Guest Blogger: See-Arr-Oh
It’s not too often JACS publishes total syntheses anymore, since their role as a general chemistry journal means every newly built molecule competes against quantum dots, in silico simulations, or new battery materials. So you can imagine my utter shock and delight at finding not one, but two almost back-to-back syntheses in the past month! First, the Porco / Ready combination of Kibdelone C, followed closely by two syntheses that didn’t receive as much play in the blogosphere: (-)-Trigonoliimines A-C (Mo Movassaghi, MIT), and (±)-Trigonoliimine C by rising newcomer Uttam Tambar at UT-Southwestern.
(I think Tambar’s 3D representation looks like a pterodactyl…you?)
These syntheses are fairly straightforward: they don’t have exotic protecting groups or cascades, and the products don’t cure cancer by the nanogram. Why make them, you ask? All the trigonoliimines show modest anti-HIV activity, but their biosynthetic hypothesis is what’s really intriguing here. Polyketide synthases may chug along; inserting propionate units to churn out giant macrocycles, but sometimes Nature sits back, considers its options, and decides to alter very simple, abundant feedstocks through coupling and rearrangements. The isolation team (Hao, OL, 2009) believes all three trigonoliimines derive from a tryptamine and a kynuramine stitched together and cyclized.
Movassaghi and Tambar both believe that the biological precursors would be two different tryptamines instead of a tryptamine / kynuramine. The natural product can be retrosynthetically derived from the hetero-indoxyl-indole dimer shown below. Both groups even have similar plans to access it, viaa [1,2]-Wagner-Meerwein shift of an alkylamino group from a suitably functionalized tertiary alcohol precursor (A), which in turn comes from bis-tryptamine B. In both syntheses, the biaryl system is formed through a Pd-catalyzed cross-coupling.
Let’s start with Tambar’s synthesis first, since both syntheses of Trig C are fairly similar. Beginning with 6-methoxytryptamine, Tambar formylates the aliphatic nitrogen, then uses Boc anhydride to protect the indole nitrogen, followed by Boc-directed lithiation / stannylation at C-2. Brominated tryptamine 2 is coupled with a dash of Pd tetrakis at 110oC, which conveniently deprotects the Boc group as well.
To access the Wagner-Meerwein intermediate A, Tambar throws the oxidative “kitchen sink” at the bis-indole. Treatment with standard reagents like mCPBA, air, or Oxone oxidizes the electron-poor indole, but the group works around the problem by coordinating the electron-rich indole to an iodonium salt, which allows water to add in at C-3, producing hydroxyindoleine A (Note: Although I’ve drawn a single enantiomer of product, this is a racemic synthesis).
Playing around with a model system taught Tambar that the [1,2]-Wagner-Meerwein could be promoted by Sc(OTf)2, but in this case, the metal drives formation of a dihydrofuran intermediate, which migrates the oxygen to the “wrong” indoxyl.
How do we fix this? Formic acid in toluene or DMF leads to oxindole or reduced products. Luckily, treatment of A with HCl / dioxane at a buck-fifty in wet DMA produces the desired indoxyl. Finishing up, hydrazine deprotection of the phthalimide, followed by Ti-promoted dehydration leads to (±)Trigonoliimine C, with a LLS of 8 steps and a 14.6% overall yield.
Movassaghi accesses nearly the same intermediate – C, from Ir-catalyzed borylation of phthaloyl MeO-tryptamine, Suzuki coupling, and oxidation – only enantiomerically enriched thanks to a Davis’ oxaziridine. He remarks here that he, too, has problems with oxindole byproducts if he uses Lewis acids to promote the [1,2]-shift. After “significant experimentation,” the alcohols are heated in TFE to produce the desired compounds in 93% yield. A similar endgame ensues, with deprotection, Ti-promoted dehydration, and formylation, to access (-)-Trigonoliimine C (96% ee), LLS = 8 steps, 9.3% overall from 6-methoxy-tryptamine (Note: The compounds are carried through the last few steps as a mixture, so he also isolates “isotrigonoliimine C”, which is the compound where the methoxy group has been swapped to the other indole)
Overall, he uses his intermediate in much the same way as Tambar to complete Trigonoliimine C, but here’s where Movassaghi turns on the afterburners – since he already has the “biosynthetic intermediate” in hand, he might as well make Trigonoliimine A and B, right? Back to intermediate C! Switching up the steps from the earlier synthesis, such that the deprotection of the phthalimides happens first, a killer bicyclo-[3.3.0] bis-pentacyclic aminal (D) emerges. Treatment with Martin sulfurane cracks the aminal and presumably forms a benzylic cation from the 3o alcohol, which the remaining amine catches to form azepane E.
Rotation of the aryl group 180o brings the aniline and secondary nitrogens close enough to cyclize together in the presence of triisopropyl orthoformate/ PPTS. Overall, the team finds a LLS of 7 steps for both compounds, 5.2% overall for Trig A (94% ee), and 4% for Trig B (95% ee).
(Editorial – this is the second guest post by See-Arr-Oh – and isn’t it a stunner! If you’d like to contribute get in touch!)