Synthetic studies directed toward the tremorgenic indole alkaloid penitrem D and the design of non-peptide peptidomimetics: Development of a novel scaffolding for proteolytic enzyme inhibitors

Chapter 1 of this dissertation describes synthetic studies directed toward penitrem D. The retrosynthetic analysis employed for penitrem D divided the molecule into a left hand aniline (64)* and right hand lactone (65) portion. Three approaches to right hand synthons were investigated. The key synthetic maneuver involved creation of the vicinal quaternary centers (C(31) and C(32)) with the correct relative stereochemistry. The first scheme proceeded in 14 steps to afford lactone 160. The key step involved sigmatropic rearrangement of 151 with subsequent trapping and alkylation of the resulting stereospecifically generated enolate. However, the synthesis proved quite inefficient proceeding in just under 2% overall yield and involving the unstable exo-methylene ketone 79. In an attempt to improve the efficiency, a second approach involving the synthesis of the more elaborate lactone 65 was attempted. The key coupling reaction between ketone 79 and halide 168, however, failed. Finally, an aldol approach to lactone 160 was successfully implemented. This scheme both avoided the unstable exo-methylene ketone 79 and provided the vicinal quaternary centers more efficiently. Retreoaldol reaction in the coupling of 149 and 173 could be avoided by quenching the reaction at low temperature or adding $\rm MgCl\cdot Et\sb2O.$ Chapter 2 presents the process whereby non-peptide peptidomimetics with novel scaffolding of enzyme inhibitors were developed. This type of mimetic differs from the non-enzyme counterpart by the need for hydrogen bonding potential in the designed framework. The enaminone functionality was selected for use in the novel scaffolds for its hydrogen bonding ability. The required binding conformation of renin inhibitors was deduced from published data on both aspartic acid proteases X-ray structures and renin active site models. The conformational data were applied to the known, potent inhibitors of renin 29-31 and MM2 calculations performed. The resulting model inhibitors were used both as design templates and for comparison to the minimized conformations of the peptidomimetic inhibitors. The inhibitor design proceeded in three stages. The first involved identification of heteroatoms participating in hydrogen bonding and subsequent deletion and replacement of the remainder with the novel scaffolding. This method generated molecules 34 and 35 which were either not synthesized or inactive. However, this first generation approach emphasized the need for synthetic practicability and lead the exploration of an iterative strategy. Next, both the 3,5- and 2,5-linked pyrrolinone scaffoldings (45 and 48, respectively) were designed; molecular modeling indicated that both pyrrolinone backbones should adopt a $\beta$-strand conformation. Synthesis of a tetrapeptide 3,5-linked tris-pyrrolinone mimetic 69 resulted in a crystalline compound which indeed adopted a $\beta$-strand conformation in the solid state and also formed both parallel and antiparallel $\beta$-pleated sheets in the unit cell. Subsequently, the nitrogen displaced pyrrolinones were employed in the design of both renin (72 and 73) and HIV1 protease (75) inhibitors. Finally, modeling suggested that the carbonyl displaced pyrrolinones can also be exploited in the design of HIV1 protease inhibitors. ftn*Please refer to dissertation for diagrams