Epoxidation of C-branched glycals: Unexpected stereochemical results and their theoretical rationale

Article abstract of DOI:10.1016/S0008-6215(02)00406-8

This paper describes the synthesis of C-3 methyl-branched glycosides by epoxidation of partially unblocked L-configured glycals. The stereochemical result depends on the orientation of the allylic hydroxyl group. A theoretical explanation is presented, based on the conformational preferences of the respective glycal half-chair conformations that were estimated by applying the BP density functional and a valence triple-ζ basis set.

Full text of DOI:10.1016/S0008-6215(02)00406-8

Carbohydrate Research 338 (2003) 231–236
www.elsevier.com/locate/carres
Epoxidation of C-branched glycals: unexpected stereochemical
results and their theoretical rationale
Christiane Ernst,a Manuel Piacenza,a Stefan Grimme,a Werner Klaffkea,b,*
aOrganisch-Chemisches Institut der Westfa¨lischen Wilhelms-Uni6ersita¨t, Corrensstr. 40, D-48149 Mu¨nster, Germany
bUnile6er Research and De6elopment, Unile6er Health Institute, Oli6ier 6an Noortlaan 120, NL-3130 AT Vlaardingen, The Netherlands
Received 26 June 2002; accepted 10 October 2002
Dedicated to Professor Hans Paulsen in the year of his 80th birthday.
Abstract
This paper describes the synthesis of C-3 methyl-branched glycosides by epoxidation of partially unblocked
L
-configured
glycals. The stereochemical result depends on the orientation of the allylic hydroxyl group. A theoretical explanation is presented,
based on the conformational preferences of the respective glycal half-chair conformations that were estimated by applying the BP
density functional and a valence triple-z basis set. © 2002 Elsevier Science Ltd. All rights reserved.
Keywords: C-Branched sugars; BP density functional; Epoxidation; Glycal
1. Introduction
2. Results and discussion
1,2-Anhydro sugars have proven to be invaluable tools
for the synthesis of glycosides1,2 and nucleoside diphos-
phates.3 Transformation into Brigl’s anhydride prior to
glycosylation, however, has been limited to protected
glycals, and, moreover, the biologically interesting fam-
ily of C-branched sugar glycosides has been omitted in
previous synthetic papers. Here, we wish to report on
Starting from the easily accessible 3-C-branched gly-
cals,
L-olivomycal (1) and
L
-mycaral (2),4 acetylation
gave the monoesters 3 and 4 (cf Scheme 1). Both
derivatives were reacted with freshly prepared
dimethyldioxirane5 (DMDO) in a solution of acetone
and dichloromethane. In the case of
expected b-facial epoxidation, which occurred cis to the
allylic hydroxy group, furnished b- -manno-configured
5. Nucleophilic attack by methanol resulted in a selec-
-mannopy-
L-olivomycal, an
the epoxidation of L-mycaral and L-olivomycal as pro-
L
totypes for C-3-branched deoxy sugars. It was our
intention to utilise the glycal route for the synthesis of
the respective activated sugars. Since the outcome of
the epoxidation step seemed stereochemically not uni-
form, we embarked on this experimental and theoretical
study to shed some light on the attack by dimethyl-
dioxirone (DMDO) on the enol ether double bond.
tive trans opening of the epoxide to yield a-
L
ranoside 6. Interestingly, polymerization of this
bifunctional substrate could not be detected.
In the case of
L-mycaral, a similar cis epoxidation
was expected as well. However, in contrast the pre-
ferred attack took place from the b-face and occurred
trans to the allylic hydroxy group. The b-L-altro-
configured isomer 10 was formed in a 4:1 ratio com-
pared to 11. The configurations of both products were
proven again using methanolysis to form methyl 4-O-
acetyl-6-deoxy-3-C-methyl-a-
L
-altropyranoside
(12)
-allopy-
and methyl 4-O-acetyl-6-deoxy-3-C-methyl-b-
L
* Corresponding author. Tel.: +49-251-83-33271; fax: +
49-251-83-36501
ranoside (13), respectively. The products could not be
separated from one another, but their H NMR and
1
E-mail address: werner.klaffke@unilever.com, werner.
klaffke@uni-muenster.de (W. Klaffke).
GCOSY spectra gave clear signals and coupling con-
0008-6215/03/$ - see front matter © 2002 Elsevier Science Ltd. All rights reserved.
PII: S0008-6215(02)00406-8

232
C. Ernst et al. / Carbohydrate Research 338 (2003) 231–236
geometries to compensate for the high rotational bar-
rier of the acetyl group were taken into account for
each conformer of 3 and 4: (i) the carbonyl oxygen
situated on the a-face, above the six-membered ring
(endo) and (ii) the carbonyl oxygen pointing away from
the ring (exo). The results of the calculations for
olivomycal 3 are illustrated in Fig. 1. Compound 3-C,
stants for both products [12: J1,2 1.4 Hz (eq, eq); 13: J1,2
7.8 Hz (ax,ax)].
The literature offers mainly two explanations on the
diastereoselectivity of epoxidations employing DMDO:
steric effects6,7 and hydrogen-bonding effects.8–10 The
hydroxy group exerts a stabilizing effect through hydro-
gen bonding on the dipolar DMDO transition state.
This effect is more pronounced in nonpolar solvents
(acetone/CCl4 mixtures) than in polar environments
(methanol), where in the latter case the solvent sta-
bilises the transition state, and, therefore, steric interac-
tions prevail. Studies by Adam and co-workers
discussed the importance of the size of the dihedral
angle a between the allylic hydroxy group and the
epoxidised double bond in cyclohexenols.10 The expla-
nation offered by these authors is that dihedral angles
larger than 130° favour in the transition state hydrogen
bonding between the allylic hydroxy group and the
dioxirane reagent. This furnishes a cis relationship be-
tween the 3-hydroxy group and the formed epoxide.
The further rationale is that with smaller angles severe
steric interference would lead to anti epoxidations.
Thus, a better understanding of the ground-state
conformations of the glycal should offer a reliable
prediction of the expected stereocontrol. A computa-
5
which has a H4-half-chair conformation is calculated
to be of the lowest energy. For both endo conformers
3-B and 3-D, the relative energies are found to be
substantially higher (6.3 and 10.3 kcal/mol). Although
the dihedral angle of 3-C (127°) is very close to the
optimal angle calculated by Adam and co-workers,10
the predicted hydrogen bond prevents an effective cis
stereocontrol brought about by the allylic OH group.
Therefore, the cis-stereoselectivity of this reaction is
exclusively due to steric restrictions. Both the axially
configured methyl group and the acetate at the C-4
position favour a b-facial dioxirane attack.
The exclusive steric control should be maintained
even if the protecting group at C-4 is changed to a
smaller substituent not capable of hydrogen bonding.
Thus, alcohol 1 was etherified with iodomethane to
yield glycal 7. Epoxidation and subsequent hydrolysis
proved to be equally stereoselective as was the case with
compound 3. Computation for 7 (cf. Fig. 2) revealed
tional study was carried out at the D
6 ensity F6 unctional
5
T6
heory (DFT) level of theory. The ground-state ge-
the H4-half-chair conformation to be energetically fa-
4
ometries of 3, 4, and 7 were considered for the H5 and
voured. Here, the methyl ether blocks the intramolecu-
lar hydrogen-bond interaction with the tertiary OH
5H4 conformers, respectively. Two additional starting
Scheme 1.

C. Ernst et al. / Carbohydrate Research 338 (2003) 231–236
233
Fig. 1. Optimised structures of 3 and relative energies of four low-lying conformers. The dihedral angle a refers to the atoms
C-1–C-2–C-3–OH.
Fig. 2. Optimised structures of 7 and relative energies of two low-lying conformers. The dihedral angle a refers to the atoms
C-1–C-2–C-3–OH.
substituents, except for the allylic hydroxyl group, are
in a pseudoaxial position, with a dihedral angle a of
about 129°. This would favour the cis transition state;
however, the steric effect excerted by the pseudoaxial
ester is prevailing and counteracting. As it is known
that DMDO is very sensitive to steric factors,7 the steric
effect seems to influence the favoured transition state
more than the directive effect of the allylic OH group.
In the experiment these competing effects lead to a
preferred trans epoxidation with a minor part of cis
epoxidation. An increased steric demand of the sub-
stituent would conclusively overrule any other effect
and shift the ratio entirely in favour of the b-facial and
thus trans epoxide.
group, and the pseudo equatorial position of the hy-
droxyl group bends the dihedral angle to a value of
131.7°.
Though the two conformers are energetically almost
equal, 7-B seems to be the more reactive species due to
the larger dihedral angle of 131.7°. As predicted, this
angle favours the association between the hydroxy
group and the attacking dioxirane, and it may be
therefore safe to assume that the cis selectivity in this
reaction can be attributed to a pre-association between
the reagent and the HO-3 group.
4
The calculations for
L-mycaral (4) indicated the H5-
half-chair conformation (4-A) to be energetically fa-
voured (cf. Fig. 3). In the favoured conformer 4-A all

234
C. Ernst et al. / Carbohydrate Research 338 (2003) 231–236
Fig. 3. Optimised structures of 4 and relative energies of four low-lying conformers. The dihedral angle a refers to the atoms
C-1–C-2–C-3–OH.
3. Conclusions
Micromass Quattro LC-Z spectrometer. Melting points
were determined on a Ta Instruments 2010 Differential
Scanning Calorimeter. Reactions were monitored by
It has been shown by the above experiments and subse-
quent calculations that H-bonding between the ester
carbonyl and the tertiary hydroxyl group has a
paramount effect on the stability of the relative con-
formers. Small changes in the substitution patterns
have large effects on the conformational preferences of
the glycal half-chairs, which consequently results in a
specific product distribution. On the computational side
the need of sufficiently large basis sets to allow definite
conclusions became evident, a fact which has been also
identified by Allinger and co-workers.11 Synthetically
this paper describes the generation of novel C-3 methyl-
branched glycosides and comprises the first example of
the dioxirane reaction with a partially unblocked glycal
without any side reaction.
TLC on silica gel plates (E. Merck Kieselgel 60 F254
)
and detected by spraying with a solution of naphthore-
sorcinol (200 mg) in sulfuric acid (2N, 100 mL) and
ethanol (100 mL) with subsequent heating. Medium-
pressure liquid chromatography (MPLC, 3–5 bars) was
performed on Kieselgel (E. Merck, 230–400 mesh,
0.040–0.063 mm). Anhydrous solvents (CH2Cl2,
MeOH) were prepared by passing solvents (analytical
,
grade, p.a.) through molecular sieves (4 A) under
argon.
4.2. General procedure for epoxidation
The glycal (1 mmol) was dissolved in anhydrous
CH2Cl2 (2 mL) under Ar, and the resulting solution
was cooled to 0 °C. A freshly prepared solution of
dimethyldioxirane in acetone (20 mL) was added, and
the reaction mixture was stirred at 0 °C for 1 h or until
TLC indicated complete consumption of the glycal. The
solution was evaporated with a stream of dry Ar, and
the residue was dried in vacuo to afford the 1,2-anhy-
dro sugar(s) in quantitative yield.
4. Experimental
4.1. Instrumentation and general methods
1H and 13C NMR spectra were recorded with a Bruker
WM 300 (1H, 300.1 MHz; 13C, 75.7 MHz) or with a
Bruker AMX 400 (1H, 400 MHz; 13C, 100.6 MHz).
Two-dimensional correlation spectra (GCOSY and
GHSQC) were measured with the Bruker AMX 400.
Chemical shifts (l) are given in ppm relative to the
signal for internal Me4Si. Optical rotations were mea-
sured at 20 °C at 589 nm (Na) with a Perkin–Elmer 241
polarimeter using a 10-cm, 1-mL cell. Electrospray
ionisation mass spectra (ESIMS) were recorded with a
4.3. 4-O-Acetyl-1,5-anhydro-2,6-dideoxy-3-C-methyl-
arabino-hex-1-enitol (3)
L-
A solution of compound 1 (412 mg, 2.86 mmol) in
anhyd pyridine (15 mL) was treated with Ac2O (325 mL,
3.4 mmol, 1.2 equiv) at room temperature and stirred
overnight. The pyridine was evaporated, and the crude

C. Ernst et al. / Carbohydrate Research 338 (2003) 231–236
235
product was three times codistilled using toluene. The
resulting residue was purified by MPLC (9:1 cyclohex-
ane–EtOAc) to obtain 465 mg (87%) of compound 3
as a white powder: mp 168 °C; [h]2D0 −60.5° (c 1.4,
J5,CH3 6.2 Hz; 13C NMR (MeOD): l 172.96 (CꢀO),
103.68 (C-1), 77.85, 76.52, 73.34, 67.44 (C-2, C-3, C-4,
C-5), 55.91 (OCH3), 21.35 (C-6), 20.04 (Ac), 18.20
(3-CH3); ESIMS (positive-ion): m/z 257 [M+Na]+.
1
Et2O); H NMR (CDCl3): l 6.24 (d, 1H, H-1), 4.93 (d,
1H, H-4), 4.77 (d, 1H, H-2), 3.97 (dq, 1H, H-5), 2.55
(bs, 1H, OH), 2.16 (s, 3H, Ac), 1.31 (s, 3H, 3-CH3),
1.29 (d, 3H, 6-CH3); J1,2 6.1 Hz, J4,5 9.9 Hz, J5,CH3 6.3
Hz; 13C NMR (CDCl3): l 171.35 (CꢀO), 142.56 (C-1),
108.15 (C-2), 79.15 (C-4), 72.11 (C-5), 69.82 (C-3),
24.82 (C-6), 20.89 (Ac), 17.45 (3-CH3); ESIMS (posi-
tive-ion): m/z 209 [M+Na]+.
4.7. 1,5-Anhydro-2,6-dideoxy-3-C-methyl-4-O-methyl-
arabino-hex-1-enitol (7)
L-
L-Olivomycal (3) (600 mg, 4.2 mmol) was dissolved in
anhydrous DMF (30 mL) under Ar, and the solution
was cooled using an ice bath. NaH (167 mg, 4.2
mmol, 1 equiv, 60% dispersion in mineral oil) was
added, and the resulting mixture was stirred for 1.5 h.
The mixture was treated with MeI (259 mL, 4.2 mmol,
1 equiv) and stirred overnight. The solvent was then
evaporated, and the residue was dissolved in CH2Cl2
and extracted with water. The organic layer was dried
over MgSO4, the solvent evaporated, and the crude
residue was purified by MPLC (9:1 cyclohexane–
EtOAc) to give 194 mg (30%) of compound 7 as a
4.4. 4-O-Acetyl-1,5-anhydro-2,6-dideoxy-3-C-methyl-
ribo-hex-1-enitol (4)
L-
A solution of compound 2 (452 mg, 3.14 mmol) in
anhyd pyridine (30 mL) was treated with Ac2O (712
mL, 7.5 mmol, 2.4 equiv) and a catalytic amount of
DMAP under Ar, with cooling in an ice bath. After
stirring the resulting mixture at room temperature for
2 days, the solvent was evaporated, and the crude
product was dried in vacuo. The resulting residue was
purified by MPLC (12:1 cyclohexane–EtOAc) to yield
362 mg (62%) of compound 4 as a white powder: mp:
1
colourless oil. H NMR (CDCl3): l 6.19 (d, 1H, H-1),
4.65 (d, 1H, H-2), 3.81 (dq, 1H, H-5), 3.36 (s, 3H,
OCH3), 3.19 (d, 1H, H-4), 1.37 (d, 3H, 6-CH3), 1.33
(s, 3H, 3-CH3); J1,2 6.0 Hz, J4,5 10.0 Hz, J5,CH3 6.1 Hz;
13C NMR (CDCl3): l 142.56 (C-1), 108.33 (C-2), 86.94
(C-4), 73.45 (C-5), 71.74 (C-3), 61.06 (OCH3), 24.26
(C-6), 17.80 (3-CH3).
1
124 °C; H NMR (C6D6): l 5.44 (dd, 1H, H-1), 5.34
(d, 1H, H-4), 5.28 (d, 1H, H-2), 4.22 (dq, 1H, H-5),
1,70 (s, 3H, Ac), 1.49 (d, 3H, 3-CH3), 1.25 (d, 3H,
6-CH3); J1,2 3.3 Hz, J1,3-CH3 1.3 Hz, J4,5 8.9 Hz, J5,CH3
6.2 Hz; 13C NMR (C6D6): l 170.69 (CꢀO), 125.53
(C-1), 89.74 (C-2), 73.16 (C-4), 72.07 (C-3), 65.95 (C-
5), 20.66 (Ac), 18.57 (C-6), 18.24 (3-CH3); ESIMS
(positive-ion): m/z 209 [M+Na]+.
4.8. 1,2-Anhydro-6-deoxy-3-C-methyl-4-O-methyl-b-L-
mannopyranose (8)
Compound 8 was prepared according to the general
epoxidation protocol. The regio- and stereoselective
product formation was proven by hydrolysis to com-
pound 9.
4.5. 4-O-Acetyl-1,2-anhydro-6-deoxy-3-C-methyl-b-L-
mannopyranose (5)
Compound 5 was prepared according to the general
epoxidation protocol. The regio- and stereoselective
product formation was proven by methanolysis to
compound 6.16
4.9. 6-Deoxy-3-C-methyl-4-O-methyl-a-L-mannopyran-
ose (9)
The formerly prepared 1,2-anhydro sugar 8 (87 mg,
0.5 mmol) was dissolved in water and stirred for 1 h
at room temperature. The solvent was evaporated, and
the residue was purified by MPLC (1:3 cyclohexane–
EtOAc) to yield 58 mg (60%) of compound 9 as a
white powder: mp 156 °C (dec.); [h]2D0 −9.0° (c 1.0,
4.6. Methyl 4-O-acetyl-6-deoxy-3-C-methyl-a-L-
mannopyranoside (6)
The formerly prepared 1,2-anhydro sugar 5 (101 mg,
0.5 mmol) was dissolved in anhyd MeOH (5 mL) and
stirred for 3 h at room temperature. The solvent was
evaporated, and the residue was purified by MPLC
(4:1 cyclohexane–EtOAc) to yield 34 mg (30%) of
compound 6 as a colourless oil. [h]2D0 −75.0° (c 1.0,
1
MeOH); H NMR (MeOD): l 4.76 (d, 1H, H-1), 3.48
(s, 3H, OCH3), 3.33 (d, 1H, H-2), 3.24 (dq, 1H, H-5),
2.99 (d, 1H, H-4), 1.19 (d, 3H, 6-CH3), 1.10 (s, 3H,
3-CH3); J1,2 1.2 Hz, J4,5 9.5 Hz, J5,CH3 6.0 Hz; 13C
NMR (MeOD): l 94.43 (C-1), 87.05 (C-2), 78.43 (C-
4), 72.16 (C-3), 62.18 (C-5), 49.9 (OCH3), 19.12, 19.02
(C-6, 3-CH3); ESIMS (positive-ion): m/z 215 [M+
Na]+.
1
MeOH); H NMR (MeOD): l 4.91 (d, 1H, H-4), 4.63
(d, 1H, H-1), 3.73 (dq, 1H, H-5), 3.48 (d, 1H, H-2),
3.36 (s, 3H, OCH3), 2.09 (s, 3H, Ac), 1.29 (s, 3H,
3-CH3), 1.19 (d, 3H, 6-CH3); J1,2 1.4 Hz, J4,5 9.9 Hz,

236
C. Ernst et al. / Carbohydrate Research 338 (2003) 231–236
4.10. 4-O-Acetyl-1,2-anhydro-6-deoxy-3-C-methyl-b-
altropyranose (10) and 4-O-acetyl-1,2-anhydro-6-deoxy-
3-C-methyl-a- -allopyranose (11)
L
-
Acknowledgements
L
Financial support by the Deutsche Forschungsge-
meinschaft (International Graduate College ‘Template
Directed Chemical Synthesis’, C.E.) is gratefully
acknowledged.
Compound 4 was treated with DMDO according to the
general epoxidation protocol. The configuration of the
formed epimers was proven by methanolysis to com-
pounds 12 and 13.
References
4.11. Methyl 4-O-acetyl-6-deoxy-3-C-methyl-a-
altropyranoside (12) and methyl 4-O-acetyl-6-deoxy-3-
C-methyl-b- -allopyranoside (13)
L-
1. Halcomb, R. L.; Danishefsky, S. J. J. Am. Chem. Soc.
1989, 111, 6661–6666.
L
2. Plante, O. J.; Andrade, R. B.; Seeberger, P. H. Org. Lett.
1999, 1, 211–214.
3. Ernst, C.; Klaffke, W. Tetrahedron Lett. 2001, 42, 2973–
2975.
4. Thiem, J.; Elvers, J. Chem. Ber. 1979, 818–822.
5. Adam, W.; Bialas, J.; Hadjiarapoglou, L. Chem. Ber.
1991, 124, 2377.
6. Murray, R. W.; Singh, M.; Williams, B. L.; Moncrieff, H.
M. J. Org. Chem. 1996, 61, 1830–1841.
7. Adam, W.; Paredes, R.; Smerz, A. K.; Veloza, L. A.
Liebigs Ann./Recueil 1997, 547–551.
8. Freccero, M.; Gandolfi, R.; Sarzi-Amade`, M. Tetra-
hedron 1999, 55, 11309–11330.
9. Abou-Elzahab, M.; Adam, W.; Saha-Mo¨ller, C. R.
Liebigs Ann. Chem. 1991, 445–450.
10. Adam, W.; Smerz, A. K. Tetrahedron 1995, 51, 13039–
13044.
The mixture of 1,2-anhydro sugars 10 and 11 (92 mg,
0.5 mmol) was dissolved in anhydrous MeOH (2 mL)
and stirred for 3 h at room temperature. The solvent
was evaporated, and the residue was purified by MPLC
(5:2 cyclohexane–EtOAc) to yield 61 mg (57%) of
compounds 12 and 1316 as a colourless oil. The two
epimers could not be separated by MPLC. NMR anal-
1
ysis gave a 4:1 ratio of 12:13. Data for 12: H NMR
(CDCl3): l 4.90 (d, 1H, H-4), 4.73 (d, 1H, H-1), 3.97
(dq, 1H, H-5), 3.58 (d, 1H, H-2), 3.46 (s, 3H, OCH3),
2.15 (s, 3H, Ac), 1.21 (s, 3H, 3-CH3), 1.21 (d, 3H,
6-CH3); J1,2 1.4 Hz, J4,5 10.0 Hz, J5,CH3 6.3 Hz. Data
1
for 13: H NMR (CDCl3): l 4.64 (d, 1H, H-4), 4.49 (d,
1H, H-1), 3.96 (dq, 1H, H-5), 3.55 (s, 3H, OCH3), 3.27
(d, 1H, H-2), 2.17 (s, 3H, Ac), 1.26 (s, 3H, 3-CH3), 1.18
(d, 3H, 6-CH3); J1,2 7.8 Hz, J4,5 9.7 Hz, J5,CH3 6.3 Hz.
11. Lii, J. H.; Ma, B.; Allinger, N. L. J. Comp. Chem. 1999,
15, 1593–1603.
12. Becke, A. D. Phys. Re6. A 1988, 38, 3098–3100.
13. Perdew, J. Phys. Re6. B 1986, 33, 8822–8824.
14. Scha¨fer, A.; Huber, C.; Ahlrichs, R. J. Chem. Phys. 1994,
100, 5829–5835.
5. Theoretical methods
15. TURBOMOLE (Vers. 5.3): R. Ahlrichs, M. Ba¨r, H.-P.
Baron, R. Bauernschmitt, S. Bo¨cker, M. Ehring, K.
Eichkorn, S. Elliott, F. Furche, F. Haase, M. Ha¨ser, H.
Horn, C. Huber, U. Huniar, M. Kattannek, C. Ko¨lmel,
All ground-state structures were fully optimised using
the BP12,13 density functional and a valence triple-z
basis set (TZVP14) on all atoms. On oxygen a TZVPP15
basis, which includes more polarisation functions
(2d1f), was applied. The calculations were performed
with the TURBOMOLE 5.315 program suite on a 600
MHz LINUX-PC-Cluster (2GB RAM).
8
M. Kollwitz, K. May, C. Ochsenfeld, H. Ohm, A.
Scha¨fer, U. Schneider, O. Treutler, M. von Arnim, F.
Weigend, P. Weis and H. Weiss, Universita¨t Karlsruhe,
2000.
16. Parker, K. A.; Meschwitz, S. M. Carbohydr. Res. 1988,
172, 319–326.