol (2S*,4R*)-4b and a (RS*,1R*,3S*)
configuration for the threo-5b epoxide may be concluded
(Figure 2, right).
Both enantiomers of the hydroxy vinylsilane 1 were
submitted to diastereoselective oxyfunctionalization by
photooxygenation and subsequent reduction of the hy-
droperoxides 2 by Ph3P to afford the diastereomeric diols
4. The latter were epoxidized by tert-butyl hydroperoxide
under VO(acac)2 catalysis. The intermediary epoxides 5
were transformed to the tetrahydrofurans 7 by intramo-
lecular attack of the C-3 hydroxy group. After oxidative
desilylation18 of the diacetates 9 (Scheme 6), the D- and
L-4,6-dideoxysorbofuranose 10a and D- and L-4,6-dideoxy-
fructofuranose 10b were obtained as anomeric mixtures
in 90 to >98% ee, which match the ee values of the
starting hydroxy vinylsilanes (R)-1 and (S)-1.
The relative configurations of the tetrahydrofurans 7
and the tetrahydropyrans 8 were determined by NOESY
experiments, those of the diols 4 and the epoxides
erythro-5 and threo-5 by chemical correlation. The rela-
tive configuration of the tetrahydrofuran 7a was found
to be (2S*,3R*,5R*), as shown in Figure 1 (left). The NOE
effects coded by the solid arrows a suggest a trans
arrangement of the C-3 hydroxy (3R*) group and the C-5
methyl (5R*) group, from which a like configuration may
be gathered for the two hydroxy substituents in the diol
(2R*,4R*)-4a . The NOE effects coded by the dashed
Discu ssion
The best results for the synthesis of the tetrahydro-
furans 7 were obtained by the metal-catalyzed (V, Ti)
epoxidation of the diols 4. Thus, the diols 4 were
converted by Ti(OiPr)4/ROOH or VO(acac)2/tBuOOH to
the epoxides erythro-5 in excellent diastereoselectivity (dr
>95:5), but which cyclized in situ to the tetrahydrofurans
7 in up to 82% yield (Scheme 3). The epoxidation of the
diols 4 by MTO/UHP or peracids was found to be poorly
diastereoselective (erythro-5: threo-5 ) 60:40 to 80:20).
The erythro-5 epoxides were converted to the tetrahy-
drofurans 7, whereas the tetrahydropyrans 8 cyclization
products were obtained from the threo-5 epoxides.
The diastereoselectivity in the metal-catalyzed [Ti-
(OiPr)4, VO(acac)2] epoxidation of the diols 4 with hy-
droperoxides as oxygen source was shown to be excellent
(erythro-5:threo-5 > 95:5). The nearly exclusive erythro
diastereoselectivity may be explained by metal-alcoholate
binding between the allylic hydroxy group and the metal
center, which requires a dihedral angle θ of ∼50-70° for
effective oxygen-atom transfer (Figure 3).9b Thus, the
erythro transition geometry is favored due to diminution

7370 J . Org. Chem., Vol. 66, No. 22, 2001
Adam et al.
F igu r e 2. Determination of the relative configuration for the tetrahydropyrans 8 by NOESY experiments.
Sch em e 7. Regioselectivity in th e In tr a m olecu la r
Ep oxid e-Rin g-Op en in g Rea ction of th e Er yth r o-5
Ep oxid e to th e F u r a n r a c-7
F igu r e 3. Postulated transition structures for the epoxidation
by hydroperoxides under titanium or vanadium catalysis.
influence on the diastereoselectivity of the epoxidation.
The only difference was the higher persistence of the
isolable (RR*,1R*,3S*)-5b from the unlike-diol 4b, whereas
the (RR*,1R*,3R*)-5a epoxide from the like-4a was
converted to the tetrahydrofuran product 7a through in-
situ ring closure.
For the nucleophilic ring opening of the erythro-5 and
threo-5 epoxides, attack at the R or â positions of the
epoxide ring is possible (Scheme 7). The erythro-5 epoxide
is converted to the tetrahydrofuran 7, whereas the threo-5
epoxide is transformed to the tetrahydropyran 8. A cross
reaction (erythro 5 to 8 or threo 5 to 7) may be precluded
as the ratio of erythro-5: threo-5 is the same as the 7:8
ratio (Table 1, entry 11 and foodnote l).
F igu r e 4. Postulated transition structures for the epoxidation
by peracids.
of 1,2-allylic (1,2A) strain. The bulky silyl group, which
was introduced in the homoallylic alcohol 1 to generate
good regioselectivity in the photooxygenation reaction,27
is also responsible for this perfect diastereoselection
between the erythro and threo transition states on
account of 1,2A strain. For the unsilylated diol, the
diastereoselectivity should drop due to less steric repul-
sion, as it is known for the direct Ti(OiPr)4-catalyzed
epoxidation of structurally related hydroperoxy homo-
allylic alcohols.28
In comparison to the vanadium- or titanium-catalyzed
epoxidations, the diastereoselectivity of the reactions
with MTO/UHP29 and peracids15 was lower. Hydrogen
bonding between the allylic alcohol and the oxidant
requires a dihedral angle θ of ∼120° in the transition
state and, therefore, 1,2A strain is less important (Figure
4). However, in this geometry the 1,3-allylic strain (1,3A)
plays a decisive role for the control of diastereoselectiv-
ity.29,30 Due to the lack of 1,3A strain in the diols 4 (no cis
substituent), only a moderate diastereoselectivity was
expected in this epoxidation reaction, as observed.
The relative configuration of the two hydroxy groups
(like and unlike diastereomers) in the diols 4 had no
The intramolecular nucleophilic attack in the erythro-5
epoxide occurs regioselectively (>95:5) at the R position
to afford the tetrahydrofuran product 7 (Scheme 7). This
SN2 attack of the C-3 hydroxy group leads to only one
diastereomer of the tetrahydrofuran 7 (dr >95:5), with
inversion of the configuration at the R position (Table 1,
entries 1, 2, 6, and 8). The preference for the R attack is
based on the more favorable five-membered-ring transi-
tion state exo-TS, which is stabilized either by hydrogen
bonding or metal-alcoholate ligation between the C-1
hydroxy group and the epoxide oxygen (Scheme 7).
Additionally, the stereoelectronic effect of the silyl sub-
stituent weakens the CR-O bond,31 and therefore tetra-
hydrofuran formation is favored. The en d o-TS′ transition
state, which would lead to the tetrahydropyran product
8, is disfavored, because the silyl group would take an
axial position in the six-membered pyran ring. Thus, no
tetrahydropyran 8 is formed from the erythro-5 epoxide.
The relative configuration (like or unlike) of the two
hydroxy groups at the C-1 and C-3 positions of the
erythro-5a and erythro-5b epoxides had no influence on
the ring-opening pathway (R versus â attack; Table 1,
(27) Adam, W.; Richter, M. J . J . Org. Chem. 1994, 59, 3341-3346.
(28) Adam, W. Nestler, B. J . Am. Chem. Soc. 1993, 115, 7226-7231.
(29) Adam, W.; Mitchell, C. M.; Saha-Mo¨ller, C. R. J . Org. Chem.
1999, 64, 3699-3707.
(30) Adam, W.; Corma, A.; Reddy, T. I.; Renz, M. J . Org. Chem. 1997,
62, 3631-3637.
(31) (a) Fristad, W. E.; Bailey, T. R.; Paquette, L. A.; Gleiter, R.;
Bo¨hm, M. C. J . Am. Chem. Soc. 1979, 101, 4420-4423. (b) Adam, W.;
Prein, M.; Richter, M. J . Tetrahedron 1996, 52, 1231-1234.

Synthesis of 4,6-Dideoxyfuranoses
J . Org. Chem., Vol. 66, No. 22, 2001 7371
Sch em e 8. Regioselectivity in th e In tr a m olecu la r
Ep oxid e-Rin g-Op en in g Rea ction of th e th r eo-5
Ep oxid e to th e P yr a n r a c-8
It should be emphasized that the silyl-substituted
tetrahydrofurans 7, accessible through the above-men-
tioned diastereoselective route, are promising precursors
for dideoxy sugars; moreover, the conversion of the
dimethylphenylsilyl substituent of the tetrahydrofurans
7 to a hydroxy group provides 4,6-dideoxyhexoses. In this
way, the oxidative desilylation of the dimethylphenylsilyl
group in the diacetate derivatives 9 of the tetrahydro-
furanes 7 with KBr-AcOOH (Fleming oxidation)18 af-
forded the 4,6-dideoxyfuranoses 10 in up to 62% yield
(Scheme 4). This transformation offers the opportunity
to prepare the hitherto unknown optically active 4,6-
dideoxyfuranoses 10 by starting form the enantiomeri-
cally enriched hydroxy vinylsilanes (R)-1 (90% ee) and
(S)-1 (>98% ee). Thus, through this newly developed
methodology, that is, the sequence regioselective photo-
oxygenation, triphenylphosphine reduction, diastereose-
lective epoxidation with VO(acac)2/tBuOOH, and oxida-
tive desilylation, the optically active D- and L-4,6-
dideoxysorbofuranoses 10a and D- and L-4,6-dideoxy-
fructofuranoses 10b have been prepared for the first time
in nearly enantiomerically pure form (Scheme 6).
entries 1 vs 6 and 2 vs 8). This is not surprising in view
of the small energy difference (1-4 kJ /mol) between the
pseudoaxial and pseudoequatorial substituents in the
transition state of the five-membered-ring, oxygen-
containing, saturated heterocycles.32 Consequently, both
erythro-epoxides (RR*,1R*,3R*)-5a and (RR*,1R*,3S*)-
5b were converted exclusively to the respective tetrahy-
drofurans (2S*,3R*,5R*)-7a and (2S*,3R*,5S*)-7b.
The threo-5 epoxide was converted exclusively to the
tetrahydropyran 8 by attack of the C-3 hydroxy group at
the â position of the epoxide ring (Scheme 8; Table 1,
entries 3-5 and 9-11).
Formation of the tetrahydrofuran 7 is unfavorable
because of the steric repulsion between the bulky silyl
group and the C-1 hydroxy group in the five-membered-
ring transition state exo-TS′. Therefore, the six-membered-
ring transition state en d o-TS, for which steric repulsion
is minimized, is preferred and only the tetrahydropyran
product 8 is formed. Again, the intramolecular ring
closure of the threo-5 epoxides to the tetrahydropyran
products 8 is independent of the relative configuration
of the two hydroxy groups (Table 1, entries 3 vs 9, 4 vs
10, and 5 vs 11).
The best results for the preparation of the tetrahydro-
furans 7 were obtained with the oxidants Ti(OiPr)4/
ROOH or VO(acac)2/tBuOOH, because no tetrahydropy-
rans 8 were formed (Table 1, entries 1, 2, 6, and 7). Since
both the epoxidation (erythro selective) as well as the
nucleophilic ring closure (exo selective) occur diastereo-
selectively (dr > 95:5), the diols 4a and 4b were converted
to only one of the possible diastereomers of the tetrahy-
drofuran 7, namely the corresponding (2S*,3R*,5R*)-7
and (2S*,3R*,5S*)-7. Epoxidation with VO(acac)2/tBuOOH
is more convenient because only 1 mol % of the cheap
VO(acac)2 catalyst is necessary, while for the Ti(OiPr)4-
catalyzed oxidation the tridentate hydroperoxide, namely
2,3-dimethyl-3-hydroperoxy-2-butanol, is not commer-
cially available and needs to be prepared.10
Con clu sion
The hydroxy vinylsilane, E-5-dimethylphenylsilyl-2-
hexen-4-ol (1), has been selectively transformed to the
hitherto unknown 4,6-dideoxyfuranoses 10a and 10b by
employing successively three versatile oxyfunctionaliza-
tion methods, namely photooxygenation, metal-catalyzed
epoxidation, and oxidative desilylation. The application
of this strategy to the enantiomerically enriched hydroxy
vinylsilane 1 has made available for the first time the
optically active (up to >98% ee) 4,6-dideoxysorbofuranose
10a and 4,6-dideoxyfructofuranose 10b. The availability
of these unusual sugar derivatives through our novel
synthetic methodology opens the opportunity to assess
their potential biologically activity.
Ack n ow led gm en t. This work was supported by the
Deutsche Forschungsgemeinschaft (SFB-347, “Selektive
Reaktionen Metall-aktivierter Moleku¨le”) and the Fonds
der Chemischen Industrie. We thank Roche Diagnostics
GmbH, Penzberg, Germany, for the generous gift of
enzymes.
Su p p or tin g In for m a tion Ava ila ble: Reported are the
synthetic details of the epoxidations, the characteristic spectral
data of the epoxidation products 5-8, and the synthesis and
characterization of the starting materials 1, 2, and 4, as well
as of the diacetates 9 and 4,6-dideoxyfuranoses 10. This
material is available free of charge in the Internet under
http://pubs.acs.org.
(32) Hartung, J .; Gallou, F. J . Org. Chem. 1995, 60, 6706-6716.
J O010549W