Home » Still In The RBF

Microcladallene B   

16 May 2007 5,938 views 38 Comments

D. Kim, Park, B. Kim, H. Kim, and D. Kim. ACIEE, 2006, Early View. DOI: 10.1002/anie.200700854.

Again with the medium rings! And what a target – I’m still surprised to see that bromo-allene in a natural product, but nature is known for her surprises… This time, however, the focus isn’t really on the eight-ring, but on the six, with a slightly tricky bromide to introduce.
The starting point for the synthesis was a Grignard addition into an aldehyde, setting up the stereochemistry from a chiral aldehyde. Formation of the ether then left the SM, with which they intended to form an enolate and then add to the allyl aldehyde. However, they had a little difficulty choosing the correct protecting group for the alcohol in some model studies.

Some work, however, resolved that the best protecting group was an anion1, and thus forming the dianion and then adding the electrophile resulted in sucess (6:1 anti/ syn). With this material in hand, formation of a second ether with ethyl propiolate, and RCM completed the 8-ring, ready for a funky samarium annulation. Awesome!


Next came a rather tricky step – displacement of the free hydroxyl with bromide with retention of configuration! Doing this with inversion (S N2 style, normally using CBr4 and P(Oct)3) is a hard reaction, so this now looks like quite a challenge. However Kim, along with many others, has read the NALG work of Lepore with some interest, so they gave it a go:


Nice work! Cleavage of the silyl ether was another benefit of this procedure, leaving them only a few more transformations to complete the synthesis. However, formation of that bromo-allene was also nice, so I’ve thrown that in for good measure! A top synthesis of an appealing target, this is a great read.

1. To me, this is known as the “Foxy protecting group”, as the academic who taught me it’s benefits is one David Fox.

1 Star2 Stars3 Stars4 Stars5 Stars (No Ratings Yet)
Loading ... Loading ...


  • The Canadian Chromatographer says:

    I couldn’t agree more. That’s one of this year’s highlight. Not extremely complicated, but just plain impressive. You read the paper, and you are in some state of bliss.

  • kiwi says:

    never seen that NALG stuff before, it’s quite clever

  • milkshake says:

    Lepore is great guy – very nice, patient and shy kind of guy; I would love to have him as an advisor. He managed to keep his small group functional, with meager resources at an disfunctional university. His one lab was like oasis of normality there, and everything he had in the lab there was hand-made and hard-earned. I hope he could transfer to a better Chemistry department.

  • Tynchtyk says:

    I wonder if the authors of this publication are relatives or is it just a coincidence that they all are “Kim”s

  • Geoff says:

    Very nice synthesis! I like the NALG step, it’s nice…

    I think the DOI isn’t right in the post Tot. Syn…

  • pi* says:

    I think you’ve got an extra methyl group on the allene

  • neo says:

    Anybody, any ideas on how the dianion chemistry works.

  • willyoubemine says:

    the more reactive anion, the second one formed, attacks the allyl iodide. thus, the alcohol is protected by the fact that is the less reactive center, as if it were a TBS ether or something.

    Also, you might expect some chelation between the alkoxide and enolate giving enhanced diastereoselectivity.

  • willyoubemine says:

    not to change the subject, but has anyone ever observed lithium halogen exchange when using allyl bromide?

  • Jose says:

    I haven’t seen the NALG manifold before, very slick. This is just out in Tet.

    Recent advances in heterolytic nucleofugal leaving groups
    Salvatore D. Lepore and Deboprosad Mondala

    Received 2 March 2007. Available online 13 March 2007.

  • HB says:

    “neo:” The principle is simple. The most acidic labile proton (alcohol) is deprotonated first, and the least acidic labile proton (pseudo-enolate) is deprotonated second.

    Since the pseudo-enolate was formed second, it is less stable than the deprotonated alcohol, and therefore more reactive – so it reacts first and consumes all the allyl iodide before the alkoxide can get to it.

    It’s really kind of a misnomer to call it a “protecting group,” but I can see the logic of that as well.

  • Hap says:

    How does this manage to go with retention only? The paper doesn’t really suggest a mechanism (the metal and the chelating group accelerate intramolecular delivery of the nucleophile) – since there are neopentyl substrates used, it probably doesn’t go by Sn2, and Sn1 doesn’t seem to make sense for these substrates. It could be using a radical anion mechanism – if the LG stays close (close ion pair), then the nucleophile will only be delivered from where the LG is and thus with retention – but I don’t know.

  • neo says:

    HB and willyoubemine : Thats cool ! Never really thought of it. I actually drew a 3D structure (energy minimized) and concluded that the alcohol may actually be doing some chelation and holding the enolate in a particular fashion to get the high de.

  • The Dude says:

    This NALG business is all old hat. It has been known for over 100 years that thionyl chloride (for example) effects replacement of secondary alcohols to secondary chlorides with retention, as seen in Walden’s original paper : Berichte 1899 vol. 32 p. 1855 and references therein (or page 420 in your handy March vol. 5.) Ion pair mechanism.

  • Hap says:

    That’s what Sni means – I thought (for some reason) it was a generic substitution mechanism rather than an ion pair mechanism. Sorry.

  • neo says:

    Well all this time I thought that the “i” in SNi is for internal.

  • willyoubemine says:

    it is for internal…

    not to go all FMO on anyone, but for this “retention” mechanism, what orbital is there for any Cl to attack? Other than either a double inversion, or Sn1 mechanism (with memory of stereochemistry), how can you explain the retention.

    I dont really buy this Sni stuff…even in ion-pair rationalized mechanisms there is some loss of stereochemistry. In the Lepore work, they confirm retention via NMR. If there is no CD, optical rotation, or even HPLC done, how can you be sure of complete retention.

  • willyoubemine says:

    Incidentally, in looking at the “Sni” reaction in this above paper, neighboring group participation can be invoked to explain the “retention” or double inversion. It makes more sense (to me at least) that the TBS O attacks the sulfonate, giving the 5-mem ring oxonium, which is opened by Br. This double inversion mech makes a lot more sense than Sni. It also explains the TBS cleavage.

  • Hap says:

    Sorry about the i.

    I don’t know if any of the missing 37% is characterized, but wouldn’t you expect some of the tetrahydrofuran product in that case? (loss of the Si from the oxonium ion should be easier than attack of Br- at the six-membered ring on the basis of steric hindrance). I would figure that the cis-tetrahydrofuran would be fairly stable.

  • TheEdge says:

    What Hap said. Also, it’s a TBDPS, which is huge, and makes the oxygen much less likely to provide anchiomeric (sp?) assistance.

    I’m impressed with the 63% they managed to isolate. TiBr4 at rt isn’t exactly mild, especially with a large number of Lewis basic sites in the molecule.

  • synthon says:

    youwillbemine…maybe this paper by Moss, Sauers and Wipf will help: http://dx.doi.org/10.1021/ol050185k

  • neo says:

    Correct me if I sound insane. I think that the oxygen on the pyran ring may do the NGP thing (via a 3 membered TS ..oxonium ion formation) and then the nucleophile attacks to give a retention of configuration.
    Secondly, if the attached sulphonyl group makes the oxygen part a really good nuleofuge then may be we have an SN1 type SNi which would then occur from a side of the samllest group on the alpha carbon i.e the “H” to give the compound as shown in the paper.

  • Hap says:

    If there is participation by the pyran O, then you might/should see ring contraction products (8.5 fused ethers with an a-Br substituted side chain). Related reactions happen with 3-(halomethyl)pyrrolidines. If that’s the case, though, conventional leaving groups might work.

  • neo says:

    Well, the sulphonyl group makes the oxygen a better leaving group.. dosent it . Also, if the 3 memebered TS (that I was talking about in my previous comment) would have to open up to give ring contraction product it woiuld mean that the Br will have to attack a hindered carbon (out of the two carbons available). I think that what you are saying also makes sense (example of pyrrolidinones) but that may be the uncharacterized compound ;)

  • provocateur says:

    Its a neat synthesis..I was just wondering if instead of the samarium-thingy, a carbene or cyanide anion mediated attack on the aldehyde carbon would result in a umpolung on the aldehyde carbon(a cannizaro type intermediate) which will do a Michael addition on the acrylate ester(I hv not drawn a T.S. to see whether it gives the right diastereoselectivity but what the heck)…but I realise it wont be as cool as the samarium..

  • provocateur says:

    ah..the name is the stetter reaction for the previous post..

  • Monkey says:

    The Stetter would give a 1,4-dicarbonyl product, so you would still need to reduce the resulting ketone in a second step…

  • willyoubemine says:

    that paper was interesting, but there are two flaws. first they assign the stereochemistry based on “computationally” determined optical rotation, which is preposterous as a stereochemical proof. second, they take the facts that there are differing levels of racemization, and retention, based on solvent to conclude that Sni is operating. That tells me that ion-pair cation formation is predominating.

    My biggest obstacle to accepting the Sni mechanism is there is no orbital for the nucleophile to occupy. Without cation formation, or backside attack, there can be no bond formation. Simple as that.

    We may have to agree to disagree on this, but is there anyone out there that can tell me which vacant orbital the electrons of the nucleophile begin to occupy in Sni?

  • synthon says:

    my thought in posting that paper wrt the orbital location was the figure…doesn’t look like a sigma* I would draw. Regardless, I don’t think you (or I) will ever be completely comfortable with the explanation given. My feeling is this falls under the same guise as “is there such thing as an empty orbital?”

  • The Canadian Chromatographer says:

    Related to post #19, I’m wondering, for the NALG bromination step, if essentially, TiBr4 doesn’t first deprotect the OTBDPS, the OTiBr3 nucleophilic moiety displaces the sulfonate leaving group, then bromide kicks in, therefore affording overall retention via double inversion… That could “easily” be probed on a “simple” alkyl-substituted analogue.

    As if anything is “easy and simple” in total synthesis…

  • willyoubemine says:

    good point, I guess we take so much for granted in terms of orbitals and their meaning, that we forget that they may not even exist. This may be a topic for the
    “Totally Existential” blog, but still interesting stuff.

    And CC’s rationale is pretty good too.

    somedays its good to be a dork.

  • chemistry is hilarious says:

    willyoubemine: My copy of March (5th edition) describes SNi as basically a tightly formed, rapid ion-pair mechanism. In talking about the retention of stereochemistry associated with using SOCl2, it says that the first chloride is displaced by the alcohol (yielding HCl), and then the RO-SO-Cl compound (alkyl chlorosulfite I think?) dissociates to give a tightly bound R+ -O-SO-Cl, which delivers the Cl- on the same face due to the speed of the process and the proximity of the carbocation. I think what the NALG stuff is doing is chelating the Br source in close proximity while the SOCl2-type chemistry does its business.

    Now, I don’t know much about SOCl2 mechanistic research, but would all that imply that SNi reactions actually become LESS stereospecific as the temperature decreases, since the lower reactivities give more time for the ionpair to drift further apart?

  • milkshake says:

    except that things tend to drift apart slower, at lower temperatures…

    The easiest test would be dependence on solvent polarity. Obviously you want some non-coordinating solvent (as to not to change coordination sphedre on Ti) but you can for example compare chlorobenzene or methylcyclohexane or even CCl4 selectivity with situation in DCM (which is quite polar). DCM should help the ion pair discociation because it can stabilise the charge separation by its dipole. So I would expect a small improvement in selectivity in MeCyhex or CCl4 or chlorobenzene.

  • aa says:

    as far as orbital overlap goes, it is important to remember that part of the sigma-star orbital (the small lobes) lies between the C-LG bond ie in the same space as the sigma orbital (in the same way that part of the bonding sigma orbital resides on either side of the bonded atoms). Therefore, it is possible to mix the HOMO of the nucleophile with the LUMO of the C-LG bond in a non-backside attack manner. of course, the overlap is much poorer, but it cannot be ruled out completely. especially in the NALG cases where the polyether is directing the nucleophile by chelation of the cation, this overlap can be maximised.

    an analogous example is the case of anionic inversion in SN2 reactions of boc-stabilized lithium anions. Although the configuration of the anion is fixed, the nucleophilic attack can occur with either retention or inversion depending on the electrophile. This occurs through overlap of either the large (retention) or small (inversion) lobes of the C-Li bond with sigma-star of the electrophile. This is described fairly well in “Frontier orbitals and organic chemical reactions” by ian fleming.

  • spottospot says:

    To invoke an orbital picture one must pick the right orbital picture. The starting material’s orbital picture is not important here. It is the orbital picture of the intermediate carboncation, which is in a solvent cage with the sulfonate coordinated with the TiBr4 (which elicited the separation in the first place). The carbocation reacts with a bromide that dissociated from the sulfonate-titanium bromide complex, on the same face, to give the retention product.
    I think that the spectulation about TiBr4 removing the TBS and participating in the rxn is not unreasonable. But as the link above, to Lepore’s work, indicates that the rxn does not need neighbouring group participation.

    Tot syn the substrate that undergoes the ketyl radical 1,4-addition cannot be the acetal, but has to be the aldehyde. Otherwise the product would be the ether and not the alcohol.

  • Spiro says:

    Hey all,

    I would like to bring an important fact to your attention. The SN reaction takes place alpha to an ether. This kind of substrate undergoes a very strong destabilisation in the transition state (electron withdrawing from the ether). In ref 13 the authors state that they first tried to displace a mesylate. And not surprisingly they had to reflux in toluene to see something happen (= a mess). A triflate would have been more appropriate. BTW this kind of triflate would be stable and could be chrommatographed.

    The Lepore leaving group worked well for them because it is a great leaving group in presence of metal halides. HOWEVER I cannot imagine an ion pair in this transition state (even a tight one). IMHO the reaction is an SN2 (the SNi’s observed by Lepore are done on substrates having a straight apolar environment).

    How to explain the stereo then? For the transformation 3 -> 2, the authors state “the prefered geometry of the transition state for this cyclization is probably described by B”, which I translate by “it is the contrary to whate we expected that happened”. With 2 having the wrong config, and an SN2 in the NALG transformation, you get the correct structure.
    Cpd 2 must be difficult to assign with all those H alpha to an ether/hydroxy…

  • TheEdge says:

    Spiro- So you’re saying that they incorrectly assigned structure 2? Interesting. It shouldn’t be that difficult to figure out which protons are which, and the difference in the coupling constants for the cis and trans compounds should be relatively large. And hopefully they have NOESY data, too.

    Let’s say they did screw up their assignment. Now, in structure 2, you’re going to either have a syn-pentane interaction between the free hydroxyl and the 8- membered ether, which should make it really difficult to put on a large leaving group like the NALG, or you’re going to have a syn-pentane interaction between the other two substituents and the incoming nucleophile. I am, admittedly discounting any boat configurations, which should be unstable anyway.

  • Smitty says:

    This a a few days late, but it is possible one of the poly-ether oxygens off the aryl sulfonate does an initial sn2 displacement of the sulfonate, forming an oxonium ion, then bromide comes in to kick that out. Obviously, ethers aren’t extremely nucleophilic, but it would certainly speed up the addition of the bromine (relative to the mesylate). I really don’t think the pyran oxygen is taking part (neo). The intermediate oxonium ion looks really really unstable. And if assistance from the ether side chain doesn’t happen, my money is on what CC suggested – deprotected TBDPS ether coming in and doing it.