A Concise and Versatile Double-Cyclization Strategy for the Highly Stereoselective Synthesis and Arylative Dimerization of Aspidosperma Alkaloids
Jonathan William Medley and Mohammad Movassaghi
The aspidosperma alkaloids belong to the family of monoterpene indole alkaloids which contains more than 2000 members. I think most of you are more or less familiar with their structures. Because of their broad structural diversity this family still challenges chemists to test new methodology. The Movassaghi group recently published this paper which contains an impressive Friedel-Crafts cyclization strategy to build up the framework in a concise manner. By the way three biogenetically related group members were synthesized and some analogous compounds.
The group planned to access all three natural products through a common precursor which can be obtained via an interrupted Bischler-Napieralski reaction. Fragment 8 was synthesized utilizing Myers asymmetric alkylation strategy.
Pseudophenamie 1 was acylated with crotonyl chloride to give amide 2 which in turn was deprotonated and alkylated to give 3. By doing so the endo-double bond was transformed into a terminal olefin. Another alkylation introduced the ethyl group while retaining the stereochemistry at the a-position.  TES protection of the auxiliary was necessary to overcome problems in the following alkylation/ring closing step.  Coupling partner 6 was obtained through methylation and chlorination of 5 in a straightforward manner. Alkylation of 6 with 4 was achieved with KH in the presence of TBAI to give acyclic precursor 7 in high yield.
Next the nosyl group was removed with PhSH. In one pot the TES group was cleaved which resulted in the expected N à O acyl transfer.  The ester then easily formed lactam 8 with complete recovery of the auxiliary in almost quantitative yield. Triflation of lactam 8 in the presence of the slightly basic 3-cyanopyridine produced the key diiminium ion shown. Depending on the following steps a lot of derivatives can be accessed.  Employing first borohydride reduction and hydrogenation (-)-N-methylaspidospermidine was obtained. Using a buffered aqueous solution of TFA the diiminium salt was hydrolyzed, the double bond hydrogenated, and the carbonyl functionality reduced with LAH to give (+)-N-methylquebrachamine.
Going half the way from the diiminium ion (which means leaving the double bond in place) coupling partners 12 and 13 were obtained. Again forming the diiminium ion from 13 in the presence of 12 iminium ion 14 was generated. Reduction with Red-Al and hydrogenation then gave (+)-dideepoxytabernaebovine.
For clarity I put the mechanism of the Friedel-Crafts chemistry below. Triflate formation is straightforward. The following spirocyclization is controlled by the quaternary stereocenter. Most likely the ethyl side chain poses greater steric repulsion and the vinyl group might exhibit some sort of attractive secondary orbital interactions. The formed indoleninium ion then underwent aza-Prins cyclization to give after HCl elimination the diiminium ion used for further modifications.
Extremely cool chemistry. I skipped to show all the analogs the group synthesized by the way but you really should have a look in the paper. It is highly recommended.
 Any guesses why the stereochemistry of the vinyl group is retained in this step? Normally it should be inverted I think…
 It was found that during the coupling step the resulting free amine after N àO acyl transfer underwent intramolecular alkylation with the chloride to close a lactone ring.
 The fast N à O acyl transfer can be explained when you look at the 3D model below:
Because of the large phenyl groups the amide nitrogen has almost no chance to overlap its non-bonding s-orbital with the antibonding p*-orbital of the carbonyl group. So the normally partial double bond character of the amide bond is weakened. On the other the free alcohol oxygen is very close to the amide carbonyl so that an acyl transfer should be really fast. I can only guess why this transfer is observed, maybe you have another explanation for that?
3D-model (click on the image to get an impression of the 3D structure):
 As nucleophiles the group employed for example Grignard reagents, allyl silanes, enol esters, or electron-rich arenes.