Abstract
Ph.D.
The objective of the research described in this thesis was to develop a synthetic
method that can be applied to the synthesis of the natural
bisbenzyltetrahydroisoquinoline alkaloids cissacapine, insularine, insularoline,
cycleanine and analogues thereof.
In this study two different strategies that allow easy entry to the precursors of these
alkaloids were developed, and these set the scene for future total synthesis of these
alkaloids. The key features of the first approach comprise the linkage of the two
appropriate rings to form the diaryl ether moiety as well as the preparation of the
11H-dibenzo[b,e][1,4]dioxepine tricyclic system. Previous approaches to the diaryl
ether formation are not suitable for large-scale reactions. We have herein described
the preparation of the diaryl ether precursors in high yields and our approach is
suitable for large-scale preparations. A search of the literature method revealed only two published methods for the
preparation of the 11H-dibenzo[b,e][1,4]dioxepine system. Both these two methods
produce compounds containing this moiety in low yields. In our studies this aspect
was addressed satisfactorily. Unfortunately, our attempts to complete the synthesis of
these alkaloids through Bischler-Napieralski reaction was met with no success, the
problem been ascribed to the unoptimised Bischler-Napieralski conditions used.
Our second approach involves the preparation of benzylisoquinoline units that are
precursors of cycleanine. The published method to the derivatives of the cycleanine
precursors is non-stereopecific and produces racemic benzylisoquinolines. Our synthetic route is a chiral auxiliary-based asymmetric version that produces the
optically-pure benzylisoquinoline monomers. The key features of this route involve
incorporation of the chiral auxiliary on the nitrogen atom, Bischler-Napieralski
cyclisation of the resultant chiral amides and finally stereoselective reduction of the
3,4-dihydroisoquinolinium ion possessing the chiral auxiliary. This route employs
both optically-pure (S)- and (R)-1-phenethylamine as the chiral source. Optically pure diastereomers were obtained. Our approach is a vast improvement compared to
the previously described non-stereospecific method since it allows easy and good
stereoselective access to both diastereomers in good yield. Unfortunately, one of the
concluding steps leading to the formation of the dimeric stereoisomers of cycleanine
through diaryl ether formation using the recently published methods was not
successful. This is ascribed to the electron-rich nature of the isoquinoline ring.