Home » Still In The RBF

Biyouyanagin A   

25 May 2007 10,019 views 33 Comments


Nicolaou, Sarlah and Shaw. ACIEE, 2007, Early View. DOI: 10.1002/anie.200701552.

I wasn’t planning on putting this synthesis up today; in fact, I’ve got a tasty number from Steve Ley to put up. However, one glance at the palladium cascade in this paper left me with my jaw on the floor, and I’d started transcribing the structures before I knew it! So, a quick look at the biological activity would be in order: selective inhibitory activity against HIV replication in H9 lymphocytes (EC50=0.798 μg) with a TI of about 31.3, along with lipopolysaccharide (LPS)-induced cytokine production (but less potent). Not stunning, but interesting none the less. On with the retro:


Perhaps the photo [2+2] is a quite an obvious disconnection, but using palladium to build the spiro bis-lactone is interesting… So we split the molecule into two fragments – the bis-lactone and the cyclohexene. The cyclohexene was built quite simply, using a pair of aldols on methyl-vinyl-ketone to build a cyclohexenone. The initial aldol used a proline-derivative, giving decent diastereomeric control.


However, it’s the other fragment that interested me: to me it seems to be a carbonylative Sonogashira coupling, followed by carbonate formation and attack into the acetylene. Then, palladium does it’s Ï€-allyl thing, and extrudes carbon dioxide as a top leaving group. Then, rearrangement of the palladium-enolate type species gives a palladacycle which eliminates to give the spiro bis-lactone.

How did they see that as a disconnection?!

So, with the fragments complete, it was time to do the photocyclisation, which went reasonably well, giving the natural product directly. I wonder if what happened to the mass-balance? Anyway, amazing stuff!


1 Star2 Stars3 Stars4 Stars5 Stars (1 votes, average: 5.00 out of 5)


  • earth23 says:

    DOI link appears to be broken :(

  • Liquidcarbon says:

    …and why are you still in 2006?

  • Tot. Syn. says:

    Oopsie with the year! But I dunno why the link doesn’t work – that’s the quoted DOI. Anyway, here’s a direct link:

  • Lost-in-Translation says:

    They did not invent the transformation, a Janpaness group did (ref. #8).

    BTW, what is the more conventional retro-synth? Dehydration from the corresponding 1,3-dicarbonyl?

  • cheyanimal says:

    what about direct attack of the ketone onto the palladium pi-allyl? reductive elimination of C-O bonds isn’t too common for Pd 2+ is it? Also, what about the reassignment of the stereocenters that they propose?

  • aa says:

    this synth is the redemption of KCN…totally sweet. i liked the organocatalysis step to make the cyclohexenone. good to see that diphenylprolinol catalysts are good for more than making repetitive JACS/ACIE papers.

  • milkshake says:

    The Pd cyclization is quite nice, IMHO it is more like a carbonylative Heck (sometimes there is a blurry line between Heck on CC triple bond and copper-free Sonogashira). The hydratation of CC triple bond should occur where it does although using CO2 for the oxygen delivery by hemicarbonate is smart.

    As previous commenters pointed out, this lactone is itself a natural product that has been studied by japanese groups quite extensively so I think KCN did a good lit search and took the most elegant but precedented approach.

  • milkshake says:

    I should add that very many things can go very wrong with photochemistry – just look at the number of double bonds in these pieces. Then we have possiility of 3 other stereoisomes from the 2+2, self-condensation and certainly some photoinduced God-awfull rearragements of the spirolactone. I am amazed they saved such a risky step for the final step. I would congratulate them for getting anything over 10% yield in this case.

  • provocateur says:

    eventhough there are a lot of ‘double bonds’ if u notice only the conjugated ones react..and these bonds can be ‘tuned’ to react depending on the wavelength..

  • Jimbo says:

    I’ve never heard of tuning to react based on wavelength.

    The phenyl substituted enone is probably the group receiving excitation, and from there it’s just reacting based on sterics. (di-substituted olefin vs. two tri-subs.)

    It would surprise me if there weren’t even a little homodimerization. Although, that mono-sub olefin is probably pretty hindered.

  • lost-in-translation says:

    Milkshake, can you elaborate on the Heck on CC triple bond and copper-free Sonogashira you mentioned? I wonder if we go Cu-free, are we still able to called a reation Sonogashira? Are we supposed to use strong bases w/o copper?

  • milkshake says:

    probably not – that was my point.

  • Smitty says:

    Tot. Syn., what did you mean by “I wonder if what happened to the mass-balance?”? Sorry, not a criticism, just a bit confused.

  • earth23 says:

    @ Smitty #13

    The yield was only 46%, he wants to know what happened to the other 54%.

  • TWYI says:

    Amazed they even got 46% personally

  • aa says:

    #10- Photochemical reactions are frequently tuned based on wavelength. An easy and common example is using a lightbulb to do radical bromination (a relatively low energy wavelength to break a weak bond). I would guess that some selectivity can be obtained between di- and tri-substituted olefins based on the difference in energy between them. Might account for the 46% yield, which is not at all the statistical outcome.

  • McPostdoc says:

    One of my favorite examples of selective excitation in photochemical reactions is between differently substituted 1,4-diarylbutadienes: J. Am. Chem. Soc. 1973, pg. 2058. Its also a nice example of absolute asymmetric synthesis.

    But as Jimbo pointed out, this [2+2] has nothing to do with selective excitation; they’re using a quartz filter, and that’s not going to allow for excitation of the cyclohexadiene, let alone a plain olefin. The vinylogous ester is excited, and it probably attacks the cyclohexadiene over either of the isolated olefins due to mostly electronic reasons. And, of course, its going to attack the side of the cyclohexadiene that doesn’t have the methyl group.

  • Jimbo says:

    #16 — Sorry I wasn’t clear… let me clarify: between different functional groups is a no-brainer (carbonyl vs. O2 vs. olefins, I2, NBS, etc.) but not tuning excitation of ends of unactivated olefins.

  • RET says:

    Do you really think that carbonate would ionize? Isn’t it orthogonal to the pi-system? I never liked mechanisms that cleaved a bond (alpha-OH) only to reform it later?

    Looks like other mechanism may be more likely.

  • provocateur says:

    hmm..WAT do u think is happening..
    a michael attack of the iodide on the acetylinic ketone forming the allenic enolate and the enolate attacking the carbon dioxide to form the carbonate and then the palladium doing its allylic thingy and the final spiro stuff forming from the tert-alcohol being the nucleophile attacking the pi-palladium intermediate..
    if iam on the rite track also this mech(if its a possible one)contains a problematic stuff..co2 acting like an electrophile….its a tuffone..

  • PotStirrer says:

    #20 Assuming that the carbonate is NOT perpendicular to the pi system, why shouldn’t it ionize? Allylic carbonates are terrific substrates for forming pi-allyl intermediates.

  • PotStirrer says:

    Actually, the carbonate needs to be perpendicular for ionization to occur.

  • provocateur says:

    the insertion of the co2 on to the congested tert-alcohol…hmm tht troubles me and the carbonic acid under basic conditions… ..but the conditions are not mild either

  • RET says:

    PotStirrer: The O-C sigma bonds needs to be parallel to the pi-system and perpendicular to the sigma framework of the allyl. In a small ring system a bond within a ring cannot overlap with the pi-system of a exo-methylene.

    The conversion of the cyclic carbonate to the product is simply hydrolysis and cyclo-condensation.

  • jimbo says:

    #23: it need not concern you… this mechanism was written by organic chemists, not organometallic chemists. You can probably imagine a mechanism where Pd helps that step along.

  • provocateur says:

    It does not matter who writes the mechanism.As you said you can write a mechansim with Pd-assisting the insertion.But again, in this case the intermediate seems to be too unstable.Understanding the most probable(not a possible one) mechanism always helps….

  • venison says:

    The selection in the photochemical step was masterful. The energies for pi->pi* transitions for most simple alkenes correspond to ~320nm. Nice work.

  • venison says:

    My last post was mangled somehow.
    The selection in the photochemical step was masterful. The energies for pi->pi* transitions for most simple alkenes correspond to wavelengths ~320nm in addition to a sens. Nice work.

  • provocateur says:

    I do not agree tht the selection of a 2+2 is some ‘masterful’ step.One of the first things tht u think abt when u see a 4-membered ring is a naturally intuitive 2+2 step.The rest is just getting the conditions rite…

  • arsenic says:

    provocateur…I wonder how many times have you seen conjugated diene entering photochemical 2+2. Have you maybe seen this in the field of tot. synthesis?

  • provocateur says:

    i donot hv acess rite now to any journals.I hv seen either a baran/nicolau paper pretty recently..i maybe wrong…will come back on this.

  • PENN CHEM says:

    KC spoke here at PENN about a month ago and talked about that crazy transformation. Apparently an undergrad in his lab came up with the idea during his summer research and then made it work the following year as a graduate student. He then went on to warn the audience that this student will be the next big thing. Guess we will have to wait and see but I thought it was interesting.

  • […] second showing for this complex little target, Zhixiang Xie’s route uses the same end-game as that developed by Nicolaou two years ago.  However, getting to that point is a very different matter, using some quite powerful chemistry. […]