Oseltamivir (Tamiflu) Pt. 6
Hayashi, Ishikawa and Suzuki. ACIEE, 2009, EarlyView. DOI: 10.1002/anie.200804883.   
It’s back! Not just Tot. Syn., reanimated refreshed after a decent winter break, but Tamiflu too (thanks to all those who flagged this one up for me. In this case, in a (marginally) different guise, as what we’ve got here is Oseltamivir-free base, not the phosphate salt normally isolated. Not an important fact, but the salt formation probably makes quite a difference in-vivo…
However, that doesn’t matter a damn when the synthesis is a sweet at this. To quote Hayashi, this is a ” three one-pot operations” synthesis, so over pretty damn quick; in fact, there’s only one synthesis scheme in the paper. However, to explain more of the chemistry going on, I’ve dismembered the scheme to show more of the intermediates, and give a better idea of what’s going on.
First up is an organocatalytic Michael reaction, using the groups own methodology… here and applications here. This installed the first two stereocenteres, and allows the group to to a futher, diastereocontrolled Michael addition into a vinylphosphonate. The intermediate produced is then perfectly set to do a ring-closing HWE, completing the cyclohexene. This approach is particularly neat, as most previous routes used Diels-Alder chemistry to install the ring, and then upwards of ten steps to functionalise the ring; as we see later, Hayashi is already imparted most of it. However, a flaw is evident; a mixture of diastereoisomers was produced a the C-5 center, bearing the nitro group. Neither acid nor base epimerisation was entirely succsessful, but heating the with toluene thiol and a spot of pot-carb (still present) did the job. This also ‘protected’ the alkene by Michael addition of toluene thiol, completing the first of the ‘one pot’ operations.
Next up is selective acidic deprotection of the t-butyl ester, formation of the acid chloride, displacement with azide and finally Curtius rearrangement and amide formation. Neatly done in one pot, but the use of azides is perhaps a drawback. Completion of the synthesis then only required reduction of the nitro group (using in-situ generated HCl with zinc), then elimination of the thiol to finish the target. This involved passing ammonia gas through the reaction mixture, generating a Zn(II)-ammonia complex (is that just to remove the zinc, or is it providing an active metal complex?), then retro-Michael.
Damn nice work; I’ve been picking though the experimental, and there’s really not much in the way of an industrial scale up of this synthesis that isn’t present in the main text (like the azide, and possibly the use of toluene thiol). Whether we need another synthesis that can be scaled is another story… resistance to Tamiflu is yet another… but lets just start 2009 with a cracking bit of synthesis.
Norzoanthamine
Kobayashi, Yamashita, Murata, Hikage, Takao, Nakazaki, and Kitahara. ACIEE, 2008, EarlyView. DOI: 10.1002/anie.200804544.   , ACIEE, 2008, EarlyView. DOI: 10.1002/anie.200804546.
When the synthesis is spaced over two papers, you know it’s a biggie; a true beast of a natural product, Norzoanthamine made the gloried pages of Science when it was first completed by Miyashita back in 2004. Bulging with interesting biological activity (the headline being prevention of decrease in bone weight in osteoporotic mice), med-chem work is apparently under way. Coupled with a lack of comments about ’scarcity’ or ‘low isolation yield’, I’d guess that it’s relatively easy to get hold of (though perhaps not is kilo-quantities), but that shouldn’t stop work on it’s synthesis.
Kobayashi’s approach is surprisingly linear, starting with the now rather familiar Hajos-Parrish ketone (discussed here a few months back); the easily installed asymmetry in this starting material is used to control the bulk of the synthesis. However, it has to be said that elaboration of the 6,5-fused system took quite a bit of effort; eighteen steps took them to a tightly functionalised, but still small core. In a few more steps, though, this was to change, as a rather nice alkylation bolted on a diene (nice stereocontrol, but this centre is perhaps irrelevent). This allowed a powerful IMDA reaction to occur, creating a further pair of stereocenters, and completing the A,B,C fragment. A hard slog, but with nine stereocenters, it was never going to be easy.
Next up was the fragmentation of the original penanone moiety found in the SM, now revealed. This was intended to become the D ring, a lactone, and thus an oxidative fragmentation was intended. However, things didn’t go as planned (which I’m sure involved a Baeyer Villiger oxidation), but the job was done using a pretty interesting series of reactions. Formation of a silyl enol ether using base and silyl chloride went as expected, but treatment of the enol ether with oxone ozone didn’t result in cleavage as expected, but as a single diastereomer of the α-hydroxy ketone. When this was treated with lead acetate, cleavage of the carbon-carbon bond was this time evident, but again the result was not quite as expected - rather than delivering homolytic oxidation, it appears that the hydroxy center is the focus of the oxidation action. The group postulate a mechanism for this chemistry, including a 1,2-hydride shift of the intermediate lead complex - worth a look…
Formation of the desired D ring was simplicity itself - treatment of the SM with unbuffered TBAF resulted in selective deprotection of the C-ring TBS ether and cyclisation in the basic environment. The last of the carbon skeleton was appended using a Horner-Wadsworth-Emmons olefination, and the group were soon ready to build the distictive bicyclic O,N-acetal. This went in reasonable yield, generating the desired stereocenter; the group must have been confident on this success, though, through model studies.
The D-ring, which had only been in place for a few steps, was then broken apart to allow oxidation of the C-ring centre. Then a final cyclisation, building two rings, completed the synthesis by formation of the second N,O-acetal. This is certainly tidier than the end-game employed by Miyashita, but a similar strategy.
Hmm. There are certainly some nice reactions used in this synthesis, but I feel somewhat underwhelmed. Don’t get me wrong - a synthesis on this scale is an extraordinary acheivement - but I don’t think I learned a lot along the way. Tell me what I missed!
Bryostatin 16
Trost and Dong. Nature, 2008, 456, 485. DOI: 10.1038/nature07543. 
Ah, back in to Nature for another nibble at organic synthesis for the biologists, and some slightly-odd looking structures. It took me rather a while to get used to COOMe rather than CO2Me… However, one thing guaranteed with chemistry in Nature is quality, and it certainly shows in Trost’s latest work.
The target is one he’s been working on for quite a while, but with good reason, as it’s quite a tasty number. To quote the paper, they speak of ‘exceptional biological activity’ and back this up with references to use of bryostatins in the clinic. However (and isn’t this always the stickler…) it’s not exactly bountiful in abundance - apparently only 18 grammes were isolated from 14 tonnes!! That’s quite a column…
Anyway, on with the synthesis - and the starting point, as usual, is assembly of the fragments. One of those was already discussed in a previous Trost paper on this molecule (last year), but a different, and shorter synthesis is used here. One step I found particularly interesting was a propargylation of an aldehyde using chemistry developed by Teck-Peng Loh a few years back. Interesting brew of reagents, but it’s a neat way to build this homo-propagylic alcohol. However, the result is still racemic, so an oxidation and CBS reduction was used to induce an enantiomeric excess.
For the coupling of this fragment with a (somewhat) related partner, Trost used his own funky ruthenium chemistry. In this case, the situation is complicated by the highly functionalised nature of the partners; ligation of the metal to the olefins in both SMs or the product is the reason given for the lowish yield. However, the starting materials were easily isolated, resubmitted and a decent amount brought through.
Transformation of the vinyl silane into a vinyl bromide was done with a bit of NBS - not a reaction I’d thought of as being so high yielding (98%) or selective in such a substrate. Next up was a rather sweet series of acid-mediated reactions in one pot, selectively removing the TBS group and performing a neat transesterification/ketalisation. Really, a very neat way to put in that ketal.
The vinyl bromide was carbonylated rather nicely using fairly traditional reagents to get a decent yield (83%). More traumatic to my eye was the required selective saponification of the beta-hydroxyl methyl ester. I’d probably be thinking about ways to use that hydroxyl to differentiate the esters, but no way would I have fell upon the reagent used by Trost. A bit of trimethyl tin hydroxide, and tickle with the bunsen did the job amazingly well, which Trost suggests is down to the lewis acidity of the reagent. This allowed it to be directed by the hydroxyl, and deliver the impressive selectivity. I’m again impressed with the robustness of the substrate, though - heating under (admittedly mildly) Lewis acidic conditions…
Last is a seriously impressive run of reactions - firstly, a palladium mediated coupling of the diyne, using Trost’s chemistry again. The selectivity, again, is hugely impressive, though as with most macrocyclisations, dilution was key to keeping the reactivity intramolecular. Quite an impressive example of a new means for macrocyclisation, though (even if the yield is only moderate). More impressive was the fact that they were now perfectly set up for the final cyclisation. After attempting this with a few different gold catalysts (a regioselectivity problem was apparent), they struck gold with the reagents shown (I should write the Angewandte captions…).
After that, all that remained was pivalation and deprotection; again done with apparent ease, rounding off an incredible synthesis, using impressive methodology in exactly the right way. A master class…
Oseltamivir phosphate (Tamiflu) Pt. 5
Shibasaki, Yamatsugu, Yin, Kamijo, Kimura and Kanai. ACIEE, 2008, EarlyView. DOI: 10.1002/anie.200804777.
A fifth appearance for my favourite drug ‘interloper’ in to this natural-product space; alarm bells shouldn’t be ringing - just cause I work in pharma doesn’t mean I’ve turned my back on natural products! Tamiflu is of course based on a natural product, shikimic acid - the starting point for the original synthesis. But as natural sources go, it’s rather hard to get hold of, and thus damned pricey (£248 for 5g on SA just now). Other routes used involved chemistry that was perhaps a mite ‘tetchy’ on scale, such as azides and aziridines. A few years back, Corey announced that the synthesis was solved (which made the national press!!!), with his cracking synthesis that I still love from a chemist’s point of view. However, from a practicle perspective, there were still a few problems, including a -78 degrees or two, which are problematic on a plant scale.
On the same day his former student Masakatsu Shibasaki published his work on Tamifly, which was also a very nice synthesis, but the azides and aziridines were still rocking around the cyclhexane. It appears that he doesn’t think of them as a problem, as this latest synthesis is still loaded with those frisky nitrogenous beasties, but there’s loads more to this synthesis.
It kicks-off with an awesome asymmetric Diels Alder, in which they scoured the top-row of the periodic table for the right addive to induce the asymmetry. Barium isopropoxide did the job, along with a pretty damned complex ligand. A quick look on SciFinder tells us that it’s a five step synthesis, using some pretty nice chemistry in itself! It looks like it’s worth it, too - the result of the DA is very nice, building that asymmetry into the cyclcohexene with apparent ease.
A few steps later - and it’s azide time. Not just one acyl-azide, though; two in this case, all set to do their Curitus rearrangement to impressively build a cyclic carbamate and the Boc-protected amine in one pot. Nice work, as this also differentiates the two amines. However, they were aware of the problems of working with such a tempramental substrate, so they optimised the reaction conditions so the azides were never isolated or removed from their solvent.
After acetylation, they needed to bolt on a carbon, and preffered to use a C1 acyl anion equivalent - in this case, a protected hydroxy malononitrile. An allylic, Tsuji-Trost style opening and decarboxylation of the carbamate did the job nicely, installing the new stereocenter in apparently complete control (not that it matters in this case).
Completeion of the synthesis was a little traumatic - substrate controlled epoxidaiton went well, as did opening of the epoxide. However, the planned displacement of the newly installed hydroxyl to install the 3-pentyl ether was very difficult. They finally surmounted this problem by displacing it with the neighbouring amine (a double Mitsunobu), forming an azidiridine. Opening of this aziridine with Lewis acid and 3-pentanol finally finished the job (after salt fomation).
Nice work, using some impressive chemistry; read the paper for full discussion of all that Diels-Alder work!
Cortistatin Pt. III
Shair, Lee, and Nieto-Oberhuber, JACS, 2008, ASAP. DOI: 10.1021/ja8071918.
A third showing for everybodies favourite androstane, this offering from Matt Shair adds to the quantity of inovative chemistry used in it’s contruction. As a quick reminder, first up was Phil Baran, back in May; then came Nicolaou and Chen in August - along with several ’studies towards papers’. However, rather than my going through it all again, have a look at this excellent review by Stefan Bräse which was in ACIEE last month.
If you read it through, you’ll notice Nicolaou’s use of the Hajos-Parrish ketone (the synthesis of which I discussed in my post on that work); Shair does likewise, reinforcing it’s application towards steroid synthesis. From the single stereocenter present in that SM, Shair first does a substrate-controlled reduction of the ketone to add one stereocenter, and then a few steps later adds a pair by selective hydrogenation, and then Rubottom oxidation. Nothing new here, but it is all very neatly controlled by the substrate.
Removal of the acetal protecting group and a bit of aldol chemistry introduced a cyclohexanone, which was triflated to provide a handle for a slightly unusual palladium-mediated coupling (or at least I found it a bit special…). In this case, the nucleophilic partners for the Kumada coupling was a silyl methylene Grignard - an interesting substrate, as I’m sure there could be a competing Hiyama process, but I guess the Kumada process is far faster. Either way, it’s a neat way to make the required allyl-silanes.
A cyclopropanation of the allyl silanes using dibromocarbene gave them the final functional group required for a rather sweet ring expansion to install the seven-membered ring. However, the nature of the silyl group was quite important, as in case a) (where TMS was used), a process of proton abstraction rather than silyl group elimination competed. Moving to the disiloxane derivative favoured attack of fluoride rather than deprotonation, and removed the competing pathway entirely, resulting in a good yield for the ring expansion.
The remaining bromide was then used as a handle for a Suzuki coupling, appending the remaining carbons for the A ring. Substrate controlled dihydroxylaion of the resulting trans olefin left them almost ready for the pièce de résistance, an aza-Prins transannular cyclisation, starting with in-situ deprotection of the MEM group (which I’ve omitted for clarity), and building several rings in one step. I must admit that I had trouble visualising the freed hydroxyl ‘reaching’ over to the olefin, but building a model relieved my fears.
Completion of the molecule (and appendage of the isoquinoline) was done much in the same way as Baran and Nicolaou did, finishing a pretty tasty synthesis. However, one criticism has to be the number of yields which were quoted over multiple steps. Most were admittedly very good, but it’s always nice to know if, say, a 65% over two steps was [65% + 100%] or [80% + 81%].
Agelastatin A
Tanaka, Yoshimitsu and Ino. Org. Lett., 2008, ASAP. DOI: 10.1021/ol802225g.
Looking at this article, the second thought I had (after being generally impressed with the key steps) was that I’d perhaps blogged this molecule before. I was right - it was one of the first I wrote, back in April of 2006 - by Trost. My writing style may have progressed since then (and I’ve certainly become more, umm, expansive, but I still knew quality when I saw it. Trost’s work centred on his manipulation of allylic chemistry using palladium catalysis - quite different to this approach by Tanaka.
Clearly I wasn’t interested in biological profiles back then (how times change…), but there’s some worth noting here - antiproliferation of ’several human cancer cell lines’, and a bit of (GSK-3β) inhibition. But a 5,6,5,5-ring system is a good incentive too, especially if one can make it easily. The first ring is bought-in, but that’s forgiveable, as it is cyclopentadiene. An oxidative cycloaddition with Boc-protected hydroxylamine resulted in a dihydro-oxazine, which was reduced to the amino-alcohol using molybdenum hexacarbonyl and borohydride. Of course, this chemistry is racemic, so the next few steps included an enymatic resolution of this cyclopentene.
Formation of the pyrrole was quite nice, and a blast-from-the-past - a Paal-Knorr condensation. Inversion of the hydroxyl was done with a typical Mitsunobu, with only a few more steps to get to the next point of discussion - and the centrepiece of the paper. Formation of an azidoformate (a tricky beast for them to complete), and then heating the crap out of it in a sealed tube gave them an impressive yield of a pretty strained aziridine.
A bit more azide (and do remember exactly how nasty sodium azide is!) opened the three-membered ring, resulting in a net formation of two stereocenters with great control; six more steps and they’d finished the target. Azidetastic.
Nakiterpiosin
Chen, Gao and Wang. JACS, 2008, ASAP. DOI: 10.1021/ja808110d.
Now that’s an interesting architecture! I’ve always had a soft-spot for homosteroids like this (and cortistatin), as their construction can often include that brilliant mid-20th century chemistry that folks like Corey and Woodward made their names with. However, for this particular natural product, Chuo Chen of the UT Southwestern Medical Center has used a real mixed bag of impressive reactions.
The target, a C-nor-D-homosteroid to be precise, is actually isolated from that marine goo that seems so prolific. And like a few of its neighbours in the goo, it’s pretty poky, with some mighty-fine P388 murine leukemia cell line busting moves. However, synthesis, in this case, was not enough, and and a bit of a stereochemical refinement was also required as the positions numbered in red.
The key to the construction was of course finding a suitable bisection point (of course it’s not a point, it’s a ‘line’, but I can’t think of the correct word…), and in this case, it’s about the cyclopenatone. That gives us a pair of chunks to get cracking with.
First up is the LHS, with its neat bicyclic ether feature. How to do that? A furanyl Diels-Alder is quite an obvious idea, but for some reason I didn’t expect it to work as well as it did. So attack of the Weinreb amide with isoprenyl grignard gave them the precursor for an intramolecular Diels-Alder (IMDA), generating the required stereochemistry about the cyclohexene rather neatly. This was controlled, of course, by the free hydroxyl, installed earlier by doing a Noyori reduction.
However, that hydroxyl is actually not in the natural product, and needs to be ess-en-two’d by bromide to give the required functionality. This was done by forming a sulfonate and displacing with lithium bromide, but isn’t as simple as this may sound. The sulfonate used was actually methyl 2-(chlorosulfonate)benzoate, a particularly electron withdrawing beasty. This may have been required to ensure inversion, as these aren’t as straight-forward as one might imagine.
Turning focus to the RHS, I was impressed at how early Chen installs the (what I would presumed to be sensitive) gem-dichloride. This was done by simply treating an aldehyde with chlorine gas (urgh…), base and phosphite (to capture that chlorate?). I’ve done the reverse reaction - aldehydes from gem-dibromides using silver nitrate - but never this, and I’m quite impressed with the result.
Exchange of the aryl bromide for the correspoding stannane gave them RHS partner, so it was time to crack out the palladium. No screening data is given, but the conditions are fairly standard, resulting in the ever impressive incorporation of carbon-monoxide, unifying the fragments.
Nice reaction, but there’s still quite a lot to be done - and several somewhat sensitive groups present, which always makes the chemistry tricky. First-up was to complete the cyclopentane, using a photo-Nazarov reaction - a classical method of forming cyclopentenones from dienones. The regioselectivity of the reaction (in terms of which side of the five-member ring has the unsaturation) is controlled by the ‘cation’-stabilisation ability of each side - and the phenyl ring is better (and also wants to remain aromatic), so results in a beautiful rearrangent and generation of a pair of stereocenters. The base was required to improve the stereoselectivity of the reaction, as the ‘α position’ was ill-defined (1:1). A bit of deprotonation allowed the natural stereochemical bias to be revealed in the desired manner. Super-smart stuff.
The last reaction I’m going to mention is the ring-expansion of the bridged cyclohexane into a caprolactol. Going back to the cyclohexane formed in the IMDA, the diol was formed by dihydroxylation (which also prevented reverse DA chemistry). This was protected-up as an acetonide, and sat inert for most of the synthesis. Freeing it up, and oxidising with a bit of hypervalent iodide resulted in an insertion of oxygen between the diol, and ring-expansion to generate a bis-hemiacetal. Selective reduction of the less-hindered bis-hemiacetal with lewis-acid mediate silane then gave the required functionality present in the natural product.
Just brilliant stuff. I loved reading this paper, and hope to read a full-paper on it in the near future.
Lysergic Acid, and definitely not LSD.
Fujii, Ohno, Inuki and Oishi, Org. Lett., 2008, ASAP. DOI: 10.1021/ol8022648.
Fancy making your own LSD? You could do a lot worse that this prep from Kyoto University. Of course, they don’t actually make the lysergic acid diethylamide, but I’m sure there’s a way to form that amide… Anyway, we should probably stick to the actual paper, and leave the class-A for another time. I’m sure we’re all aware of the biological activity of some of the members of the ergot alkaloid family, but did you know about synthetic members with anti-Parkinson’s disease properties? Certainly gives them a more legitimate reason to work on the synthesis.
For me, it gets interesting with a gold-mediated Claisen rearrangement. Done on a propargyl alcohol, this generates an allene in cracking yield. This is quoted over two steps, as they did an in-situ reduction of the initially formed aldehyde. It took a little effort to find these conditions, as their initial attempts with just a bunsen-burner resulted in both a shoddy yield and d.r. Switching to microwave irradiation boosted the yield, but left the d.r. unimproved, whist a bit of the oxo-gold complex did the job nicely.
However, separation of the diastereoisomers was ‘difficult’, so the mixture (a few steps on) was dumped into an impressive palladium-mediated domino cyclisation. A variety of conditions were applied, also varying the protecting group on the pendant amine from nosyl to the more optimal tosyl, finally resolving in the conditions shown. The yield is pretty decent for what is an impressive transformation, and the d.r. also.
I might have taken those yields and moved on, but the group were considerably more thorough, and did a useful mechanistic study. For this, they separated the mixture of allenic starting materials, and repeated the reactions with diastereomerically pure starting materials. Using the purified, desired SM, a worse d.r. (but better yield) was obtained for the cyclisations; the group suggest that this could result from two competing pathways.
They suggest that the preferable pathway (on the right) is such because of the disfavoured, strained palladacycle present in the first step of the minor pathway.
Impressive as this is, I must admit that my attention wavered as the group went on to make the title compound, and a pair of diastereomeric compounds, lysergol and isolysergol. However, I was somewhat taken by the use of thioglycolic acid along with lithium hydroxide to remove the tosyl group. I knew I’d seen that stuff in a bottle back home, and I was right - my grandmother uses it to ‘perm’ her hair…
Oh, and if you’re started ordering reagents to follow the prep, can I dissuade you, and suggest reading (or watching) ‘Fear And Loathing In Las Vegas‘ as an alternative to a life time of freaky flash-backs.
Panacene
Canesi, Sabot and Berard. Org. Lett., 2008, 10(20), 4629-4632. DOI: 10.1021/ol801921d.
I nearly missed this paper, published online at the beginning of last month, but it’s an incredibly neat solution for the synthesis of some funky little natural products. The parent molecule, panacene, has some fairly unique biological properties, as it is a shark-antifeedant. In otherwords, it puts sharks of their food, and makes the sea-hare in which it is produced that little bit less attractive to the peckish big-fish. Okay, so it doesn’t cure cancer, but the next time you’re swimming in the Med…
I suspect (though I’ve no justification) that this paper was a get-the-methodology-then-find-the-target type of study, but it’s certainly methodology that seems to work rather well. The key motif is a umpolung addition of furan to phenolic-type systems (or more largely electron-rich aromatics), using iodobenzene diacetate to generate a cationic intermediate that allows attack of furan. Sort-of reminiscent of Friedel-Crafts chemistry. The oxonium ion produced is then rearranged by attack of the ketone to provide a 6,5,5-fused system in a reasonable yield. There is a selectivity issue as to which position ortho to the phenolic hydroxyl is fused, but this was overcome by using a TMS blocking group.
With the desired system in place, it was time to add the desired allenic sidechain. Two different approaches were used to impart the different sidechains, with the unhalogenated desbromopanacene made most succinctly in only two steps from a common (desilylated) intermediate. Firstly, the DHP DHF was oxidised hydrated to the hemiacetal using an oxo-mercuration (rather them than me…) / borohydride process. Treatment of this with a propynylsilane under Lewis-acidic conditions allowed a Sakurai reaction, neatly completing the target in good yield.
For the parent molecule, panacene, the brominated allene moiety meant that a more complex route was required. Using the same hemiacetal shown above, they did a Wittig olefination to give a enyne. Again, a spot of mercuric acetate was used, but this time the product was a mercuric allene. To give the target, they did an in-situ protiodemercuration using ethandithiol, which was impressively stereoselective. Nice stuff!
Dibromoagelaspongin
Feldman and Fodor. JACS, 2008, ASAP. DOI: 10.1021/ja807020d. JOC, 2007, 72(21), 8076-8086. DOI: 10.1021/ja807020d.
This one looks, at least from the JACS communication, to be short piece of work over only two pages, but it’s very reliant upon methodology and ideas presented in a JOC full-paper from back in 2007, so perhaps we be best to read that one first. Although this synthesis is racemic, that isn’t actually an important issue, as the natural product may be racemic itself. The lack of certainty seems a bit odd to me, as it doesn’t take much material to get a quick alpha-dee, but at least we know the relative stereochemistry for sure, from an x-ray.
The starting material for the chemistry I’m going to ramble about was made in seven steps, from simple precursors, splitting about the ever-attractive amide bond. Treatment of the functionalised imidazole with triflic anhydride resulted in a Pummerer reaction (remember that from the first time you tried to do a Swern oxidation), which could proceed in either of the two mechanisms below. Postulate a is a vinylogous Pummerer, where a sulfonium ion is firstly formed along with an iminium ion. Something about this intermediate (i.e. the two charges) makes me a bit uncomfortable… The other option is a more standard Pummerer, where the amide nitrogen attacks (hmm…) the imidazole directly, resulting in a spiro-fused intermediate that must do a 1,2 N-shift to give the 6,5-fused target.
Either way, the yield is fairly good for such a piece of chemistry!
Next up was a deprotection of the SEM group, which gave the intermediate required for the next cyclisation, which would complete the 5,5,6,5- ring system. To do this, they simple dropped the SM into a bit of NCS, which led to a cracking yield of the target. However, the mechanism for this reaction is again lacking clarity.
This time, though, the group were able to be a bit more proactive; repeating the reaction with other halide sources led them to conclude that a further Pummerer process was at play, probably beginning with chlorination of the sulfane. Elination of chloride (and overall oxidation) proceeds via sulfonium ion formation (and another dication), which can be attacked by the imidazole to form the desired ring system. The iminium ion can then be quenched by attack of the chloride ion (which has been hanging around…).
A couple of deprotections, and formation of a guanidine left them with the natural product, but I did think that the displacement of the methyl-sulfoxide by azide (in the presence of zinc iodide) was quite interesting. Any thoughts on what is going on there?
|
