Efficient Synthesis of 2(S)-[1(S)-Azido-2-phenylethyl]oxirane

Abstract

Various HIV protease inhibitors have been developed as promising chemotherapeutic agents for the treatment of AIDS. The importance of hydroxyethylamine isosteres in the this field was well witnessed by antiviral agents such as saquinavir, amprenavir and nelfinavir. In connection of our studies on the development of an improved drug delivery system, we were interested in the synthesis of amprenavir and related compounds. Among many reports on the synthesis of this class of inhibitors, we focused on the synthetic scheme that relies on the use of protected aminoalkyl epoxides. On the way to the synthesis of crucial 2(S)-[1(S)-azido-2-phenylethyl]oxirane (1) according to the known procedure, we needed to devise a more efficient route to this key intermediate to avoid tedious experimental conditions. Here we wish to report a brief synthesis of oxirane 1 which is more efficient and suitable for preparation on a large scale. In Scheme 1, our modified synthesis starts from acetonide 2 which was readily and cheaply available in large quantity from D-isoascorbic acid. Mesylation and subsequent reduction with NaBH4 converted acetonide 2 to a selectively mesylated diol. Sequential treatment of NaH with the mesylate and addition of PhMgBr to the resulting oxirane in the presence of CuI produced the known alcohol 3 in 80% overall yield. Next, alcohol 3 was mesylated and hydrolyzed to the diol which was then converted to the known diolazide 4 via azide displacement with CsN3. This azidation was superior to the known Mitsunobu azidation of alcohol 3. The resulting diolazide 4 was treated with 2-acetoxyisobutyryl chloride and NaOMe successively to give the known key intermediate 1 in 60% overall yield from alcohol 3. In conclusion, 2(S)-[1(S)-azido-2-phenylethyl]oxirane (1) was efficiently synthesized from acetonide 2 as a starting material by 9 steps with two purification steps at alcohol 3 and oxirane 1 respectively in overall yield of 48%. This protocol avoids troublesome LAH reduction and Mitsunobu azidation that were involved in the known method, providing an expedient route to the facile, practical and large production of the desired epoxide 1 and other related structural motives from the cheaper D-isoascorbic acid.