B.R.S.M. All this happened, more or less.

30Jun/1213

(+/-)-Echinopine B

Vanderwal (ˈvændəˌrwɑːl)

v. Vanderwalled, Vanderwalling, Vanderwals

v. tr. To complete an impossibly short, but racemic, synthesis of a popular target e.g.

 

A Synthesis of Echinopine B

Vanderwal et al., Angew. Chemie. Int. Ed., Early View [GROUP] [PDF] [SI]

DOI: 10.1002/anie.201203147

Since their isolation in 2008 the echinopine sesquiterpenes have proved quite popular targets for total synthesis, probably thanks to their unusual and compact molecular architectures. Johann Mulzer, one of my favourite living synthetic chemists, succeeded in a beautiful first total synthesis just a year after their isolation, asymmetrically synthesising both natural products (starting from cyclooctadiene!) and assigning their absolute configurations in an excellent paper that somehow ended up in Organic Letters. The year 2010 saw rather a lengthy synthesis by Nicolaou as well as a paper  by Chen outlining studies that developed into a total synthesis last year, which I commented on at the time.[1] Until a couple of weeks ago, Mulzer led the pack needing a mere 20 steps from commercial materials, with Chen a close second at 24, and KCN's ponderous 39 step route languishing at the back. At the time it came out, I didn’t think Mulzer was doing too badly, given the dearth of oxygen in the target and the lack of obvious disconnections. However, the synthetic bar has just been raised by Vanderwal and co-workers, who recently reported a nifty 12 step, albeit racemic, total synthesis that I think should stay at the top for a little while.

The route began with the construction of a guaiane-type precursor from a cheap commercial ketoester. The seven-membered ring was formed first, by conversion of the ketone to the corresponding TMS enol ether, followed by cyclopropanation and treatment with ferric chloride to effect ring expansion to the enone. Cuprate addition then gave a good yield of a mixture of diastereomers, which, although separable, were carried through the following sequence together.[2] The TBS protected alcohol was converted to the corresponding tosylate, and formation of the five-membered ring by intramolecular alkylation then gave the required cis-bicyclo[5.3.0]decan-8-one.[3] Next, Wittig olefination began the process of removing unwanted oxygen functionality from the molecule, and this was continued by conversion of the ester to the aldehyde and thence the alkyne through use of the handy Ohira-Bestmann reagent.

With the guaiane-type precursor in hand, it was time to test the last few bond forming steps to contruct the final two rings. I notice that the group refer to the approach they take as ‘bioinspired’, presumably in reference to the order of bond construction rather than the reagents and conditions used. After functionalisation of the alkyne with suitable substituents for later development into the ester, a rather nifty platinum catalysed cycloisomerisation reaction was used to simultaneously form both remaining rings. In fact, the group developed two very similar approaches, depending on which ester equivalent was used. The higher yielding option found to be a methoxymethyl group, introduced by alkylation of the alkyne with MOMCl. When this substrate was heated platinum(II) chloride in toluene, the cycloisomerisation cascade gave the enol ether shown,[4] which could then be oxidised directed to the methyl ester using a little PCC. Alternatively, a slightly more direct (but lower yielding) route using trimethyl orthoformate in place of MOMCl was also described. Exposure of the resulting dimethyl acetal to the same conditions temporarily gave the dimethyl ketene acetal, which upon work-up to directly yielded the target molecule.

A great piece of work, admirably concise from the title of the paper right through to the final step!

 

Tangential Information

1. Totally Synthetic covered the Nicolaou synthesis, which I wasn’t particularly impressed with, so I’ll just put the graphical abstract up here, to serve as inspiration for the on-going Colouring Competition.

The fact that KCN's 39 step route to these targets got into JACS when Mulzer's 20 step one, which was also the first, ended up in Org. Lett. seems like a bit of an injustice to me.

2. Due to the conformational flexibility of the system, the relative configuration of the two centres could not be readily determined using NMR methods, even after formation of the second ring.

3. The yield obtained for the intramolecular alkylation is a little on the low side, but was accompanied by recovery of some 30% of the starting materials. Due to the lack of a strong cis/trans preference in these types of systems, prolonged reaction times were often accompanied by epimerisation so the reaction had to be stopped at quite low conversion. In fact, these 5-7 systems can be pretty tricky to work with when the opportunity to epimerisation is open to them:

From T. Hudlicky et al., Synlett, 2005, 2911-2914

4. I'm loathe to try and draw a mechanism for this, as the last two times I've proposed mechanisms for reactions on this blog I've been wrong. Oh, go on then... you guys can correct me in the comments:

Update 30/06/12:  See Arr Oh's suggested mechanism:

Comments (13) Trackbacks (0)
  1. Ahh, the neverending issue of the community frowning upon KCN publishing next to anything in JACS. Note, however, that Echinopines were published as an JACS Article, not Communication, which lessens the novelty requirement a bit for acceptance. I would not be at all surprised if the authors pushed for a communication but got bashed by reviewers and had to rewrite it as a full paper.

    While I admire Vanderwal’s work (a few nicely spotted disconnections and very interesting enyne cycloisomerisation as a key step), I dare postulate that making the starting ketone chiral could be cumbersome to say the least and might increase the step-count a bit. All in all, great job this one, definitely earned its place in a top journal like ACIE.

    • I don’t so much think that the Nicolaou paper didn’t belong in JACS, as the Mulzer paper did more/as well. I do, however, have no JACS papers myself. I also realise that the initial retro for the Nicolaou route was probably not 39 steps, but probably got extended along the way, as so often happens. I think you’re right about the fact we won’t see an asymmetric version any time soon; just like his strychnine synthesis, it’s not easily amenable to it. Still, both are excellent and well executed syntheses!

    • The asymmetric step would have to be formation of the enol ether – I don’t have literature access right now, but pages 8-12 of this seminar suggest it’s very much possible: http://stoltz.caltech.edu/seminars/2005_Ebner.pdf

      Also, I agree with See Arr Oh on the Pt step.

      • the nicolaou synthesis is most certainly NOT an excellent synthesis; the work was done in the Chen Lab, and if KCN didn’t volunteer to put his name on it there is no way it would have made its way into JACS. This molecule, while very intricate, contains fewer than 20 atoms (not including H), which means that this synthesis required, on average, more than 2 steps per atom. While step count/molecular weight ratio may not be a fair way to analyze a synthesis, I think in this case it emphasizes the lack of efficiency in this synthesis. Alternatively, you could look at the number of protecting group manipulations and oxidation state adjustments made in the synthesis; it’s atrocious. The Nicolaou/Chen paper is a good example of how NOT to do natural product synthesis.

        • also, the only reason it was published as an article is because there is no way you could fit a 39 step synthesis in a communication. Most JACS articles take time to explain key insights in detail and discuss the significance of the work carried out; here the entire 4 pages is devoted to delineating the step-by-step sequence required to complete the synthesis

        • Agreed. I was rather ambiguous in that comment – I was trying to say that both the Vanderwal synthesis (i.e. strychnine and this one) are excellent, despite the fact that I can’t see an easy way to make them asymmetric. I definitely would not describe KC’s route as excellent. Hell, I called it ponderous in the post. I used to think Mulzer’s synthesis, which was the first, should have been the last, but I quite like Vanderwal’s.

  2. #4 – So close! In my mind, Pt(II) is a good pi-Lewis acid, so the first step woul dseem to be coordination to the alkyne. 6-exo cyclization (from the “southern” olefin) creates a temporary 7-6-5 ring system, with a cation at the ring juncture and an sp2-bound Pt. This Pt-olefin acts like an enolate (like you have drawn) to slam shut on the cation, closing the cyclopropane, and generating a Pt carbenoid. The rest of the mech finishes as you’ve suggested.

    (I sent a ChemDraw to your inbox…if you like it, feel free to post it!)

    • Also, note that the SI gives only a combined 55% yield of echinopine and its bypdt. The other 45% probably results from “wrong” olefin closure, or early cracking and degradation of the acetal.

      Caveat synthesizor.

  3. not keen on curly arrows from a ‘Pt’ vinyl complex to form a carbene complex.

    • Search the literature for carbenoid mechanisms involving platinum, gold, copper, or silver. You’d be surprised how commonly they’re invoked.

      • no issue with that, but think about the electron counting. If you go from a vinyl complex to a carbene complex, that is probably best represented as a reductive elimination from the metal centre. That’s not consistent with a curly arrow from the metal centre.

        • the mechanism drawn above is missing a formal negative charge in the vinyl metal species, so going from that intermediate to a carbenoid involves no net change in oxidation state

  4. There are no oxidation states on Pt, so it’s hard to say if there is or not.


Leave a comment


2 × = fourteen

No trackbacks yet.