THIOGLYCOSIDES: THE 0-MANNOSYLATION REACTION
Several years ago in Chicago, we were engaged in developing a solution to the well-known ^-mannoside problem  involving the inversion of the much more readily accessible a-mannosides by a sequence of hydrogen atom abstraction, radical inversion, and diastereoselective quenching (Scheme 5) [20,21].
With the fundamental chemistry, which was developed using commercially available a-methyl mannopyranoside as substrate in hand, it became necessary to prepare a genuine a-disaccharide for inversion. We selected Kahne's excellent sulfoxide method [22- 24] for numerous reasons, which included the reported excellent yields for coupling to extremely hindered alcohols at low temperatures, the absence of any metal salt as promoter, and the implication that the coupling proceeded through quenching of the oxacarbenium ion, hence should provide the a-mannoside. Other than a brief footnote in their original communication noting that the anomeric stereoselectivity was a function of the stereochemistry at C2 of the sulfoxide donor , Kahne and his coworkers had not described the application of their method to the mannose series. Jarmila Brunckova thus prepared S-ethyl 4,6-benzylidene-a-d-thiomannopyranoside by standard means and converted it to the 3-0-benzyl-2-0-ferf-butyldimethylsilyl derivative (14) by the aegis of dibutyltin oxide and benzyl bromide, then TBDMS triflate. Oxidation with mCPBA then gave the sulfoxide (15) required for coupling . Brunckova noted the highly selective sulfoxidation process, which gave essentially a single diastereomer, in contrast to the unselective oxidations ^-thioglycosides observed previously , but was unable to assign configuration at the time. In fact it was several years before Jan Mataka and Sanxing Sun were able to prepare crystalline derivatives suitable for X-ray analysis and so assign the configuration as SR . Brunckova then mixed the sulfoxide with methyl
2,3,4-tri-O-acetyl-a-d-glucopyranoside (16) and 2,6-di-ferf-butyl-4-methylpyridine (DTBMP), a hindered base, in diethyl ether at — 78°C and activated the sulfoxide by dropwise addition of triflic anhydride. In line with our expectations, a good yield of a 10:1 mixture of glycosides (17) favoring the a anomer was obtained (Scheme 6) . The TBDMS protecting group was removed and was replaced by the radical precursor, and the radical inversion procedure was conducted with moderate success, comparable to that seen with the a-methyl mannoside .
Sanxing Sun, a new student, sought to prepare more of the a-disaccharide (17) and subsequently to improve the radical inversion process. He repeated Brunckova's preparation with the minor, but fortuitous, difference that the sulfoxide was activated with triflic anhydride before addition of the acceptor alcohol. To our amazement, an excellent yield was obtained of an anomeric mixture favoring the ^-mannoside by a factor of roughly 10 (Scheme 7) [26,27].
Consideration of this unanticipated reversal of diastereoselectivity led us to propose a mechanism in which, under all conditions, the sulfoxide 15 (Scheme 8) undergoes rapid sulfonylation leading to a sulfonium ion 18. We then postulated that this sulfonium ion collapses to the oxacarbenium ion 19 which, when generated in
Was this article helpful?