Home » Still In The RBF


2 July 2011 19,319 views 23 Comments

Trauner, Cakmak, Mayer. Nat. Chem., 2011, 3, 543. DOI: 10.1038/nchem.1072 Article PDF Supporting Information Group Website

Simple tricyclic morpholine derivative with four stereocenters. Can’t take that much effort, can it? Well, yes it can – and it’s amazing how tricky this little beastie is! Trauner explains in the informative introduction that only one successful synthesis has been completed – taking twenty steps to get to the racemic target. Clearly that leaves the team in Germany with a lot of room for improvement.

The team made the first intermediate in the scheme above by using a bit of metathesis to form that eight-member ring. The diol stereochemistry was unsurprisingly installed using a bit of Sharpless’ chemistry – in this case a desymmeterising asymmetric epoxidation, followed by opening of the allylic epoxide to install the amine.  Treatment of the diol with a thionyl chloride then results in a slightly unusual reaction (I’d perhaps expect allylic displacement and installation of a chloride), forming a cyclic sulfite.  This is still a rather neat leaving group, as treatment with lithium azide results in formation of the azide – apparently completely selective for the allylic azide.

At this point it helps to consider the third-dimension – and I’d have illustrated it with a bit of Chem-3D, but there’s something wrong with my installation.  The point I’m failing to make, though, is that the structure of the medium ring is heavily puckered, mean that the secondary amine is actually quite close to the olefin.  Thus, addition of a little bromine in methanol allows formation of the bromonium ion, which is trapped by the amine to ultimately give the bicyclic tertiary amine after neutralisation of the salt.  Interestingly, the neutralisation was conducted with lithium chloride present, which does the Appel reaction thing, displacing the bromide for a chloride with inversion of stereochemistry.  Whether this reaction really is an sN2 is a matter of some consideration – but not right now…

What to me is a little amazing is that the azide group seems remarkably unconcerned with the reaction conditions used so far.  Even more surprising is that heating it up in a microwave is also tolerate, meaning the group can complete the morpholine ring by intra-molecular displacement of the chloride by the proximal hydroxide (again, try to consider the ring pucker allowing the hydroxide to easily attack the secondary  chloride.

Examination of the target shows that the group was almost there – reduction of the azide by hydrogen / metal led to Temuline (or norloline), whilst the same reaction in the presence of Boc anhydride results in the protected derivative, which was reduced to Loline via hydride reduction.


1 Star2 Stars3 Stars4 Stars5 Stars (11 votes, average: 2.73 out of 5)
Loading ... Loading ...


  • BRSM says:

    The 20 step synthesis that Trauner mentions in the intro is actually asymmetric; the other, older, synthesis he references is the racemic one. Is the displacement of the bromide to the chloride an Appel reaction? In my mind that’s the formation of a bromide or chloride from an alcohol using CBr4 or CCl4 and PPh3. If I had to call it anything I’d side with Trauner and call it a Finkelstein, although the reaction conditions are far from classic.

    • Stef says:

      yeah nice piece of work, I agree on the bromide-chloride exchange, it is not an Appel-reaction. Appel would be exactly what you describe. I would also say it is somehow a Finkelstein but probably we should just call it a nucleophilic substitution…

  • nitrosonium says:

    sometimes you have to watch out with those allylic azides….especially with heat!!
    nice to see it survived in this case.

    • milkshake says:

      I agree but in this case they already got to non-allylic alkyl azide before heating. Alkyl azides with 6+ carbons are typically very stable, surviving reflux in methoxyethanol. Its the aryl azides and acyl azides that are most sensitive, falling apart at temperatures around 100C. I remember Sharpless once made that distinction when describing how his group had few explosions and it was always with aryl azides or nitro compounds.

      • gippgig says:

        They had to heat the lithium azide-sulfite mixture to make the allylic azide in the first place.

        • ... says:

          Allylic azides are prone to [3,3]-sigmatropic rearrangements. As gippgig mentions, in the supp. info, the azide introduction was conducted using LiN3 in DMF at 130 C, and Trauner says they observe no SN2′ products! Its likely that the displacement is an SN2′ process and then that allyic azide rearranges via a [3,3] to the thermodynamically preferred allylic azide they isolated because there can be an intramolecular hydrogen bond between the alcohol and the azide. This hydrogen bonding effect in allylic azides has been documented, see: Trost, TL 1995, 36, 8737.

          • PotStirrer says:

            @… If the initial displacement by azide was an SN2′ process, one might expect it to be “syn” with respect to the sulfite leaving group (while syn stereospecificity of SN2′ reactions is not always seen, I think it is rare if not unknown to see complete anti stereospecificity). The resulting [3,3]-sigmatropic rearrangement would then most likely occur on the same face of the ring, giving the epimer of the desired product, with the azide and hydroxyl syn with respect to each other. Because Trauner et al. get such a high yield of their desired product, I think this argues reasonably well against an SN2′ mechanism.

          • ... says:

            @ PotStirrer
            There is nothing wrong with an anti SN2′ process, especially if the stereochemistry of the molecule dictates it (very common in compelx molecule synthesis). However, I’m not saying that in this system there is any obvious reason from looking at a model that an anti-SN2′ process would be favored. I’m not sure why you think one should expect the SN2′ to be syn to the leaving group. Do you have some reference to back this up? Regardless, there is no doubt that allylic azides undergo the [3,3] (again the ’95 TL paper and references therein). Even if the substitution was 100% SN2 (as Trauner suggests), at 130 C in DMF, one would expect the [3,3] to occur. The fact that they only observe one of the two possible products can be explained by the preference for the isomer containing the hydrogen bond (again the TL paper). All I’m saying is that the SN2′/[3,3] is another reasonable explanation, especially since even if it is purely an SN2, the formal SN2′ product could have been observed from a [3,3]. These possiblilities should have been discussed in the paper, and the appropriate literature cited.

          • PotStirrer says:

            @… I agree that [3,3]-rearrangement may be occurring and that the thermodynamically more stable allylic azide is therefore isolated. I also agree a more thorough discussion of the mechanism was warranted. My point is simply that it seems unlikely the SN2′ is exclusively anti, which it would have to be to give the observed product. I’m no expert in this area, but from reading this review (Tetrahedron, 1980, 1901-1930) it appears as if the majority of studies have shown a strong or slight preference for syn SN2′ over anti. Those examples where anti SN2′ was exclusive were generally intramolecular cases with no chance to go syn. Do you think there is an overwhelming preference for anti SN2′ with Trauner’s molecule? If not, then I would predict at best a mixture of syn and anti addition, which should result in a mixture of epimeric products after [3,3]-rearrangement (who knows, maybe they got a small percentage of the epimer and discarded it, but they didn’t report that). Because there is no question about the stereospecificity of SN2 reactions, I support initial SN2, followed by [3,3] “equilibration” which ends up giving the same product anyway. I would be interested to know of papers where exclusive anti intermolecular SN2′ has occurred (non organometallic/catalyzed examples).

    • nitrosonium says:

      thanks for the clarification, milkshake! i just got finished wading though an awful mess here in my facility with warming (40 C) an allylic azide in DMF. This was a terrible experience (obviously the synthesis did not work out). I saw an allylic azide here and later saw heat…..merged steps based on current negative outlook on the synthesis scheme i had “brilliantly” put together in my head.

      just love talking/thinking about synthesis!!

  • Young Padawan says:

    Hey TotSyn. Good to see you back!!! Nice post, too.

  • IrII says:

    RE: whether the reaction really is an SN2

    I am not sure what the alternatives are (I have not read the paper – is this discussed?):

    having that adjacent doubly substituted carbon is going to make it more hindered (OK); that one of the substituents is (electronegative) nitrogen would disfavour both SN1 and SN2 – presumably the NaOH is there to deprotonate the ammonium salt (having the nitrogen protonated would be less favourable than unprotonated for SN1 mechanisms. For SN2 I don’t know – alternative explanations for the soi-disant beta-oxygen effect give contradictory answers: if interaction with a negatively charged nucleophile pre-attack (populist explanation) is the dominant deactivating factor, then having a positively charged nitrogen beta should actually enhance SN2, but if the reason behind beta-oxygen is truly the disfavourablisation of delocalisation of electrons into the leaving group, then this should indeed be worse with the positive nitrogen).

    Maybe the question is why there is no equilibration of bromide stereochemistry (due to the presence of bromide anion) prior to chloride substitution, and in that case there are two answers: i) it is, or ii) what is the solvent. (although in the preceding (cyclisation) reaction an alternative explanation based on the so-called beta ammonium effect may be relevant.)

    OK I give up, why is it not SN2?

  • JM says:

    I think there are two reasons for the Finkelstein to proceed exactly like this. 1) is an steric argument, because the possible reattack of a bromide has to proceed from the “downside” of the [3,3,0] bicycle, which is very unlikely. 2) from 3D-modell one can assume that the C-Br bond might be activated by a stereoeletronic effect. The C-Br s*-bond looks quite parallel to the C(tert)-N bond, which is not the case after Br-Cl exchange an inversion of the stereocenter.

    thanks for beeing back tot. syn. nice post about a nice synthesis.

  • Young Padawan says:

    Hey out there,
    How can I get into Tenderbuttons archive’s? Anyone? I often find some links to it, mostly here or in the pipeline…

  • john says:

    Does any one know a way to suppress the chlorination in Tosylation of alcohols with TsCl/MeLI ( I do not to use any organic bases since they do not push the tosylation to completion of my hindered alcohol.

    • HPCC says:

      a) Did you try KH or NaH? Although if you’re sure you’re quantitatively deprotonating with MeLi, maybe just play with the tosylating agent, and leave the base alone.

      b) There’s Ts2O, whose leaving group is TsO-, and totally non-nucleophilic. It’s fairly pricey, probably 20x the price per mole of TsCl, but it would solve your problem. However, I’ve never tosylated a tertiary alcohol, so I don’t know whether Ts2O would be electrophilic enough for your system.

      I’m also thinking silver salts. There must be some prep out there that adds 1.1 equiv of AgOTf or some other salt to scavenge the evil halide…

      Good luck!

    • xx says:

      While it’s not a protecting group, Greene’s protecting group book has a nice section on tosylation.

  • john says:

    Thanks HPCC, and XX.

    I am trying silver salt and checking the luck.

  • asian chemist says:

    Can you use mesylation instead, if the alcohol is hindered? Have you tried potassium tert-butoxide with TsCl? I found this combination quite powerful personally.

  • Imardio says:

    Of course it is not an Appel, but a Halex reaction (call it Finkelstein if you want).
    A nice piece of work… and I’m still impressed about the lack of explosions coming from that azide… Or maybe is this what Sharpless calls ‘Azidophobia’?