2330 - update: Just noticed the Tot. Syn. writeup. Check it out for an alternative discussion! Funny that this work's two weeks old, yet we both write about it today!
An Efficient Synthesis of (+)-Loline Alkaloids
I just noticed this very efficient enantioselective synthesis of a trio of the loline pyrrolizidine alkaloids reported by Trauner and coworkers back in June. Although loline has been known for more than a century, prior to this work only one enantioselective synthesis had been reported, and at 20 steps in length - that's well over two synthetic operations per carbon atom - it's as good an example as any to illustrate the point that size isn't always what makes a target challenging. Trauner, with characteristic German efficiency, took only 9 steps to reach loline through clever use of some simple and largely well established chemistry. Reading this paper reminds me of one of the most important axioms for designing anything (syntheses, websites, software, machinery...): that a thing should only be as complicated as it needs to be - and no more.
The route starts from the homochiral allylic alcohol below, which is available by enantiotopos- and diastereoface-selective Sharpless allylic epoxidation of achiral, and surprisingly expensive, 1,4-pentadien-3-ol. The epoxide was then opened with the homoallyl amine, the newly formed secondary amine was protected as the benzyl carbamate, and Grubbs' 2nd generation catalyst was used to effect RCM to give the 8-membered ring. Reaction of the syn-diol with thionyl chloride formed the cyclic sulfite in essentially perfect yield, and this was then opened regioselectively with lithium azide at the more reactive allylic position. Despite the fact that azide is an excellent nucleophile, the reaction required heating to 130ºC in DMF, but the yield obtained was still impressive.
The resulting azido alcohol was then treated with a single equivalent of bromine at 0 ºC in methanol, effecting the key transannular cyclisation, and delivering the bromopyrrolizidine (as the hydrobromide), in near quantitative yield. Etherification required inversion of the halogen center, and this was accomplished by an unlikely looking Finkelstein reaction. Upon treatment of the alcohol with potassium carbonate and heating in the microwave, an intramolecular Williamson-type SN2 ether formation occurred, closing the morpholine ring and establishing the skeleton of the natural product. Conversion of the azide to the methyl amine was done with typical Teutonic efficiency (and a neat two step sequence I hadn't seen before). Reduction of the azide using H2 and catalytic Pd/C proceeded smoothly, but the reaction was performed in the presence of di-t-butyldicarbonate (Boc2O). You might think that Boc protection at this stage is a bit of an odd move, until the next step where the group completed the synthesis by smashing the carbamate all the way down to the final methyl with excess LiAlH4 in refluxing in THF to give loline. N-Formyl loline was made by formylation of loline with acetic formic anhydride and temuline was accessed by simple hydrogenation of the tricyclic azide without any Boc2O present, reducing it to the expected primary amine. Lovely.
The paper starts with a discussion of the different metrics we use to assess the efficiency of a synthetic route, things like: overall yield, Trost's atom economy, Wender's step economy, Baran's ideality and redox economy, and the generally accepted tennet of minimal use of protecting groups. However one looks at this, I think the Trauner group did pretty well here.