Cool Syntheses

My biggest complaint about the Science & Nature section is a distinct lack of organic chemistry. So I thought I'd take it upon myself to change that, and create a thread where I hope to summarize a cool synthetic study once in every while (spare time & motivation permitting). The advantages I see (beyond satisfying my own interest and learning a thing or two along the way) is bringing interesting synthetic research to people who don't have access to scientific journals (probably most of you), and hopefully to stimulate conversation about the exciting field which currently garners very little discussion here.
 
Who knows, it could interest no one else but me, but it can't hurt. So I will start with a recent paper published in Organic Letters by the Gademann group from the University of Basel, Switzerland, outlining their elegant and protecting group-free syntheses of taiwaniaquinone F and taiwaniaquinol A.
 
These two natural products belong to the taiwaniaquinoid family, which are diterpenoids boasting unique 6-5-6 tricyclic structures. These have attracted wide interest from the scientific community, and their biological activity is noteworthy (including potent cytotoxicity against epidermoid carcinoma (KB) cells). The former statement is exemplified by the fact that this paper I speak of was published only weeks prior to a second publication from an independent group in Shanghai who published their own synthesis (which I will not cover*).
 

 
Gademann initiated the syntheses of both from the commercially available (-)-abietic acid (which is a major constituent of resin from pine trees, is what violinists rub on their strings, and conveniently possesses two of the required stereocenters - chiral pool strategy). This was converted to the natural product (+)-sugiol methyl ether in 36% yield over nine steps on multigram scale, following literature procedures. This intermediate was then diazotized using p-acetamidobenzenesulfonyl azide (p-ABSA) and DBU (a base), to lay the foundations for the first key step in the total synthesis.
 

 
With the diazoketone intermediate in hand, Gademann employed a Wolff rearrangement. Here, the carbene intermediate generated from light underwent a particularly interesting ring contraction (I attempted to show the mechanism), thus establishing the 6,5,6 fused tricyclic scaffold together with the connected carbonyl group as a methyl ester. Indeed, Wolff rearrangements are catalysed by light, the source of which was a mercury lamp here (presumably the fastest acting source), but the author noted that the reaction proceeded under sunlight with identical results - remarkable (and more on solar power to come)!
 

 
Although the reaction proceeded with 20:1 diastereoselectivity, the unwanted isomer was favoured. In English - they got the mirror image of what they desired. The author suggested that was due to steric effects, whereby protonation occurred at the less hindered face of the molecule leading to the unwanted isomer. Unperturbed, the chemist(s) treated the methyl ester with sodium methoxide in methanol under microwave irradiation (100 degrees Celcius), effecting epimerization (flipping the group pointing the wrong way to the right way) to provide the desired diastereomer (not the mirror image) in quantitative yield, with a diastereomeric ratio of 33:1. This compound was further reduced with lithium aluminium hydride to the corresponding alcohol in quantitative yield, setting the stage for the next key step in the synthesis.
 

 
I'd like to note that lithium aluminium hydride (LiAlH4) - affectionately known as "LithAl" by many including myself, is one of my favourite lab reagents given how reliable it has been to me (and to Gademann, giving him a 99% yield, beautiful stuff). But it is a force to be reckoned with - a previous member of my lab received 3rd degree burns to the face after stupidly cleaning out a plastic bag covered with it, using methanol. Methanol + LiAlH4 = raging fire.
 
Anyway, back to the story now. With the requisite alcohol in hand, Gademann could employ standard procedures to brominate the ring, which would then undergo a notably succint oxidation to the phenol using dioxygen (a stream of O2 is bubbled through the reaction mixture) with n-BuLi (a very strong base) and TMEDA (a ligand to activate the aforementioned base). This novel approach is particularly attractive as standard procedures would employ a relatively clunky hydroboration strategy, which the author originally sought, before developing this simplified method. A second oxidation to the quinone again employed dioxygen, alongside Co(salen) - a cobalt ligand that acts as a vehicle for oxygen. It should be noted that these steps and purification of the light-sensitive intermediates were carried out in completely light-free conditions. I can tell you that it's particularly frustrating when dealing with light sensitive compounds... chemistry is hard enough in a brightly lit room, so my hat goes off to the poor bastard who did all of this in the dark... particularly the purification part, as this was carried out by conventional chromatography. I imagine that they used brown glass columns.
 

 
All that stood between Gademann and (-)-taiwaniaquinone F was a simple oxidation of the hydroxymethyl group to the required aldehyde, which was carried out in quantitative yield using Dess-Martin periodinane (DMP). The story does not end here, however. As what I (and likely the authors) find to be the most novel aspect of this study is the conversion of taiwaniaquinone F to taiwaniaquinol A. As I mentioned, synthesis of intermediates in the previous diagram were marred by light-sensitivity. The keen chemist(s) performing these reactions did not overlook the fact that one impurity's NMR had a striking resemblance to taiwaniaquinol A! They leapt upon this observation by intentionally exposing an ethereal solution of the light-sensitive taiwaniaquinone F to sunlight, remarkably affording a 30% yield of taiwaniaquinol A.
 

 
Mechanistic insights into this transformation are forthcoming from the authors, however they postulate that an elaborate C-H functionalization takes place, whereby an alkoxy radical (generated by irradiation) may participate in a 1,5-H abstraction of the methoxy group. In other words, the oxygen radical rips off one of the hydrogens attached to the nearby methoxy (OMe) group, forming the bridgehead. However true that might be, it is clear that this natural product readily undergoes photolysis to furnish a methylenedioxy group in one way or another - without the use of any complex materials or protecting groups, just sunlight will do. The authors acknowledge that this is quite likely the mechanism employed by Mother Nature - which is corroborated by the fact that taiwaniaquinol A is present in the leaves of the producing tree (Taiwania cryptomoides), whereas taiwaniaquinone F is only isolated from its roots. This entails that other non-enzymatic biosyntheses of these methylenedioxy-containing natural products could be occurring elsewhere, too.
 
And that concludes my summary of this elegant natural product synthesis. I'd like to again acknowledge that this is purely for a little education, and of course all credit goes to the authors of the original publication (which I did not plagiarize, everything here was reproduced honestly). I hope someone out there enjoys this chemistry as much as I did, and hopefully I'll produce another story from the literature in time.
 
*When I originally conceived the idea to cover this synthesis I was unaware of this paper. Those chemists were unlucky as they were writing their manuscript when Gademanns' paper was published, and did not manage to take prize for being the first to synthesise taiwaniaquinol A (Gademann earned that). Their synthesis is impressive, though. They went a step further and started from the extremely simple precursor 1,2,4-trimethoxybenzene, rather than 'cheating' with a chiral pool precursor like abietic acid. Furthermore, their synthesis is divergent and accessed not only taiwaniaquinone F and taiwaniaquinol A, but also taiwaniaquinols B and D. Interestingly, they too arrived at a nearly identical Wolff rearrangement step discussed here, however they found difficulty with methanol and opted for the benzyl derivative. Their synthesis suffers in that it is racemic, and is certainly not as sleek as the approach I outlined above (no bromide-oxygen exchange or sunlight-induced C-H activation!).

 

Thanks for the encouraging words, I'm very glad that it was appreciated by you both!
 
Chiefton I certainly see the appeal of letting bacteria do the hard work, we're always 2 steps behind Nature and with progress in biotechnology it's not unthinkable that more and more complex synthesis will be achievable through engineering bacteria effectively. It's scary to think that one day my job might be taken by a microorganism!
 
It's really impressive that the enzyme you speak of is able to tether an aliphatic group like isobutane to an aromatic ring. Could you let me know what paper this is? I'd like to check it out. I'm a bit out of my depth discussing exotic enzyme-catalyzed reactions, but I presume the oxygen forms an activated oxygen/iron species analogous to that of Cyp450 enzymes, which would facilitate C-H activation of the isobutane. I couldn't possibly say for sure (maybe someone smarter could!?) but it also seems likely to me that radical intermediates lead to the tethering of this unlikely pair.
 
Like this? http://www.nature.com/nchem/journal/v3/n5/full/nchem.1038.html

 

 
I'm glad you're taking an interest, good to hear you getting into it. Sorry the Wolff step could have been fleshed out a bit more, it's not obvious what's happening, from those drawings.

I've attempted to show (or speculate) what goes on above. Light cleaves off the diazo which is happy to leave as elemental nitrogen, with a very reactive carbene left behind. That collapses to the five membered ring in a Wolff rearangement, ejecting the carbonyl group (C=O) and forming a ketene (C=C=O). The new bond that formed is bolded above. Ketenes don't hang around for the party, being highly susceptible to attack from anything around them including methanol (with three hydrogen-bonded molecules of methanol participating above).
 
As for your question about preparing for class. I don't think there's really too much you need to do beforehand, other than to prepare yourself for a barrage of information and structures!

 

 
They possess "potent cytotoxicity against epidermoid carcinoma (KB) cells" - i.e. naturally occurring anticancer agents.

 

The holiday season has granted me enough spare time to muster up another addition to this thread. I thought about covering Prof. Baran's impressive synthesis of ingenol, but it occurred to me that it might be more appreciated by others here if I was to cover a synthesis of THC, or more specifically, (-)-Δ⁹-tetrahydrocannabinol.
 
I doubt this post requires much elaboration of what THC is, where it's found in nature, or its biological activity. So I'll cut right to the chase. While there are many previous syntheses of THC, very few have accomplished an approach utilizing the holy grail - asymmetric catalysis. In other words, using a catalyst (typically a transition metal like palladium, ruthenium, etc.) to produce only one three-dimensional structure (stereoisomer). This means you can avoid relying on expensive starting materials that already contain the chiral centre, among other advantages. The group of Q.-L. Zhou accomplished this in 2013 with their asymmetric synthesis of THC. 
 
Starting from a commercially available cyclic enone 1, straightforward iodination with iodine, followed by a Suzuki coupling with an arylboronic acid derived from 3-fluoro-5-bromoanisole afforded enone 3, which was reduced to the racemic ketone (rac-4) by simple hydrogenation under atmospheric pressure. All known chemistry, not to gloss over that Nobel-prize winning Suzuki coupling. Those kinds of palladium-catalysed C-C bond forming reactions are truly remarkable, and I am from a generation that surely take them for granted. For those of you who might not be aware, the squiggly bond in rac-4 denotes the fact that it is racemic (rac) – i.e. a mixture of two stereoisomers (known as enantiomers in this case). Any carbon centre bonded to four different groups (i.e. sp³ geometry) give the molecule three dimensional properties, better known as chirality. Chirality can be the difference between a drug or brick dust, or the smell of lemons or oranges.
 

Moving on, things start to get interesting. The group managed to convert rac-4 into pure S-4 via diastereomer 5, using their own protocol of asymmetric hydrogenation via dynamic kinetic resolution, followed by routine oxidation. What happens in a ‘normal' hydrogenation of ketones is that dihydrogen (H[SUB]2[/SUB]) will add unselectively across the C=O bond, reducing it to form an alcohol (C-OH), generating a chiral centre in the process. So if you had a racemic substrate as in this case, you'd end up with four possible outcomes (as you make another chiral carbon, there are 2^2 possible stereoisomers known as diastereomers in this case). However, using this highly specialized catalyst (C1), you can selectively reduce one enantiomer whilst ignoring the other.

To explain, the bulky and chiral nature of the catalyst (C1) blocks it from interacting with one enantiomer. As such, hydrogen can only be delivered to one face, and subsequently only one possible enantiomer can be made. The other enantiomer is simply left alone by the catalyst. The story doesn't end here, though. The clever chemists devised conditions in which isomerization of the enantiomers (rac-4) will occur  â€“ so that the non-participating enantiomer (R-4) will flip to the reactive enantiomer (S-4). What happens next is that S-4 is swiftly and irreversibly reduced to the alcohol (5). As the two enantiomers are in equilibrium, i.e. will remain in similar ratios, the effect will eventually work to convert all of the non-participating enantiomer resulting in one optically pure product in excellent yield. That is dynamic kinetic resolution, and is an incredibly useful and efficient reaction.
 
Moving on, the alcohol 5 was then re-oxidized back to the ketone (S-4) in a Swern oxidation which set the scene for a Wittig reaction. Both are textbook reactions, and both were carried out in excellent yield, providing alkene 6 in a cis/trans ratio of 4:1. The mixture of isomers is neither here nor there, as the alkene is soon to be destroyed.

6 was subjected to aqueous acetic acid leading to hydrolysis of both the acetal and enol ether groups, forming a ketone and aldehyde respectively (7). This was followed by a Jones oxidation, converting the newly formed aldehyde to a carboxylic acid (8) which was immediately methylated with methyl iodide to give the methyl ester (rac-9) as a mixture of cis and trans isomers in a 1.4:1 ratio. Isomerization with sodium methoxide in methanol then provided the trans isomer (R,R-9) in a 31:1 ratio.

Subjecting (R,R-9) to methyl magnesium bromide, a Grignard reagent, led to additions of methyl groups at both hemispheres of the molecule resulting in diol 10, which was ripe for the formation of the tricyclic core, where the molecule begins to actually look like THC. Now we will see why that fluoride atom was so cleverly introduced early in the synthesis. Unlike just about any atom or leaving group I can think of, fluoride will undergo S[SUB]N[/SUB]Ar reactions – in other words, it may be substituted by an appropriate nucleophile even when directly bonded to an aromatic ring like benzene. Although in my own experience, these reactions are a pain in the ass, it served these guys well indeed with a 94% yield after only one hour, forging the tricyclic structure of the desired cannabinoid. Just for a laugh, they then revisited their conditions to incorporate a tandem demethylation of the phenolic oxygen in the same pot, which worked well to deliver the phenol 11 in 90% yield. All they needed to do was throw some 2-diethylaminoethanethiol into the same reaction. Following that, all that was standing between them and THC was simply one hydroxyl group.

The removal needed to be selective in order to afford the correct regioisomer. Incorrectly, and you get boring old Δ⁸-THC… which they did anyway, because why not (by treating 11 with tosic acid in benzene under reflux). To get Δ⁹-THC they needed to use a slightly more exotic set of conditions, which feature the “Lucas' reagent” (first time I've heard of it) which swap out the hydroxyl group for chloride, which is rapidly converted to the alkene by stirring at 65 C with potassium tert-pentoxide in benzene for fifteen minutes. I'd have to wager that they found that last step particularly challenging before consulting old Lucas' reagent, as those conditions are weird as hell and it appears that they dug them out of an article written in German from 1969! But that is the fun of it all, I suppose, blowing the cobwebs off old relics to find that perfect set of conditions.  

And then they had it, (-)-Δ⁹-THC, in an overall yield of 30% over fourteen steps which is an impressive feat. Just for the record, and for people who sometimes ask around this forum, THC is not crystalline! I don't know why people discuss "crystals of THC", I suppose the somewhat crystalline appearance of trichomes may be to blame. But pure THC is most often characterized as being a colourless oil. From reading elsewhere, cooling the oil will produce an amorphous solid (glassy or waxy, I couldn't tell you, I'd bet on the latter). Most molecules with long greasy carbon chains (that C[SUB]5[/SUB]H[SUB]11[/SUB] bit) are bound to be anything but nice crystals.

 

 
Hey, thanks. I'd love to update it more often, just hard pressed for time usually. Even if no one reads it, I get a lot out of making the posts, so it's great that people have actually stopped to have a read.
 
The point your raised is something I learned from reading this paper. Using standard acid-catalyzed dehydration (E1 elimination) of 11 will invariably provide the more thermodynamically stable delta-8-THC. The use of Lucas' reagent actually proceeds through a chlorinated intermediate, shown below. Treating that with base then follows through an E2 elimination which ensures formation of the delta-9 analogue