Obviously this style of post is borrowed from Derek Lowe's Things I Won't Work With. I'm not the most original person, and this certainly isn't the most original blog.
Hopefully many of you will have now read Baran's axinellamine paper (my take here) - it's an impressive piece of work and worth a few minutes of your time. I've noticed that I keep having the same conversation about it with different people over the last day or two:
BRSM: Have you seen Baran's new axinellamine paper? It's the shit!
Chemist: Yeah, it's pretty cool, but you couldn't pay me to work on that project. Those guys must have spent a lot of time staring off the roof of the Scripps chemistry building.
BRSM: But he's so sexy!
Chemist: I have to go now, I, uh, have a TLC running...
Or variations on that theme. Anyway, I realised that actually: 1. I'm probably inclined to agree, and 2. I find myself thinking this quite a lot. As I told commenter gippgig yesterday, I saw Baran give a talk about this work back in 2009. I remember him describing the silver(ii) picolinate oxidation and saying they had to screen a massive number of oxidants to find this successful reagent. I think he described one of his students as having a 'black tar phase' of 6 months or so, where everything presumably kept degrading when they attempted to install that troublesome hydroxyl. Silver(ii) picolinate obviously wasn't high on the list of things to try as they had to make it specially, and none of the references they give for its use are later than the early 1970s. Periods of faliure come to everyone who practises chemistry long enough, but when you're pushing the boundaries of what's known, as Baran was here, you can spend months in the lab going nowhere at all.
A couple of other syntheses from this year with similar heroic efforts at optimisation come to mind. I should say now that I think they're both fantastic, but I'm not sure I would have had the dedication and mental fortitude to see them through.
I'm sure many of you will have read the Tot. Syn. coverage from when the paper came out. I'm only going to talk about the last step here - the amazing Brook rearrangement followed by conjugate addition to the enal. This is a very cool simultaneous disconnection for the strychnine D and F rings, but, perhaps unsurprisingly, required a lot of experimentation to realise. Saying nothing of how long it took find conditions that actually gave any product at all, Vanderwal says that over 100 experiments were tried to attempt to improve the yield, and yet 5-10% was the best they ever got. I bet that was a depressing couple of months for the one student on the paper, although I suppose having their name on the world's shortest route to the racemic natural product offered some small comfort afterwards.
However, this task seems positively straightforward compared to the sisyphean effort of...
Obviously, when targeting dimeric compounds, a late stage dimerisation is the way to go, and this is exactly the retrosynthesis that Herzon and coworkers applied to the lomaiviticins:
The carbon-carbon bond forming reaction they decided to use was the oxidative dimerisation of enolates, which is pretty well known. The problem was the instability of the system that they chose to apply it to - enolising a ketone with a b-leaving group, for example, is rarely a recipe for happiness, especially when you start adding potentially Lewis acidic metal oxidants to the mix, as elimination, tautomerisation and irreversible aromatisation are all waiting to happen, given half the chance. And as if that's not hard enough, why not include a diazo group in the monomer for an extra challenge?
During extensive investigation of the dimerisation reaction the group found that the use of a cyclic protecting group for the vicinal diol was essential as the silyl enol ethers of compounds bearing acyclic protecting groups were unstable above 0 ºC. Attempts to perform the oxidative coupling at lower temperatures using conventional oxidants such as CAN or copper chloride returned only aromatised material with all substrates tried. Eventually, the Manganese catalyst shown was found to be uniquely effective for this transformation. Herzon attributes its singular success to its solubility and stability in benzene, and its low Lewis acidity, factors which minimise the competing b-elimination-aromatisation. Anyway, after just 1500 experiments the group found that the dimerisation could be performed in 26% yield. To put that in perspective for non-bench chemists, that's how many reactions I did in the first two and half years of my PhD. This is a truly incredible feat of endurance for the authors. Mightn't you have been tempted to stop after the first 100, 500, or 1000 failed reactions? I would have. In the words of Samuel Goldwyn's famous malapropism, 'gentlemen, include me out'.
Has anyone else seen a recent synthesis that makes them feel the same way?
1. However, I imagine the work ethic at Yale is a bit different to where I'm studying in the UK.