The Fürstner Synthesis of Amphidinolide F
The amphidinolides, having zero, one or (as exemplified by Amphidinolide F ) two tetrahydrofuran rings, have shown interesting antineoplastic activity. It is a tribute to his development of robust Mo catalysts for alkyne metathesis that Alois Fürstner of the Max-Planck-Institut Mülheim could with confidence design (Angew. Chem. Int. Ed. 2013, 52, 9534. ) a route to 3 that relied on the ring-closing metathesis of 1 to 2 very late in the synthesis.
Three components were prepared for the assembly of 1. Julia had already reported (J. Organomet. Chem. 1989, 379, 201. ) the preparation of the E bromodiene 5 from the sulfone 4. The alcohol 7 was available by the opening of the enantiomerically-pure epoxide 6 with propynyl lithium, followed by oxidation following the Pagenkopf protocol. Amino alcohol-directed addition of the organozinc derived from 5 to the aldehyde from oxidation of 7 completed the assembly of 8.
Addition of the enantiomer 10 of the Marshall butynyl reagent to 9 followed by protection, oxidation and addition of, conveniently, the other Marshall enantiomer 11 led to the protected diol 13. Silylcupration-methylation of the free alkyne set the stage for selective desilylation and methylation of the other alkyne. Iodination then completed the trisubstituted alkene of 14.
Methylation of the crystalline lactone 15, readily prepared from D-glutamic acid, led to a mixture of diastereomers. Deprotonation of that product followed by an aqueous quench delivered 16. Reduction followed by reaction with the phosphorane 17 gave the unsaturated ester, that cyclized with TBAF to the crystalline 18. The last stereogenic center of 22 was established by proline-mediated aldol reaction of the aldehyde 19 with the ketone 20.
To assemble the three fragments, the ketone of 21 was converted to the enol triflate and thence to the alkenyl stannane. Saponifaction gave the free acid 22, that was activated, then esterified with the alcohol 18. Coupling of the stannane with the iodide 14 followed by removal of the TES group led to the desired diyne 1. It is noteworthy that the Mo metathesis catalyst is stable enough to tolerate the free alcohol of 1 in the cyclization to 2.
The plan from the beginning had been to selectively hydrate the alkyne of 2 by Pt-mediated intramolecular addition of the free alcohol. Inspection of models had led to the decision that the R diastereomer would cyclize more readily than the S, so that was the one prepared. In the event, cyclization followed by hydrolysis of the intermediate enol ether gave the desired ketone 23, that was carried on to Amphidinolide F (3).
It is instructive to compare the work described here with the complementary synthesis of 3 recently reported (Angew. Chem. Int. Ed. 2012, 51, 7948, ; J. Am. Chem. Soc. 2013, 135, 10792, ) by Rich G. Carter of Oregon State University. In both approaches, enantiomerically-pure subunits were assembled to make 3.