The separate, and combined, infhrences of ether, silyl ether and cyclic ketal groups at positions 3 and 5 on the mode and efficiency of Gheptenoyl radical cyclizations have been studied with a view to the eventual synthesis of 1~@dihydroxyvit D,.
With a view to the eventual synthesis of the important vitamin Da metabolite lo!,25-dihydroxyvitamin D,
(1)' we have undertaken a detailed study of the cyclization of acyl radicals onto alkyne? and alkenes' and into the effect of oxygen based substituents on the efficiency and mode, m or &, of such processes. In this paper we report in full the results of our exploratory work in this area.
Our initial retrosynthetic analysis of ( ) (Scheme 1) reflected that established' by Lythgoe and reexamined5 recently by Baggiolini in which the A ring in the form of phosphme oxide (2) is coupled to the hydroxylated Grundmann/Wiidaus ketone (3)" in a Wadsworth-Homer-Es reaction. Further disconnection led to the enone (4) as primary target molecule which it was hoped could be elaborated to (2) by Wittig or corresponding chemistry. Alternatively coupling of (3) and (4) could conceivably be achieved by means of acetylide chemistry and subsequent reductive elimination with low valent titanium as employed' by Solladit in a recent synthesis of dihydrovitamin 4."
Further examination led to the hypothesis that ketone (4) might readily be prepared by cyclization of an acyl radical onto an alkyne as indicated in scheme 2. Moreover it was considered that the acid (5), a precursor of any appropriate acyl radical source would be available in a minimum of steps from 3-hydmxyglutsric acid via the known" anhydride (6) and chemicallo*" or enzymic" resolution.
When this investigation was begun, and despite the popularity of radicals in organic synthesis, there existed
only isolated examples of acyl radical cyclizations onto alkenes" and of alkoxycarbonyl radicals onto alkenes" and alkynes". However several studies contemporary with the work described here have now been published".
With these aims and background in mind we began our investigation by preparation of a model alkynoic acid (9) by reaction of bis(trimethylsilyl)acetylene with the acid chloride derived from the monomethyl ester of glutaric acid in the presence of tin IV chloride at -78 "C in dichloromethane, essentially according to Nicolaou", giving ester (7) in 75% yield. Ketalisation procc&d smoothly to give (8) in 97% yield and saponification with concomitant protcdesilylation'9 furnished ( 9) in 78% yield. An alternative shorter approach, based on the knowt?' aluminium trichloride promoted tea&on of phthalic anhydride with bis(trimethylsilyl)butadiyne, involving heating glutaric anhydri& with bis(trimethylsilyl)acetylene and tin IV chloride in dichloromethane was sluggish and provided only minor quantities of a dihydropymnone (15).
6581