[Archer]

Hi All,

This reaction has been troubling me for some time now and I would like the oportunity to discuss it with the wider chemical community.

It is taken from this publication

http://pubs.rsc.org/en/content/articlelanding/1966/j3/j39660001615

The authors did not include a mechanism, just a scheme and experimental conditions.

On first glance it would look like a simple cyclysation but the proton would have to be extracted from what is, in my opinion, an unlikely source.

Would anyone care to comment on this? There is no doubt that the reaction works. If I recall the reaction does not yield much product but this seems an unlikely mechanism.

Best wishes

Archer

[Dan]

Quick question - is the starting material enantiopure and if so do we know whether the stereochemistry at the tertiary centre is retained?

I am also very skeptical of this deprotonation, especially given the pKa of the H flanked by 3 carbonyls is probably somewhere around 8!

[Archer]

Quick question - is the starting material enantiopure and if so do we know whether the stereochemistry at the tertiary centre is retained?

No, interestingly enough, as this was the total synthesis of a molecule which contains two chiral centres. I did look, many years ago, at trying to attempt this with an enantiopure starting material but never got round to it.

I am also very skeptical of this deprotonation, especially given the pKa of the H flanked by 3 carbonyls is probably somewhere around 8!

I very much doubt that this would even exist exclusively as the tricarbonyl (this is a paper from the 1960's not many high res NMR's around in those days), more likely one of the ketones would predominantly enolise. I have seen enol protons from these types of substances with chemical shifts of δ=19 ppm, that's a pretty desheilded proton!

No one is more skeptical than me about this mechanism, what I have put here is the best anyone could come up with and I think "dubious" is an understatement.

No one I have spoken to has been able to suggest anything else though, in fact most people question whether or not I have written the structure incorrectly!

It is a very academic excercise as the end prodcut was a seven step reaction with an overall yield <1% but the authors must have had some reason to believe that this would work (which of course it did) as it is step 5/7.

I am hoping for a better suggestion from someone on here. 

[opsomath]

So I am thinking elimination of OH- to form a αβ-unsaturated compound, then the OH- adds to the ketone on the left (the one that starts off with the alpha-hydroxy) to form a hydrate/gem-diol. The gem-diol gets deprotonated and rearranged, one electron pair comes in and forces the OH group to migrate to same carbon that initially had an OH group, pushing the electron pair in the double bond into the ester carbon in a concerted step.

Crazy, I know.

[Dan]

Hmmm. The only thing I thought of is anionic methyl migration. If you deprotonate the tertiary alcohol and use the negative charge to do a 1,5-methyl shift onto the ester (think benzilic acid rearrangement, but 1,5 instead of 1,2), then you get a methyl ketone that is an aldol addition away from the observed product.

I have never seen such a 1,5-shift documented though. Just thought I'd throw it out as an idea.

[Archer]

I have never seen such a 1,5-shift documented though. Just thought I'd throw it out as an idea.

That's much easier to swallow than what had been suggested to me previously with the deprotonation mechanism.

It's a big jump acros the molecule, it could go in two steps 1,2- then 1,3-. It was  a very low yielding reaction so anything is possible. I would love to know what the rational was for deciding this route though. I can't believe that they would have predicted a 1,5-shift in their retrosynthesis!

[Dan]

Actually, now I think that stereoelectronically my proposed 1,5-shift is complete nonsense - I can't see a transition state that makes sense.

This is an interesting reaction - I will think about it a bit more and see if I can make any sense of it.

[Archer]

*Grammar edit*

I included this reaction in my PhD thesis, my external asked about the mechanism and when I said I had no idea he asked me to make sure (for my corrections)  that I had drawn it correctly.

Of course I check but because it had stumped my external and my internal examiner I have been thinking about it ever since.

It has been nearly 8 years and still I am thinking about it. I know the authors retired many years ago so I can't ask them. I would be so pleased to finally put this one to bed.

[Archer]

Ok so now I am pushing the boundaries of my knowledge on sigmatropic shifts but the 1,2- thermal shift to the adjacent carbonyl is feasible. From there, if the ester enolises then a subsequent thermal 1,3- sigmatropic shift would also have a feasible transition state.

I am on my phone now so I have hand drawn it but the orbitals of a methyl group should allow this to proceed, in principal.

What do you think? Or am I way off with my orbital theory (it's been a long time!)

[Archer]

Looking at my drawing I realise I have inadvertently drawn an ethyl group in the 1,3- transition state. Below is what I think the transition step for the 1,3- shift would be.

[magician4]

maybe helpful, maybe just rubbish ...:


you know how it is if you" feel " something, but just can't get a grip on it?

... that's what just happens here : Ireland-Claisen and Cope rearrangements keep on yelling at me to be considered, but I just can't make the molecule behave right, even if I monodeprotonated it first, or if I started with the en-diol form of the acyloin substructure...


... but maybe someone else can?


regards

Ingo

[Dan]

Yeah, this is looking better. I have changed my mind again after a bit of pericyclic revision.

Actually, the 1,5-shift I originally suggested, which is a [1,4]-sigmatropic shift, is thermally allowed by FMO - so perhaps not nonsense...

A 1,2-shift seems reasonable. The following [1,3]-sigmatropic shift is also OK in terms of FMO theory.

I have attached a scheme of what I see as theoretically plausible pathways (HOMO C-C σ; LUMO alkene/enone).

Now, the only examples of these kind of sigmatropic shifts with carbon I have seen involve C-C σ bonds involved in cyclopropanes or cyclobutanes - precedent nevertheless.

[Archer]

That'll do for me! Thanks for your input.

My FMO theory is very rusty. I always wanted to read Fleming's book on FMO's but there are so many books I need to read on something and not enough time to read them all. I think I should start with Fleming's OCP book on pericyclic reactions.

[Dan]

Yeah, I always found pericyclic chemistry the hardest area of organic. Fleming's OCP is very good, but I preferred this one and keep a copy of it at home.

[Archer]

That's the one I meant, I thought the OCP might complement it as I need to blow the cobwebs out first.