Cubanes! Get Your Cubanes, Here!

A cube of med-chem cubanes**

It's not often you see drug candidates advanced with cubane substructures.* A collaborative group from the UK branch of AstraZeneca and Oxford University aims to change that: a report in Organic Letters indicates they've produced gram-scale quantities of cubane derivatives with common med-chem handles already "baked in."

Starting from a 1,4-cubane diester, Burton, Davies, and Wlochal take turns elaborating each side into thiazoles, beta-diketones, nitriles, amine salts, and ynones. Though more cautious readers might expect the cubed synthons to go up in smoke, they actually seem to survive fairly rigorous conditions - boiling water, phosphoryl and oxalyl chloride, TMS triflate, LDA.

I'd be most interested to hear from any readers who integrate these new building blocks (sorry, couldn't resist!) into their research. It's hip to be...cubic.

* On Twitter, CJ (rightly) points out that the authors could have done some DSC (stability) and solubility assays, just to put minds at ease that these synthons really could be integrated into a drug campaign.

**Not a real substance (but wouldn't that be something???)

Friday Fun: Sweet Cardamom (Peroxide)

Rice pudding. Ginger snaps. And...malaria?

That's what'll be going through my head next time I cook with cardamom, thanks to Tom Maimone and coworkers (UC-Berkeley) and their under-the-wire JACS ASAP from yesterday afternoon. The title and abstract scratch all the Baran lab alumni itches: 1) biosynthetically inspired, 2) novel mechanisms, 3) scalable, 4) just four steps! And hey, we're making stable endoperoxides, which all the cool kids are into nowadays.



Not their actual abstract graphic...

As Maimone points out, the latent symmetry of the final product offers a really neat assembly strategy. The group McMurrys together two units of (-)-myrtenal, then hits it with singlet oxygen, initially forming a 6-membered endoperoxide they fragment / rearrange with base. A gentle oxidation (DMP) sets them up for the wild step: stitching together a 7-membered endoperoxide using Mn(III)*, a radical source, a silane reducing agent, and even more oxygen. Simple phosphine reduction knocks down the last hydroperoxide into an alcohol, and the whole target (7 stereocenters!) falls out as a single stereoisomer.

Pretty sweet.

P.S. - Since the group's made over half a gram in just this first push, I'd assume an efficacy paper against live Plasmodium parasite hot on the heels of this one...

*We're apparently already calling this the "Shenvi catalyst"...wasn't this only two months ago?

What-a-Ouabagenin! Grams on Demand

I can't believe I got back from New Year's without finding a single post on ouabagenin* [wah-bah-jenn-in], the latest from Phil & Co. in Science this past week:

Ouabagenin, a polyhydroxylated (>5 -OH groups) cardenolide (steroid with an appended lactone) positive inotrope (helps heart pump more forcefully) had been completed only once before, in a 40+ step relay synthesis by Deslongchamps in 2008 (got all that?!?). Only a few mg were prepared, and those of you familiar with the Baran group know that the only real way to make natural products is with a shovel and bucket - gram-scale, baby!

Hulkster - Quite interested in
gram-scale ouabagenin precursors...

So, we start out with 20 g of cortisone acetate - just one reduction shy of Preparation H - and two steps later have a fully protected version of adrenosterone. The group first tries a porphyrin-catalyzed C(19) hydroxylation (the bottom-left methyl), which doesn't work, so they opt instead for some fancy solid-state photochem to generate a cyclobutane ring, which selectively pops open with NIS under sunlamp irradiation.

Selective de-ketalization and iodide hydrolysis sets the stage for a three-step sequence (peroxide, SeO2, peroxide) to generate a diepoxide (right, top), a.k.a. the "most difficult transformation to secure on scale." They toss a "gamut of conditions" at the molecule, only to receive mixtures of enones. Finally, they find that using in situ Al-Hg amalgam (we're talking foil / scissors here!) combined with Sharpless "on water" suspension produces the desired triol, which they wrap up as an acetonide.

Next, "superhydride" reduction both reduces the ketone and protects - as a boronic ester - the remaining two hydroxyl groups. A little Saegusa-esque dehydrogenation, a fluorous solvent-enabled bond migration, and a Co-catalyzed hydration produces 'protected ouabageninone' (right, bottom).

Endgame - We're not out of the woods yet, folks! Conversion of that lone ketone into the vinyl iodide (hydrazine, iodine, TEA) followed by a modified Stille returns a butenolide diene. They again toss in a 'kitchen sinkful' of reductants, only to find that dicobalt-borane (cool!) followed by Barton's base (N-tert-Bu-TMG) produces the correct butenolide orientation (3:1 dr). A touch of HCl in methanol liberates the natural product.

Despite the fact that they report the last few steps on just 30-60 mg, the group claims that they have >0.5 g parked at the protected ouabageninone (vide supra). With this synthesis, Baran also alludes to the overall usefulness of his "redox-relay" strategy, which has certainly served him well before.

*Bonus - In the Scripps press release, Phil calls ouabagenin "probably the most polyhydroxylated steroid known on planet Earth." Billions of yet-undiscovered microorganisms could not be reached for comment.