identified the chlorohydrin product 4f derived from acid-
promoted ring opening of the epoxide as the product. We
reasoned that the trapping agent was preferentially reacting
with the low-valent titanium reagent instead of the anhydro
sugar. As illustrated in Scheme 4, the combination of
Scheme 2
Scheme 4
is useful for disaccharides as well (entry 9). This suggests
that the methodology could also be applied toward oligosac-
charide synthesis. Also, we are interested in utilizing this
chemistry in the construction of glycoproteins.10 Entry 5
shows our initial efforts in this regard employing a dehy-
droalanine11 derivative as a trapping agent. We intend to
further study these reactions to improve both yield and
stereochemical control.
An interesting and potentially useful point is the formation
of a free C-2 hydroxyl group on the sugar. Although the
selective formation of an allyl R-C-glycoside and subsequent
deprotection of the neighboring C-2 benzyl ether has been
described,12 our method allows for the introduction of the
C-glycoside and the concomitant formation of a free C-2
hydroxyl unit in a single transformation. In addition, we can
introduce greater molecular diversity into the C-glycoside
through the use of a variety of trapping agents. The resulting
free alcohol can be manipulated to form disaccharides or
glycosylamines or be orthogonally protected, if desired. As
seen in Scheme 3, we have achieved this orthogonal
titanium(IV) and manganese(II) salts are apparently Lewis
acidic enough to promote the ring opening of the epoxide;
capture by chloride affords 4f.13
These results suggested that the process might be limited
by the reduction potential of the trapping agent. In an effort
to provide a quantitative assessment of this view, we
examined the reduction potential of the electron-deficient
alkenes shown in Table 2. We observed that there is a
definitive window within which the potential must reside in
order to effectively trap the anomeric radical. Acrolein and
â-nitro styrene are preferentially reduced by the titanocene-
(III) complex, preventing the free-radical cleavage of the
epoxide. Unactivated olefins such as styrene and 1-hexene
are apparently not electrophilic enough to react with the
nucleophilic anomeric radical (our attempts at trapping with
these olefins were unsuccessful).14 Thus, we conclude that
if the redox potential falls within the range of ca. -2.7 to
-2.9 (vs Ag/AgNO3) the anomeric radical will be success-
Scheme 3
(8) For examples, see: (a) Evans, D. A.; Trotter, B. W.; Coleman, P. J.;
Cote, B.; Dias, L. C.; Rajapakse, H. A.; Tyler, A. N. Tetrahedron 1999,
55, 8671. (b) Rainier, J. D.; Allwein, S. P.; Cox, J. M. J. Org. Chem. 2001,
66, 1380. Rainier has published a report using boron- and aluminum-
mediated alkylations to form R-C-glycosides: Rainier, J. D.; Cox, J. M.
Org. Lett. 2000, 2, 2707.
(9) In addition to deuterium trapping and NMR data, to our knowledge
all examples of trapping reactions of anomeric radicals leads predominantly,
if not exclusively, to R-configured products.
protection in situ by adding acetic anhydride and triethy-
lamine to the reaction mixture after the initial epoxide
cleavage and trap.
(10) For reviews, see: Dondoni, A.; Marra, A. Chem ReV. 2000, 100,
4395. (b) Taylor, C. M. Tetrahedron 1998, 54, 11317.
(11) Ferreira, P. M. T.; Maia, H. L. S.; Monteiro, L. S.; Sacramento, J.
J. Chem. Soc., Perkin Trans. 1 1999, 24, 3697.
(12) Nicotra, F.; Cipolla, L.; Lay, L. J. Org. Chem. 1997, 62, 6678.
(13) Gansa¨uer has reported that the ZnCl2/Cp2TiCl reagent is capable
of opening epoxides, though the MnCl2/Cp2TiCl system appears to be less
Lewis acidic. See ref 5b.
(14) For a review on anomeric radicals, see: Descotes, G. J. Carbohydr.
Chem. 1988, 7, 1. Also see ref 2.
Of note are attempted trapping reactions with acrolein and
â-nitro styrene (entries 5 and 6, Table 1). In both cases we
(6) Danishefsky, S. J.; Halcomb, R. L. J. Am. Chem. Soc. 1989, 111,
6661.
(7) This representation of 3 may be overly simplified, but it allows one
to see the impact of anomeric stabilization leading to the observed products.
For more on the conformations of these types of radicals, see refs 3 and
14.
(15) Reduction potentials were recorded in freshly distilled and degassed
THF using a glassy carbon working electrode and a 0.01 M Ag/AgNO3 in
acetonitrile reference electrode. This electrode has a potential of ca. 0.3 V
versus SCE.
Org. Lett., Vol. 4, No. 9, 2002
1441