1306
Serebryany and Beigelman
Derivatives of 1,2-dideoxyribose can be synthesized from 2-deoxy-D-ribose by a
number of methods, however the high cost and poor availability of 2-deoxyribose
restricts the use of these approaches. We decided to employ thymidine as a starting
material in the preparation of the target compound because it is relatively inexpen-
sive and readily available in bulk quantities due to production of azidothymidine.
It has been shown that thymidine and its 50-O-silylated derivatives undergo elim-
ination of thymine base upon treatment with ammonium sulfate in hexamethyldi-
silazane (HMDS) under reflux to give cyclic enol ethers, related to dihydrofuran
and so called furanoid glycals.[2] Apparently, the double bond in the glycal molecule
can be hydrogenated to provide the desired 1,2-dideoxyribose structure. Our syn-
thetic approach is shown in the Sch. 1.
In our initial experiments 50-O-(tert-butyldimethylsilyl) thymidine 2 was reacted
with 0.2–0.5 eq. (NH4)2SO4 in HMDS under reflux as described[2] to furnish unstable
glycal 3. We observed the yield of this reaction to be variable and highly dependent
on reaction conditions and scale. When reactions were performed on more than
10–15 g scale, the yield dropped and decomposition of 3 was detected. The likely rea-
son is instability of 3 in the presence of acidic ammonium sulfate. This prompted us
to investigate other compounds that can potentially be used as catalysts for this
reaction.
We found that treatment of 2 with trimethylsilyl chloride, trifluoroacetic acid
and tin tetrachloride in HMDS did not afford glycal 3, resulting only in 30-O-silyla-
tion of starting compound 2. Sulfuric acid, p-toluenesulfonic acid, methanesulfonic
acid and trimethylsilyl methanesulfonate gave results similar to those obtained using
ammonium sulfate. Trimethylsilyl triflate and triflic acid furnished fast elimination
of nucleobase but gave rise to a complex mixture of products. On the other hand,
we have found that treatment of 2 with methanesulfonamide along with a small
amount of acidic catalyst, such as methanesulfonic acid furnished fast and clean con-
version of 2 to 3. Under optimized conditions (0.05 mol. eq. of MeSO3H and 0.5 mol.
eq. of MeSO2NH2) the desired compound 3 was obtained in more than 80% yield.
No signs of decomposition of 3 during the reaction were detected. It is worth noting,
that increasing the reaction time or scale did not result in reduction of the yield.
Scheme 1. Reagents and conditions: i) TBDMS-Cl, Py, 0ꢀC, 93%; ii) HMDS, MeSO3H
(0.05 eq), then MeSO2NH2 (0.5 eq), reflux, 2 h; iii) H2, Pd=C, 1 h; iv) PyꢁTFA (0.05 eq),
MeOH, 30 min 80% from 2; v) DMT-Cl, Py, DMAP, 87%; vi) NaOH, EtOH-H2O, reflux;
vii) succinic anhydride, Py, DMAP, then Et3N, 82% overall from 6.