The Epoxy Diol Route

Epoxidation of 47 using buffered m-CPBA led to a-epoxy lactone 48 in 60% yield. The corresponding P-epoxy isomer was also isolated in 10-20% yield but could be readily separated from a-epoxy isomer 48. The relative stereochemistry of 48 was assigned using nOe experiments. 

It is noteworthy that stereochemical control of the two new stereogenic centers at C4a and Cl2b ultimately stems from the stereochemistry at Cl through the intramolecular Diels-Alder reaction. Hence, an optically pure 46 should lead to an enantioselective preparation of a-epoxy lactone 48, thereby ultimately leading to an enantioselective synthesis of arisugacin A [1], LAH reduction of ( )-a-epoxy lactone 48 led to epoxy diol 49 in 78% 

However, under our conditions20,35,37 as well as a variety of other conditions,53,54 the formal [3 + 3] cycloaddition reaction of the iminium salt 51 with pyrone 12 failed to provide any desired pentacycle 52. This failure prompted us to think that 51 might be sterically too demanding, thereby obstructing the reaction pathway. 

Acetylation of epoxy diol 49 followed by Dess-Martin periodinane [DMP] oxidation afforded epoxyketone 53 in 90% overall yield. Deacetylation of 53 led to the formation of epoxy lactol 54, and PCC oxidation successfully oxidized 54 to give the desired keto-aldehyde 55 in 71% yield. Attempts to go from 49 to 55 directly via a double Swem oxidation were not successful. 

Keto-aldehyde 55 proved to be suitable for constructing the pentacycle 10.55 The iminium salt intermediate 56 was generated from 55 using 0.5-1.0 equiv of piperidinium acetate in the presence of Na2S04 at 80 °C for 1 

---

To avoid the C4a-C5 epoxide issue, a new route to the keto enal "73-a" was pursued [Scheme 15]. Triol 71 could be attained via LAH reduction of epoxy diol 49, and acylation of 71 would lead to the acetate 72 whose stereochemistry was assigned by X-ray analysis. The subsequent TPAP oxidation of 72 followed by deacylation and PCC oxidation led to the keto-enal "73-a" in 73% overall yield. 

Two methods of end-group construction were used in the synthesis of methyl azafrin [methyl 5,6-dihydroxy-5,6-dihydro-10 -apo-/3-caroten-10 -oate (72)]. The first route used the epoxy-intermediate (73) which on treatment with acid gave the 5,6-diol (74). In the second procedure the trimethylcyclohexanone (75) was converted into the hydroxy-ketone (76) which reacted with the Cg fragment (77) to give the acetylenic compound (78). The Wittig salt (79) was made from either (74) or (78) and used in the condensation to form methyl azafrin. 

As an alternative route to the sesquicarane-type sesquiterpenoids, Hortmann and Ong have examined the carbanionic opening of the epoxy-ester (70) derived from perillaldehyde (71). The two products of this reaction are (72) and (73) in the ratio 3 1. Further elaborations of these two compounds have still to be carried out. Plattner and Rapoport have now reported the preparation of both (+)- and (— )-sirenin. This was accomplished by preparative g.l.c. separation of the two diastereoisomeric pairs of ketals derived from the synthetic ketone (74) and D-( —)- and L-( -1- )-butane-2,3-diols. Synthetic procedures for the conversion of racemic (74) into both racemic sirenin and racemic sesquicarene had already been developed. Furthermore, these authors have firmly established by chemical correlation and c.d. studies that naturally-occurring sirenin and sesquicarene have the same absolute stereochemistry, i.e. (75) and (76) respectively. 

Highly selective and easy to direct routes to 1,3-diols are of special importance for synthetic endeavors aiming at macrolide antibiotics and marine natural products in general. Therefore, the repetitive transformation of epoxide 376 into the epoxy-hexol-derivative 377 developed by Mori and Suzuki [122] merits special mention. 

A flexible route was developed. Commencing with racemic epoxy tosylate and mono-protected butane-l,2-diol, the synthesis of the precursor proceeded smoothly through several intermediates (Scheme 83) and finally yielded the desired acyclic precursor 323. When compound 323 was added to a solution of Pd2(dba)3-CHCl3 and DPPE or P(OEt)3 ligand, the nine-membered rings 324 and 325 were formed in 71-88% overall yield. Stereocontrol of all three chiral centers has therefore been accomplished in the key steps. 

Pan and collaborators [87] reported a more conventional approach to aigialomycin D exploiting a Yamaguchi lactonization for macrocyclization. Key features of the route included a SAE and selective opening of epoxy alcohol 222 to install the C5, C6 -diol and two Julia-Kocienski couplings to establish the -configured double... 

---