Home » Still In The RBF

6-deoxyerythronolide B   

30 August 2009 16,670 views 42 Comments


White, Stang. Nature Chem., 2009, AOP. DOI: 10.1038/nchem.351. Article PDF Supporting Information Group Website

Nice to see a bit of total synthesis in Nature Chemistry this month – but to be fair, this is more about the organometallics than the macrolide.  But if you’re going to prove some methodology – especially, as in this case, macrolactonisation, this is a good target to work on.  6-deoxyerythronolide B (6-dEB) is a typical polyketide macrolide, and I mean that chemically, as the configuration of the stereocenters is very much the norm.  This meant two things – 1) the group could use typical polyketide chemistry to build the stereocenters – and 2) this would serve as an excellent example for macrocyclisation chemistry.


The first task, of course, was to build the linear polyketide, unsurprisingly using an aldol-tastic route.  Working right-to-left, the group first used a Myres alkylation to append an allyl group in a stereoselective manner.  Two Evans-type boron syn-aldols allowed construction of the central chunk, whilst a further Myres alkylation allowed the C-6 stereocenter to be installed along with an aldehyde, ready for the next aldol.  This final aldol was between two larger, and quite highly functionalised units – with several competing stereochemical interactions.  Using standard conditions (see this Tetrahedron), a bit of tickle-four returned a reasonable yield of the product (49%, 7:1 d.r.).  I’d have been fairly happy with that, but the group went that bit further, improving matters considerably with Ti(Oi-Pr)Cl3.


Barring a little reduction and removal of the chiral auxilliary, the linear precursor was done.  With a free seco-acid, they were all set to macrocyclise – except they were lacking the C-13 hydroxyl with which to form the ester.  However that was whole point – to do a stereoselective C-H oxidation and form the lactone in situ.  Quite a lofty goal, but one which the group have worked on for some time.

Throwing the starting material in with a bit of catalyst (10 mol%) returned only a very low yield.  Boosting the loading, and also the concentration, allowed the group to isolate a 34% yield, along with 45% recovered SM.  Apparently, efforts to boost this result were unsuccessful, as they resorted to recycling the SM.  However, this allowed isolation of a 56% yield, and a near perfect d.r.  Damn nice work.


The paper then goes on to discuss the rationale for this stereochemical result, using a rather interesting method.  Presuming that the reaction proceeds via a macrocyclic tether, the group using a source of flouride to deactivate this selectivity.  The fluoride presumably attacks the catalyst, disrupting the tether, and is evident in the dramatic drop in diastereoselectivity.  The suggestion is that it is the conformation of this tether, and the product-like transition-state that account for the stereoselectity, bourne-out by computational studies which suggest a 3 kcal mol-1 preference for the desired stereochemistry.  When forming the pi-allyl intermediate, this complex can epimerise back and forth, with the desired conformation leading to product.


An interesting further study was to take the SM, and hydroxylate it.  The two hydro-acids were then reacted under typical (Yamaguchi) conditions; whilst the SM with the desired stereochemistry cyclised nicely (87%), the C-13 epimer did not.  It’s nice when nature works in your favour!

I should mention that a further three steps were needed to get to the close, but after this, do you really think it troubled them?

1 Star2 Stars3 Stars4 Stars5 Stars (4 votes, average: 4.00 out of 5)


  • milkshake says:

    in the Pd-oxidative cyclization scheme you are missing vinilic C=C (which they then hydrogenated into ethyl)

    The synthesis finale (after the Pd-oxidative cyclization step) is also noteworthy: 1. Pearlman/H2 to cleave off PMB acetals with vinyl C=C reduction at the same time 2. Selective oxidation of C9-OH to keto with TPAP (with unprotected C11-OH in the molecule) 3. HCl deprotection of the acetonide – that does not ruin the stereochemistry on C8 and C10

  • The Next Phil Baran says:

    Nature! Woodward is rolling in his grave.

  • earth23 says:

    How many polyketides must be synthesized before they become this generations steroids?

  • T says:

    I think this synthesis really makes one appreciate the advances in synthetic methodology over the past 30 years. Recall when Corey synthesized Erythronolide B, the synthesis of large lactones was regarded as nearly impossible. Also, only the absolute top tier chemists could devise ways to asymetrically alkylate molecules. Nowadays, with Evans Aldols and Myers Alkylations – as well as the myriad of ways to make lactones now, quite complex molecules can be made relatively quickly and efficiently.

    Methodology really has come a long way. It’ll be interesting to see what the organic reaction toolbox looks 30 years from now.

  • gilgerto says:

    I think it is Myers, not Myres…:)

  • loaded says:

    Elegant use of Evans chemistry! Same number of steps as many other approaches.

    So whats the advantage here of C-H activation?

  • Bitter After Taste says:

    So what you’re telling me is that I could (a) install the hydroxyl from the begining (readily accomplished) and get an 88% yield in the key cyclization or (b) fail to do so, use some lame surrogate and cut my yield by >30%. Yes… powerful methodology indeed for solving problems that do not exist.

    • a-non says:

      I guess if you were kidnapped by terrorists and told to synthesise 10g of this product you would not use this approach, but if this paper had never appeared how would you have known that it wouldn’t be better than the existing methodology. I’d be willing to bet that there will be countless examples where this type of approach works better than the existing alternatives.

      We all want our results to be the best ones. The thing is that sometimes it is just as interesting when the exeriments fail or fall short of the mark.

  • The Next Phil Baran says:

    Could the diastereoselectivity for the closure be the result of an equilibration of the products due to a small amount of Pd(0) in the pot? Any crosschecks indicated in the paper? I can’t access it! The immediate product of the macrocyclization is an allylic acetate, which would be labile with a Pd(O) species, any thoughts? They should have checked enriched samples of the product under these conditions.

  • TB says:

    Bitter After Taste, this is a beautiful application of White’s allylic palladium oxidation/macrocyclization methodology on a well-studied polyketide model system.

  • loaded says:

    agree with bitter. by the way has anyone run one of those allylic oxidations? they are very messy and need high palladium loadings.

    • milkshake says:

      The related Stolz oxidative cyclization methodology using Pd(py)2(TFA)2 and sieves and Na2CO3 in toluene and O2 baloon is actually is quite clean and high-yielding.

  • bob says:

    Don’t agree with bitter – Nice new method that avoids the use of some PG chemistry and closes a macrocycle with different functional groups. The more alternatives to the -yawn- macrolactonization approach the better.

    And also if you have ever tried closing macrocycles (which some of us have) you would know that an 80% yield is hardly a given with the standard protocols.

  • The Next Phil McGroin says:

    First off, I liked this paper. I thought it was a good target to showcase their method, and according to their report, the route compares favorably with those prior reported syntheses.

    But, their key step uses a catalyst loading of 30% and achieves a 34% yield of product. That 4% is certainly within expt error of being an aberration which would make the rxn stoichiometric in Pd. I am NOT suggesting any dishonesty, and perhaps I am missing something, but that seems a bit poor.

  • Bitter After Taste says:

    New methodology is not justified BECAUSE it is new. Its worth is demonstrated by transformative, strategy level simplification (if you’ve ever seen White talk, she MAKES THIS POINT and has better examples to back it up). As to the yield question,”whilst the SM with the desired stereochemistry cyclised nicely (87%), the C-13 epimer did not.” Yes, I have run a macrocyclization, and if this method DOES allow the poorly yielding ones to work better, I’ll be the first cheerleader, but if you want to show off your new method, pick a target where it radically improves the previous state of the art. This is incremental at best and frankly White’s better than that (as is Nature normally, but I digress).

  • John Wood says:

    Why do I not have access to Nature Chemistry while I am enrolled in a Chemistry Program?

    Something seem off? or am I just missing the right way of accessing the articles?

    • The Next Phil McGroin says:

      As a professor of chemistry at Colorado St., I would think you should be able to access any journal you want!

    • beef supreme says:

      Perhaps you attend a broke-ass public university in California that can’t afford a subscription.

  • stir_bar says:

    I think some schools just haven’t decided to get a subscription yet…mine included. You are still able to access the reviews from Nature Chemistry though.

    On another note, why is difficult to close these large macrocycles? Obviously it isn’t the easiest of chemistry by past methods… With normal protocols you have something like an activated acid and one unprotected alcohol to react (or the opposite activation). I don’t understand where the difficulty comes in. Is it just the large ring formation or maybe the confirmation of the final macrocycle?

  • Remev says:

    Stirbar- macrocycles like this were difficult in 1980. Now because of people like evans and myers they are easy.

    Because of yamaguchi and others closing them is easy too. Don’t forget grubbs who basically revolutionized macrocyclic formation.

  • ... says:

    Anyone who says macrocyclization is easy is very naïve and has likely never actually tried to perform one. The fact that there are so many methods developed for macrocyclization yet still it can take months of optimization on a system which is difficult to close and still only get poor yields attest to the difficulty of macrocyclizations. Stir-bar, the challenge with closing large rings is typically an entropic problem (see Chem Rev. 2006, 106, 911 for a nice review). A significant reduction in the degrees of freedom available to the uncyclized molecule occurs upon ring closure. Because of this, intermolecular reaction (dimerization) is typically preferred. However, intermolecular reaction can be overcome by running the reaction under high dilution conditions. Finally, steric effects play a critical role in ring-closure, and therefore, your success in performing your macrocyclization in high yields in a month vs. a year will depend on each individual case.

    • The Next Phil Baran says:

      Entropic problem! What the hell. So there are too many degrees of freedom with a macrocycle, or the open chain? The entropy contribution to the thermodynamics of this process is minor at best The problem is getting the two ends to meet, which is compromised by the backbone. There is no good feel for how these molecules sit in solution. There may be only one conformation it sits in because of all the substituents present, therefore not an entropy problem. The change in entropy between the open chain and macrocycle is tiny compared to other relevant energies.

      • PotStirrer says:

        NextPhilBaran. You are most likely correct in most cases when you say “the entropy contribution to the thermodynamics of the process is minor…” and “the change in entropy between the open chain and the macrocycle in tiny….” However, entropy plays a very important role in the kinetics of macrocyclizations. This is the problem of “getting the two ends to meet” as you say, and is reflected in the entropy of activation for the macrocyclization. When you dismiss entropy by saying that a long chain molecule may exist in solution in only one conformation is really ignoring a lot of evidence to the contrary. We should all know that molecules are highly dynamic and most are constantly undergoing conformational changes even though they may not look like that on the NMR time scale. Do you suppose that your precursor’s presumed one conformation is so close to the transition state geometry for cyclization that no entropic restrictions must occur? A good reference: Acc. Chem. Res. 1981, 14, 95.

  • UBChem says:

    Paul – please blog MacMillan’s minfiensine synthesis. I really want another conversation on a paper being JACS worthy.

  • European Chemist says:

    I would have to agree with McGroin. It was a nice target to exemplify their methodology, but… it just didn’t work out as planned. This is essentially a stoichiometric reaction. I think that it is the boldness in pursuing what is clearly a risky approach to the natural product which deserves highlighting here. Don’t forget that one of the first uses of RCM in a successful total synthesis (Hoveyda, Fluvirucin B1) was also an extremely risky plot – forming a trisubstituted alkene in 1994 with (by today’s standards) the first rudimentary carbene catalysts. I can’t help but thinking that if that one hadn’t worked out at all it might have taken a bit longer for that Nobel Prize… This particular case just shows that White still has work to do on these systems before people will really start using them.

  • Erlenmeyer says:

    White points out in her paper that one reason for this approach is the possible versatility of being able to form both ester linkages at a late stage, rather than having to go back to the beginning to reset the stereocenter. Another point she makes is that at least in cases like this, it’s easier to carry around an olefin than a protected alcohol which requires a good orthogonal protection strategy. Not saying that all her claims are legitimate, but these are a few of the rationales that haven’t been mentioned yet.

  • Knut Jacksworthy says:

    I don’t care what people say, macrocyclizations can be excruciatingly tough. You screen countless conditions, get a 30% yield and are rewarded by people saying “well that was obvious”. Look at Trost’s Soraphen A for instance.

    Performing an allylic oxidation as a macrocyclization step takes balls.

    • European Chemist says:

      It sure does.

      • loaded says:

        Yes, it takes balls OR access to Scifinder. Macrocyclization via Pd-allyl complexes are as old as dirt. see for example (of dozens out there): JACS 126, 11254.

        • Tok says:

          Generating a pi-allyl from an allylic carbonate as in your reference and doing an allylic oxidation from an un-functionalized alkene as in what White does are two very different beasts.

  • ian says:

    something NEW for once in an organic chemistry publication – the optimisation will come and is being done, I’m sure.

    I enjoyed this paper and good on them for having the guts to try it

    Or we can all sit around and cry as competitive grant money is wasted on aldol-aldol-aldol-Yamaguchi/RCM = macrolide. :cry: :roll:

    • European Chemist says:

      Well, they spent theirs on Myers alkylation-aldol-aldol too, before the key-step…