phenylselenium ions.6 Dominant in the approaches used to
date for the synthesis of N-glycosides has been the fact that
the base is always introduced on the cyclic sugar. Acyclic
N-glycosides have even been prepared from cyclic inter-
mediates through ring cleavage of the sugar.7
Lacking in literature have been strategies involving a base
coupling step that precedes a cyclization as illustrated in
Scheme 2. A rare example of such an approach is Liotta’s
speaking, an electron-withdrawing group should be present
on the nitrogen to render the nitrogen electron deficient,
making more energy demanding the creation of an iminium
ion intermediate in a competitive SN1-like process. We
hypothesized that silylated purines and pyrimidines would
be judicious choices of amines and that an intramolecular
SN2 displacement of the sulfur on the thioaminyl by the
hydroxyl group at C4 should lead to N-glycosides (Scheme
2b).9 The preliminary results of our study are depicted herein.
Dithioacetal 1 was used to evaluate the reactivity of
different thiol activating reagents in the silylated base
coupling step. Hg(OAc)2/TMSOTf 10 and dimethyl(meth-
ylthio)sulfonium tetrafluoroborate (Me2S(SMe)BF4)11 have
both proven to be successful conditions as seen in Table 1.
Promising ratios of 4:1 and 13:1 were obtained in favor of
syn product 2a when silylated uracil was used (Table 1,
entries 1 and 2). Replacing the methyl with a silyl ether
(OTBS) resulted in a significant increase in diastereoselec-
tivity (cf. entries 1 and 3 as well as entries 2 and 4, Table
1).12 Similarly, excellent diastereoselectivity and yield were
obtained when thioacetal 3 was treated with either thymine
(Table 1, entries 5 and 6) or cytosine (Table 1, entries 7 and
8). The silylated adenine was coupled effectively with
dithioacetal 3 using Hg(OAc)2/TMSOTf (Table 1, entry 9).13
Scheme 2
Substrates bearing a primary or secondary alkoxy group
at C4 were also considered. With the exception of 12 (Table
1, entry 12),14,15 all substrates tested gave excellent results
(Table 1, entries 10, 11, 13, and 14).16
synthesis of AZT (R1 ) OBz, R3 ) N3, R4 ) H, X)OBn),
wherein an SN1 pathway is responsible for the anomeric
stereochemistry (Scheme 2a).8 The challenging strategy we
propose (Scheme 2b) is based on a diastereoselective
introduction of the silylated base to an acyclic dithioacetal
intermediate and a subsequent SN2-like intramolecular dis-
placement.
A few years ago, we became interested in inducing on
acyclic molecules stereogenic centers bearing a C-N bond.
One of the approaches we have developed in this context is
based on a diastereoselective transformation of thioacetals.
The first step of the approach to be depicted herein involves
the formation of a thionium ion intermediate to which an
amine is added in order to generate the thioaminyl acetal
(Scheme 3).
A couple of conclusions can be drawn at this point. The
use of Hg(OAc)2/TMSOTf and Me2S(SMe)BF4 in the first
step was effective in selectively activating the sulfur.
Additionally, the presence of an alkoxy group (OR1, 1,2-
induction, Scheme 3) on the stereogenic center R to the
dithioacetal17 during addition of the silylated base was
successful in promoting excellent diastereoselectivity. Two
(9) The synthesis of disaccharide proposed by Hindsgaul bears a similarity
to our approach in that cyclization occurs as the final step: McAuliffe, J.
C.; Hindsgaul, O. J. Org. Chem. 1997, 62, 1234.
(10) Bra˚nalt, J.; Kvarnstro¨m, I.; Niklasson, G.; Svensson, S. C. T.;
Classon, B.; Samuelsson, B. J. Org. Chem. 1994, 59, 1783.
(11) Trost, B. M.; Murayama, E. J. Am. Chem. Soc. 1981, 103, 6529.
(12) Proof of the structure was obtained through crystallization of 4a
followed by X-ray analysis. See Supporting Information.
(13) The coupling of this base was less successful in the presence of
Me2S(SMe)BF4, giving complex mixtures of thioaminyl acetals. Similar
results were obtained with guanine in the presence of either thiol activating
reagent, an issue that will have to be revisited with modified purine bases.
Regioselectivity in the synthesis of purine nucleosides is discussed in ref
1a. Additional approaches such as the blocking of nitrogen functions were
not considered in this study.
Scheme 3
(14) Bernardi et al. have shown that allylmagnesium bromide addition
to 2,3-syn- and 2,3-anti-dialkoxy aldehydes also takes place with different
degrees of stereocontrol, the syn substrate giving the highest ratio: Bernardi,
R.; Fuganti, C.; Grasselli, P. Tetrahedron Lett. 1981, 22, 4021.
(15) A mixture of SEt/SMe compounds was obtained from 12. A control
experiment involving cyclization of the corresponding isolated SEt and SMe
products in the presence of Me2S(SMe)BF4 gave products 23a:23b with
similar ratios, indicating that the SMe/SEt exchange does not occur after
the introduction of thymine and is not responsible for the erosion of
diastereoselectivity.
(16) The 1,2-anti relative stereochemistry of 9b was confirmed by X-ray
analysis. See Supporting Information.
(17) For more on the use of dithioacetals in diastereoselective processes,
see: (a) Shibata, N.; Fujita, S.; Gyoten, M.; Matsumoto, K.; Kita, Y.
Tetrahedron Lett. 1995, 36, 109. (b) Mori, I.; Bartlett, P. A.; Heathcock,
C. H. J. Org. Chem. 1990, 55, 5966.
Diastereoselectivity is induced through the transfer of
stereochemical information originating from the stereogenic
center adjacent to the thioacetal (1,2-induction). The resulting
thioaminyl acetal is further transformed through an SN2-type
process to preserve chirality. For the latter to occur, tactically
(6) (a) D´ıaz, Y.; El-Laghdach, A.; Matheu, M. I.; Castillo´n, S. J. Org.
Chem. 1997, 62, 1501. (b) D´ıaz, Y.; El-Laghdach, A.; Castillo´n, S.
Tetrahedron 1997, 53, 10921.
(7) Elkattan, Y.; Gosselin, G.; Imbach, J. L. J. Chem. Soc., Perkin Trans.
1 1994, 10, 1289.
(8) Hager, M. W.; Liotta, D. C. J. Am. Chem. Soc. 1991, 113, 5117.
242
Org. Lett., Vol. 4, No. 2, 2002