B.R.S.M. Yield isn't everything


If I Had A Blog Back Then 1: (+/-)Acanthodoral

More updates soon; I'm trying to finish my thesis this week as I have to start work on Monday!

Zhang and Koreeda, Org. Lett., 2004, 6, 537

DOI: 10.1021/ol0363063

Yes, there's been plenty of good stuff in the literature recently, so much so that I'm hoping to soon resurrect my old habit of making 'Top 5 syntheses of the month' lists this week, retrospectively starting with April. However, as I'm a bit busy right now, here's a short post on some slightly older work. A co-worker showed me this paper last week, which he'd found trawling the ACS website looking for good problem session material, but was worried might be 'a bit mental'. I'm not so sure, but then my last few problem sessions have included dodecahedrane and discodermolide. There are quite a few interesting steps, especially for an Org. Lett., but some of the chemistry you'd just never expect to work as it does.

The route began with a fairly standard reaction - the addition of methylmagnesium bromide to an enone in the presence of catalytic copper(I) and LiCl, followed by quenching with gaseous formaldehyde. This was followed by a rather useful but little known reaction; the Takai-Oshima-Lombardo methylenation of the ketone.[1] The newly formed alcohol was then oxidised under Swern conditions and treated with propen-2-yl magnesium bromide, followed by methyl chloroformate, to give the expected allyl carbonate.  Next came the route's sexy step - a cool carbonylative metal ene-reaction. Thus, heating the carbonate in wet acetic acid with palladium(II) acetate and triphenyl phosphine under a single atmosphere of CO gave the hydrindane acid in quite reasonable yield. So far so good.

The carboxylic acid was the converted to its methyl ester by alkylation with iodomethane; I guess a Fisher-type esterification here is probably not a good idea. This was followed up by a ring expansion using dibromocarbene, which formed the expected cyclopropane, followed by treatment with p-methoxythiophenol and potassium carbonate in hexafluoroisopropanol to opening up the new ring with loss of bromide.[2] Finally, the thioether was oxidised to the sulfone with two equivalents of m-CPBA. This is where things a get a bit silly, as treatment with thiophenol in aqueous hydoxide - methanol then give the acid shown... with replacement of bromide by thiophenol![3] The acid was then coupled with selenophenol, using the horrific looking phenyl dichlorophosphate as a coupling agent. Ewww. Irradiation of the selenoester in the presence of hexabutyl ditin (!) then lead to a non-reductive radical cyclisation onto the nearby double bond with loss of the sulfone by fragmentation. Low temperature desulfurisation with RaNi and hydrogenation of the remaining double bond with Adam's catalyst gave the tricyclic ketone in impressive yield. The group were now only a Wolff rearrangement and a reduction away from victory. Introduction of the diazo group required for the Wolff rearrangement was carried out in two steps, first making the oxime by treatment with isoamyl nitrite and then using ammonium hydroxide and bleach to form the α-diazoketone. Finally, photolysis in methanol carried out the expected ring contraction to give the cyclobutane methyl ester that was then converted to the required aldehyde. Nature doesn't make many cyclobutanes, but I'd say Koreeda's synthesis of this one is pretty neat.



1. As far as I know you can only carry this out as a methylenation (i.e. to introduce a =CH2), like a Tebbe reaction. Also like the Tebbe, it seems to be very useful where other olefinations fail. A friend of mine had trouble methylenating some halogenated 3-acetyl indoles - normal Wittig, Tebbe and Peterson conditions just didn't do anything, nor did the compounds react well with methylmagnesium bromide or methyllithium. The Takai-Lombardo worked great, though!

2. This kind of ring expansion is quite popular; there's an example in my coverage of Yang's schindilactone A (3rd scheme)  and a similar thing at the start of Baran's synthesis of taxadiene.

3. In retrospect, with the product and starting material shown it's not too hard to see what's going on. Presumably the double bond moves into conjugation with the sulfone, addition-elimination of bromide occurs, and then the double bond moves out of conjugation again (for some reason). Not sure I'd want to try and explain why to a room full of chemists. So, why even bother with the thiophenol? Well, the authors say that they originally just tried aqueous hydroxide in methanol - typical hydrolysis conditions - but these resulted in a similar substitution of bromide by methoxide, but with the double bond remaining in conjugation with the sulfone! Unfortunately, this messed up the next step, which required the double bond where it was, so the conditions had to be changed. How they knew that addition thiophenol would help (or why they'd try it if they didn't) is beyond me.

Comments (3) Trackbacks (0)
  1. Good luck with finishing your thesis! I’m sure selenophenol is delightful. Also, you deleted the gem-dimethyl in your second figure.

  2. Well, concerning the WTF step – if one assumes the double bond to move into conjugation that would mean the generation of a tertiary carbanion, which to my mind is not keen of being attacked by a nucleophile… Or am I wrong?

    Though, I have no idea at all… :-/

  3. So you are in Trauner group? that pagodane took me hour and half to work trough…..

Leave a comment

2 × = six

No trackbacks yet.