Catalytic oxidation-phosphorylation of glycals: Rate acceleration and enhancement of selectivity with added nitrogen ligands in common organic solvents

Article abstract of DOI:10.1021/ol047417k

(Chemical Equation Presented) The first general catalytic oxidation of glycals has been developed to afford useful glycosyl phosphates in high yield and selectivity in a domino process with the use of catalytic methyltrioxorhenium, urea hydrogen peroxide, stoichiometric dibutyl phosphate as nucleophile, and a substoichiometric amount of nitrogen ligands, such as pyridine or imidazole, in organic solvents.

Full text of DOI:10.1021/ol047417k

ORGANIC
LETTERS
2005
Vol. 7, No. 4
725-728
Catalytic Oxidation−Phosphorylation of
Glycals: Rate Acceleration and
Enhancement of Selectivity with Added
Nitrogen Ligands in Common Organic
Solvents
Gianluca Soldaini, Francesca Cardona, and Andrea Goti*
Dipartimento di Chimica Organica “Ugo Schiff”, Polo Scientifico,
UniVersita` di Firenze, ICCOM C.N.R., Via della Lastruccia 13,
I-50019 Sesto Fiorentino (Firenze), Italy
andrea.goti@unifi.it
Received December 16, 2004
ABSTRACT
The first general catalytic oxidation of glycals has been developed to afford useful glycosyl phosphates in high yield and selectivity in a
domino process with the use of catalytic methyltrioxorhenium, urea hydrogen peroxide, stoichiometric dibutyl phosphate as nucleophile, and
a substoichiometric amount of nitrogen ligands, such as pyridine or imidazole, in organic solvents.
Glycals 1 are very useful and versatile substrates for the
synthesis of glycoconjugates and oligosaccharides.1 Their
conversion involves an initial epoxidation to give the
corresponding 1,2-dehydrosugar derivatives 2, used as gly-
cosyl donors with appropriate acceptors or as precursors of
more stable glycosyl donors, which can be used succes-
sively.2
nected with its potential explosiveness. Therefore, the
development of a practical alternative for this transformation
is highly desirable, especially for large-scale syntheses. In
particular, no general catalytic method had been reported for
the epoxidation of glycals before we recently disclosed that
methyltrioxorhenium (MTO)4 is able to catalyze this oxida-
tion in methanol, with the complex urea hydrogen peroxide
(UHP)5 as the stoichiometric oxidant, giving the correspond-
ing methyl glycosides 3.6,7 In the presence of dibutyl
phosphate (DBP) in ionic liquids, the oxidation furnished,
(1) Danishefsky, S. J.; Bilodeau, M. T. Angew. Chem., Int. Ed. Engl.
1996, 35, 1380-1419.
(2) Gordon, D. M.; Danishefsky, S. J. Carbohydr. Res. 1990, 206, 361-
366.
(3) Halcomb, R. L.; Danishefsky, S. J. J. Am. Chem. Soc. 1989, 111,
6661-6666.
(4) (a) Beattie, G.; Jones, P. J. Inorg. Chem. 1979, 18, 2318-2319. (b)
Herrmann, W. A.; Kratzer, R. M.; Espenson, J. H.; Wang, W. Inorg. Synth.
2002, 33, 110-112.
Dimethyldioxirane (DMDO) is practically the only reagent
used for the epoxidation step.3 DMDO is unstable, has to be
freshly prepared, and poses serious safety problems con-
(5) Heaney, H. Top. Curr. Chem. 1993, 164, 1-19.
10.1021/ol047417k CCC: $30.25
© 2005 American Chemical Society
Published on Web 01/15/2005

with moderate chemo- and stereoselectivities, the corre-
sponding glycosyl phosphates 4,6 which are extremely useful
and convenient glycosyl donors.8 Novel methods for their
preparation are very timely, due to their extensive use in
the first automated oligosaccharide syntheses recently de-
veloped by Seeberger.9
In this paper, we report that the use of nitrogen ligands in
this reaction gives tremendous beneficial effects in terms of
conversion of substrate, rate acceleration, product selectivity,
and diastereoselectivity of the epoxidation and allows
standard organic solvents to be used. This results in a novel
practical and convenient method for the synthesis of glycosyl
phosphates.
Although MTO/H2O2 (or UHP) has been used for the
epoxidation of a large variety of differently substituted
alkenes,10 its use with enol ethers is barely documented, while
a properly modified procedure has been reported for oxida-
tion of silyl enol ethers.11 Glycals proved quite sluggish
substrates toward MTO/UHP. When an up to 3-fold excess
of DBP was used in the domino epoxidation-phosphoryl-
ation, the conversion of tribenzylglucal (5) chosen as a model
substrate was incomplete in THF, CH2Cl2, CH3CN, and other
organic solvents. Dimethylimidazolium tetrafluoroborate
([BMIM]BF4) is necessary in order to force the oxidation to
completion. Ionic liquids are difficult to obtain with repro-
ducible purity and especially completely anhydrous; there-
fore, the competing ring-opening by water was always a
serious drawback and acceptable yields of glycosyl phos-
phates 6 and 7 were obtained only with a large excess of
DBP (g5 equiv). With such an excess of DBP, however,
the reaction went to completion also in common organic
solvents, affording better yields of glycosyl phosphates
(Table 1, entries 3 vs 2 and 7 vs 5). Moreover, the selectivity
of the reaction turned out to be quite solvent dependent, the
higher gluco/manno selectivity occurring with the more
coordinating solvent THF (Table 1, entry 7).
It is worth noting that, while 7 was obtained as the
R-anomer only, phosphate 6 was always produced as a
mixture of R- and â-anomers. However, this is not a critical
point, since both anomers participate in subsequent glycosyl
transfer affording the same product.12 Moreover, R-glycosyl
phosphates can be formed from the corresponding â-isomers
by temperature-dependent acid-catalyzed anomerization,
promoted by the excess of DBP in the reaction mixture, as
already shown by previous work.13 Therefore, anomeric
equilibration affords R/â ratios which are variable with
temperature and time, with the more stable R-glucosyl
phosphate prevailing at longer reaction time and higher
temperature. In our case, for instance, when the reaction was
performed at 0 °C (Table 1, entry 4), R-glucosyl phosphate
R-6 and â-glucosyl phosphate â-6 were obtained in 1:1.5
ratio. The same reaction afforded a 4:1 ratio in favor of the
R-anomer when carried out at 25 °C (Table 1, entry 3). The
R/â ratios found for each reaction are reported in the
Supporting Information.
The effect of DBP on the conversion suggested that DBP
might work as a donor ligand toward Re in the catalyst
precursor and/or catalytic species, thus influencing the
turnover rate and/or catalyst lifetime. We then decided to
test other better ligands for Re in this reaction, and
particularly nitrogen ligands, such as pyridine (Py), pyrazole
and 3-cyanopiridine, which have already shown beneficial
effects in epoxidation of alkenes, enhancing the reaction rate
and limiting the hydrolysis of epoxides to diols.10 Moreover,
the use of ligands should limit the DBP to be used to a
stoichiometric amount. Then, 1.2 equiv of DBP as nucleo-
phile and 0.5 equiv of the ligands reported in Table 2 have
been used in the model reaction in CH2Cl2 or THF.
The results in Table 2 (see the Supporting Information
for results with a larger number of ligands) demonstrate the
effectiveness of nitrogen ligands in promoting the reaction
with stoichiometric amounts of DBP. Oxygen ligands, such
as N-oxides, gave poorer results (entries 13 and 14). In the
absence of ligand, the reaction does not go to completion in
CH2Cl2 or in THF (Table 2, entries 1 and 2). With added
ligands, complete conversion is attained. Some of them are
particularly remarkable in terms of conversion of substrate,
rate acceleration, and facial diastereoselectivity of epoxida-
tion, which determines the selectivity at C-2 (gluco/manno).14
Table 1. “One-Pot” Synthesis of Glycosyl Phosphatesa
(6) Soldaini, G.; Cardona, F.; Goti, A. Tetrahedron Lett. 2003, 44, 5589-
5592.
DBP
(equiv)
time
(h)
conv (%)
(yield, %)b gluco/manno
epox select.c
(7) For a different example of one-pot oxidation-nucleophile addition,
see: Di Bussolo, V.; Kim, Y.-J.; Gin, D. Y. J. Am. Chem. Soc. 1998, 120,
13515-13516.
entry
solvent
1d
2
3d
4
5
6d
7
[BMIM]BF4
CH2Cl2
CH2Cl2
CH2Cl2
THF
THF
THF
CH3CN
5
3
5
5
3
5
5
5
3.5
5
100 (58)
80
5.5:1
(8) Palmacci, E. R.; Plante, O. J.; Seeberger, P. H. Eur. J. Org. Chem.
2002, 595-606.
1.5
6.5
4.5
1.5
7
100 (75)
100 (74)
60
3.8:1
2.7:1
(9) Plante, O. J.; Palmacci, E. R.; Seeberger, P. H. Science 2001, 291,
1523-1527.
(10) Adolfsson, H. In Modern Oxidation Methods; Ba¨ckvall, J.-E., Ed.;
Wiley-VCH: Weinheim, 2004; pp 32-43 and references therein.
(11) Stankovic, S.; Espenson, J. H. J. Org. Chem. 1998, 63, 4129-4130.
(12) Plante, O. J.; Andrade, R. B.; Seeberger, P. H. Org. Lett. 1999, 1,
211-214.
60
100 (65)
100
9.8:1
3.8:1
8
7
(13) Plante, O. J.; Palmacci, E. R.; Andrade, R. B.; Seeberger, P. H. J.
Am. Chem. Soc. 2001, 123, 9545-9554.
(14) The found selectivities and reaction rates have no apparent relation-
ship with pKa values of the ligands, nor with equilibrium constants for the
coordination of used ligands with MTO, recently reported: Nabavizadeh,
S. M. Inorg. Chem. 2003, 42, 4204-4208.
a Conditions: (a) MTO (4 mol %), UHP (3 equiv), dry solvent, DBP, 0
°C; (b) Py, Ac2O, rt, overnight. b Isolated yields. c Calculated by integration
of the 31P NMR spectra of the crude mixtures. d Reaction performed at room
temperature.
726
Org. Lett., Vol. 7, No. 4, 2005

Py and ImH gave the best results, leading to complete
conversions and good diastereoselectivities in both solvents
used, and were selected for further experiments. The reaction
is faster in CH2Cl2 (Figure 1a) than in THF (Figure 1b). The
two ligands display a considerable accelerating effect in
CH2Cl2, with Py being most effective and driving the reaction
to completion in only 10 min (Figure 1a). Notably, under
these conditions a reduced amount of MTO (1 mol %) was
sufficient to give a complete conversion of substrate in 1.5
h. Conversely, in THF the ligands did not show a similar
effect on rate, with ImH giving an opposite decelerating
effect (Figure 1b). Nevertheless, their use was beneficial to
catalyst lifetime, allowing the reaction to go to completion.
The best result in terms of C-2 selectivity (gluco/manno)
was obtained with ImH (up to 14:1 with 0.5 equiv of ImH
in THF (Table 2, entry 10).
Table 2. “One-Pot” Synthesis of Glycosyl Phosphates 6 and 7
from Tribenzylglucal (5) in the Presence of Nitrogen Ligandsa
epox
select.c
ligand
(0.5 equiv)
time conv (%) gluco/
solvent (h) (yield,%)b manno
entry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
CH2Cl2 2.0
THF 1.5
CH2Cl2 0.2
THF 1.5
CH2Cl2 1.0
THF 1.5
CH2Cl2 0.7
THF 2.0
CH2Cl2 1.0
THF 1.5
50
81
pyridine
pyridine
3-cyanopyridine
3-cyanopyridine
pyrazole
pyrazole
imidazole
imidazole
100 (76)
100 (68) 10.5:1
100
55
100
90
100 (75)
100 (69) 14.0:1
100
100
100
81
7.0:1
4.1:1
7.8:1
2.9:1
11.1:1
9.0:1
The effect of the amount of added imidazole on selectivity
in CH2Cl2 and THF was also investigated (Figure 2). This
1,10-phenanthroline CH2Cl2 1.2
1,10-phenanthroline THF 1.0
CH2Cl2 2.0
THF 2.0
4.7:1
8.3:1
4.7:1
9.4:1
bipyridine N-oxide
bipyridine N-oxide
a Conditions: (a) MTO (4 mol %), UHP (3 equiv), dry solvent, DBP
(1.2 equiv), ligand (0.5 equiv); (b) Py, Ac2O, rt, overnight. b Isolated yields.
c Calculated by integration of the 31P NMR spectra of the crude mixtures.
Contrary to most of the ligands used, imidazole (ImH)
has no precedent in the literature in MTO-catalyzed epoxi-
dations. More importantly, the strong ligand and solvent
effect on the diastereoselectivity of the reaction is com-
pletely unprecedented.15 As a rule, epoxidation selectivity
is always better in THF than in CH2Cl2 with any ligand
(Table 2).
Figure 2. Influence of the amount of imidazole on the facial (gluco/
manno) selectivity in the MTO-catalyzed oxidation-phosphoryl-
ation of tribenzylglucal (5) in CH2Cl2 and THF (other conditions
as in Table 2).
allowed establishment of an optimal amount of 0.8 equiv of
ImH to increase the facial selectivity in THF up to a
synthetically meaningful 16.7:1 ratio. It is worthy to note
that all the isomeric ratios have been evaluated by integration
of the singlet signals from 31P NMR spectra of the crude
reaction mixtures. This method allows an extremely accurate
and reliable determination, even of very minor isomers which
might elude different techniques.
This methodology is of general scope, as proved by the
synthesis of glycosyl phosphates 8-14, in good to excellent
yields and up to complete stereoselectivity, from the corre-
sponding glycals derived from D-glucose, D-galactose, L-
rhamnose, and D-arabinose, either protected with benzyl or
acetyl groups (Figure 3). The nature of the protecting groups
influences strongly the rate of reaction. Substrates with
electron withdrawing acetyl groups were oxidized much more
sluggishly and generally were not completely converted in
THF. Use of dichloromethane and pyridine as additive
proved the best conditions for such substrates. Pyridine
ensured shorter reaction time compared to imidazole, without
Figure 1. Conversion of tribenzylglucal (5) in the MTO-catalyzed
oxidation-phosphorylation in (a) CH2Cl2 and (b) THF (conditions
as in Table 2).
(15) For an example of a solvent-dependent diastereoselective oxidation
of an allylic alcohol with free OH, see: Adam, W.; Mitchell, C. M.; Saha-
Mo¨ller, C. R. Eur. J. Org. Chem. 1999, 785-790.
Org. Lett., Vol. 7, No. 4, 2005
727

observed that parallel those reported above for tribenzyl-
glucal. Thus, the selectivity at C-2 could be increased to a
10:1 ratio when the domino process was carried out in THF
in the presence of imidazole (see the Supporting Informa-
tion).
In summary, the first general catalytic oxidation of glycals
has been developed to afford useful glycosyl phosphates in
high yield and selectivity in a domino process with stoichio-
metric DBP and substoichiometric Py or ImH in organic
solvents.
Acknowledgment. We thank MIUR, Italy, for financial
support (PRIN 2002) and Ente Cassa di Risparmio di Firenze,
Italy, for granting a 400 MHz NMR spectrometer. Technical
assistance from Brunella Innocenti and Maurizio Passaponti
is gratefully acknowledged.
Figure 3. Reagents and conditions: glycal, MTO (4 mol %), UHP
(3 equiv), DBP (1.2 equiv), pyridine (0.5 equiv), CH2Cl2, then
pyridine, Ac2O, rt, 15 h. Yields, C-2 selectivities, and reaction
times: 8, 85%, 3.3:1, 6.5 h; 9, 69%, >50:1, 1.5 h; 10, 83%, 39:1,
17.5 h; 11, 70%, 6:1, 0.5 h (63%, 9.9:1, 1 h in THF with 0.8 equiv
of imidazole); 12, 85%, 3.3:1, 6 h; 13, 70%, >50:1, 2 h; 14, 85%,
>50:1, 5 h.
Supporting Information Available: Experimental pro-
cedures, complete data on selectivities, characterization data,
and NMR spectra of compounds 6-14 and of minor
diastereoisomers. This material is available free of charge
via the Internet at http://pubs.acs.org.
affecting the selectivity to a substantial extent. For tribenzyl-
L-rhamnal, precursor to 11, effects of ligands have been
OL047417K
728
Org. Lett., Vol. 7, No. 4, 2005