New Factors Governing Stereoselectivity in Borohydride Reductions
FULL PAPER
index detector was used. Mixtures of gluco/manno-epimers were
separated on LiChrosorb Si 0.5 µm column (0.4 ϫ 20 cm) with
different (appropriate) ratios of hexane/EtOAc as the mobile phase
flowing at a rate of 1 mL minϪ1 at 40 °C. Effluent was monitored
at 254 nm for compounds bearing benzyl ethers and by refractive
index detector for compounds containing non-UV absorbing pro-
tecting groups. Products were identified by using relevant stand-
ards. Solvents were HPLC grade.
cosidulose 4 shown in entries G, H, and J. These experi-
ments afforded stereochemical results identical with those
obtained for ulose 4 (entry K, L, and M). Thus, the yield
obtained in the preparation of ethyl (3-O-allyl-4,6-O-isop-
ropylidene-β--mannopyranosyl)-(1Ǟ4)-3,6-di-O-benzyl-2-
deoxy-2-phthalimido-1-thio-β--glucopyranoside
(12)
could be increased from 53% up to 81% (entry C and L).
The suitably protected disaccharide-thioglycoside building
block 12, which contains a β-mannosidic linkage, has been
successfully applied in the synthesis of complex fucosylated
and nonfucosylated core structures of xylose containing
carbohydrate chains from N-glycoproteins.[3e,9]
In order to establish the stereochemical outcome of the
carbonyl reduction in the absence of the 4,6-O-acetal func-
tion both in the presence and absence of "activated"
Methyl 3-O-Allyl-4,6-O-isopropylidene-β-D-glucopyranoside (5): p-
Toluenesulfonic acid (100 mg, 0.53 mmol) was added to a stirred
suspension of methyl 3-O-allyl-β--glucopyranoside[5] (1 g,
4.27 mmol) in 2,2-dimethoxypropane (5 mL, 40.66 mmol) at room
temperature. The reaction was stopped after 20 min by adding so-
dium hydrogen carbonate. The mixture was diluted with dichloro-
methane and the organic layer was washed with water, dried, and
concentrated. Column chromatography (6:4 hexane/EtOAc, Rf
DMSO, the preparation of β--glycosidulose 13 was carried 0.51) of the syrupy residue afforded 5 (1.08 g, 92%). [α]D Ϫ28.5
1
(c ϭ 0.34, CHCl3). Ϫ H NMR (CDCl3): δ ϭ 1.42 and 1.50 (2 ϫ
out according to the procedure described for compound 4.
Borohydride reduction of 13, obtained by workup protocol
A, and reduction of 13, obtained by workup protocol B,
provided a stereoselective and a highly manno-selective (an
essentially stereospecific) course of reduction (entries N and
O), respectively. In monosaccharide-uloside cases 2 and 3,
obtained by workup protocol B, carbonyl reductions pro-
ceeded in a stereospecific manner (entries P and Q).
s, each 3 H, CMe2), 2.73 (br. s, 1 H, OH), 3.26 (m, 1 H, H-5), 3.43
(m, 2 H, H-6a and H-6b), 3.56 (s, 3 H, OMe), 3.65 (m, J2,3
ϭ
10.0 Hz, 1 H, H-2), 3.78 (dd, J3,4 ϭ 10.0 Hz, 1 H, H-3), 3.93 (dd,
J4,5 ϭ 5.5 Hz, 1 H, H-4), 4.28 (d, J1,2 ϭ 7.5 Hz, 1 H, H-1), 4.31
(m, 2 H, CH2ϭCHϪCH2O), 5.25 (m, 2 H, CH2ϭCHϪCH2O),
5.95 (m, 1 H, CH2ϭCHϪCH2O). Ϫ 13C NMR (CDCl3): δ ϭ 19.1
and 29.1 (CMe2), 57.3 (OMe), 62.1 (C-6), 73.3 (CH2ϭCHϪCH2O),
99.3 (CMe2), 104.2 (C-1), 117.0 (CH2ϭCHϪCH2O), 135.1 (CH2ϭ
CHϪCH2O). Ϫ C13H22O6 (274.14): calcd. C 56.90, H 8.09; found
C 56.83, H 8.14.
Conclusion
Methyl 3-O-Allyl-4,6-O-isopropylidene-β-D-mannopyranoside (6)
(Table, entry B): A solution of 5 (750 mg, 2.73 mmol) in 1:2 acetic
anhydride/methyl sulfoxide (6 mL) was kept at room temperature
for 16 h and then concentrated. The residue was dissolved in
dichloromethane and washed with water, dried, and concentrated.
The crude 2-ulose and its hydrate 3 were dissolved in 1:1 dichloro-
methane/methanol (6 mL) and cooled to 0 °C. Sodium borohydride
(517 mg, 13.67 mmol) was added to the reaction mixture in one
portion. After stirring for 1 h at room temperature, the mixture
was diluted with dichloromethane, washed with water, dried, and
concentrated. TLC (6:4 hexane/EtOAc) of the residue showed com-
plete disappearance of 3, traces of 5 (Rf 0.51), and a major product
6 (Rf 0.30). Column chromatography of the residue gave 6 (638 mg,
85%). [α]D Ϫ61.9 (c ϭ 0.27, CHCl3). Ϫ 1H NMR (CDCl3): δ ϭ
1.42 and 1.52 (2 ϫ s, each 3 H, CMe2), 2.52 (br. s, 1 H, OH), 3.18
(m, 1 H, H-5), 3.47 (dd, J3,4 ϭ 9.3 Hz, 1 H, H-3), 3.56 (s, 3 H,
OMe), 3.90 (m, 2 H, H-6a and H-6b), 4.08 (dd, J4,5 ϭ 9.6 Hz, 1
H, H-4), 4.12 (m, J2,3 ϭ 3.3 Hz, 1 H, H-2), 4.22 (m, 2 H, CH2ϭ
CHϪCH2O), 4.41 (d, J1,2 Ͻ 1 Hz, 1 H, H-1), 5.25 (m, 2 H, CH2ϭ
CHϪCH2O), 5.92 (m, 1 H, CH2ϭCHϪCH2O). Ϫ 13C NMR
(CDCl3): δ ϭ 19.2 and 29.2 (CMe2), 57.2 (OMe), 62.1 (C-6), 71.3
(CH2ϭCHϪCH2O), 99.7 (CMe2), 101.4 (C-1), 117.4 (CH2ϭ
CHϪCH2O), 134.7 (CH2ϭCHϪCH2O). Ϫ C13H22O6 (274.14):
calcd. C 56.90, H 8.09; found C 56.94, H 8.11.
New factors governing stereoselectivity in borohydride
reductions of β--glucoside-2-uloses were observed. The
substitution pattern influenced the stereoselectivity of the
reduction as follows: the 3-O-benzyl and 3-O-allyl ethers
vicinal to the carbonyl to be reduced had a similar influence
on the steric outcome of the carbonyl reduction. The pres-
ence of a 4,6-O-acetal function resulted in lower stereo-
selectivity in the monosaccharide-uloside cases and low ste-
reoselectivity in the disaccharide-uloside cases, while the ab-
sence of a 4,6-O-acetal group provided distinctly higher ste-
reoselectivity.
A
peculiar
effect
of
acetoxydi-
methylsulfonium acetate (‘‘activated’’ DMSO) was ob-
served. In all cases, the presence of this species strongly in-
creased the manno-selectivity of the reduction even in the
presence of a 4,6-O-acetal function. A simple, preparatively
expedient, generally applicable protocol has been elaborated
for achieving high manno-selectivities and, hence, satisfact-
ory yields. A systematic study of the nature of the selectivity
enhancement by ‘‘activated’’ DMSO is in progress.
Experimental Section
Ethyl (2,4,6-Tri-O-acetyl-3-O-benzyl-β-
3,6-di-O-benzyl-2-deoxy-2-phthalimido-1-thio-β-
D
-glucopyranosyl)-(1Ǟ4)-
D
-glucopyranoside
General: Optical rotations were measured with a PerkinϪElmer 241
polarimeter. NMR spectra were recorded with a Bruker WP-200
SY spectrometer for solutions in CDCl3 (internal Me4Si). The reac-
tions were monitored by TLC on Kieselgel 60 F254 (Merck, Darm-
(7): A solution of ethyl 3,6-di-O-benzyl-2-deoxy-2-phthalimido-1-
thio-β--glucopyranoside[3f] (720 mg, 1.35 mmol) and 2,4,6-tri-O-
acetyl-3-O-benzyl-α--glucopyranosyl
bromide[6]
(1.86 g,
4.05 mmol) in dichloromethane (13 mL) containing powdered 4A
stadt) with detection by UV light and/or by charring with 50% molecular sieves (3 g) was stirred for 30 min under argon. A solu-
sulfuric acid. Column chromatography was performed on silica gel
60 (63Ϫ200 µm). For HPLC a Hewlett Packard 1090 series II Li-
quid Chromatograph equipped with a diode array and refractive
tion of silver triflate (1.40 g, 5.44 mmol) in toluene (33 mL) was
added dropwise in the dark during 1.5 h at Ϫ45 °C and stirring
was continued for 1 h at Ϫ40 °C. Pyridine (2 mL) was added and
Eur. J. Org. Chem. 2000, 3931Ϫ3935
3933