Part I. Synthesis of Garner's Aldehyde and Efforts Toward the Synthesis of Tapentadol via an Asymmetric Hydroformylation/Reductive Amination Sequence Part II. Development of a Rhodium-Mediated Domino Annulation and Efforts Toward the Total Synthesis of Linderagalactone C
A novel, catalytic synthesis of either enantiomer of the widely used synthetic building block Garner's aldehyde from a single alkene was developed. Rhodium-catalyzed asymmetric hydroformylation (AHF) with the bis(diazaphospholane) (BDP) ligands developed by Landis and coworkers affords each Garner's aldehyde enantiomer in high yield and enantiomeric purity in an atom economical, regio- and facially-selective alkene hydroformylation. This is the first reported AHF with a 1,2-disubstituted alkene with a different heteroatom on each carbon. Three AHF-based synthetic strategies toward the analgesic tapentadol are described. A styrene AHF strategy was abandoned due to poor regioselectivity producing the desired aldehyde as a minor component. A diene AHF strategy was attempted but resulted in very poor enantioselectivity due to rhodium allylic rearrangements during the AHF catalytic cycle. AHF of a trisubstituted olefin resulted in a complex mixture of aldehydes. A rhodium-mediated domino annulation (RMDA) of delta-alkynyl ketones and α-boryl-α,β-unsaturated esters producing fused pyranones was developed. Transmetallation of the vinyl boronic ester with the rhodium catalyst produces a vinyl rhodium intermediate that undergoes a highly regioselective syn addition to the alkyne to produce a second vinyl rhodium intermediate that immediately attacks the pendant ketone, forming a new five-membered ring. The rhodium alkoxide formed by the cyclization transesterifies with the α,β-unsaturated ester to form a fused lactone ring system. Overall, two new carbon-carbon bonds and one new carbon-oxygen bond are formed, creating two new rings. Several examples are shown exploring substrate functional group tolerance and steric limitations. A total synthesis of the recently isolated natural product linderagalactone C was initiated, but production of both components of the RMDA met significant synthetic roadblocks. A planned Eschenmoser-Tanabe fragmentation failed to form the necessary δ-alkynyl ketone and the boronic ester for the RMDA was unable to be accessed from α-halo esters due to the severe electron withdrawing effects of the carbonyls.