Totally Synthetic by Paul H. Docherty, 25 July 2006

Total Synthesis of Vindoline

Boger

H. Ishikawa, G. I. Elliott, J. Velcicky, Y. Choi, D. L. Boger, J. Am. Chem. Soc. 2006, 128, 10596-10612.

More outstanding cyclisation action, this time from the Boger labs at Scripps. Working towards a whole class of bisindole alkaloid structures, including Vinblastine and vincristine (known for their potent antitumor action), they have completed the total synthesis of vindoline. This structure, also found in nature, is a synthetic and biosynthetic precursor to the former targets. Along with an accompanying methodology paper, this work describes their forays into the synthesis of this interesting motif.

The work hinges upon an immense [4+2] / N₂ elimination / [3+2] sequence that creates three rings and sets six stereocenters in one pot. In the first case, they completed the synthesis of the related structure, minovine. The cyclisation sequence went particularly well (in 74%); they then had to remove an extraneous carbonyl group, which they completed using Lawesson’s reagent to convert to the thiolactam, and then reduction with Raney nickel.

They then opened the ether bridge and used Burgess’ reagent to provide a good leaving group for the liberated alcohol, which allowed then to eliminate, but unfortunately as a mixture of natural product and the undesired alkene isomer.

They were then able to showcase this methodology and strategy further, with short syntheses of vindorosine, 4-desacetoxyvindorosine, N-methylaspidospermidine, and finally, vindoline, which itself was completed in a remarkable eleven steps! Using relatively familiar methodology to get to the cyclisation precursor, they produced both cis- and trans- material, and cyclised. Both gave the expected products, with the trans cyclising faster, and the enantiomers separated using HPLC. Then, reducing the carbonyl as before, along with opening of the ether bridge returned the target.

Selected Comments

26 July, 2006 at 23:51, Hap says:
The diazo compounds don’t need 180°C, but even if they did they probably wouldn’t be able to stand it. The thermal conditions aren’t great, but the compound should be bench-stable (and unreactive under most conditions); it also doesn’t have a whole lot of complexity, and any complexity present can be assembled piecewise. The rhodium catalyst for diazo compound cycloaddition runs about $160/g (Aldrich, ’03 - probably about $20 for a 5 mmol-scale reaction), and the diazo compound is going to have to be generated from something else more stable. Unless the reaction is messy and has lots of byproducts, thermal (that doesn’t require FVP or any of its ilk) reaction of a stable precursor seems preferable to the generation of a diazo compound followed by RT reaction - but YMMV.