Baran, Chen. Nature, 2009, AOP. DOI: 10.1038/nature08043.
They say that the best defence is a good offence, so I’m going to start latterly – and remind (somewhat forcefully, if necessary) all prospective commenters that the folks at Nature decreed that this paper is worthy. That’s good enough for me, so let’s avoid the tired arguments about suitability… </pre-emptive rant>
Anyway, what we’ve got here is typical of Baran – re-examination of chemical methodology of the past, and modernisation to solve problems in an orthogonal fashion. Breaking that down, what he seems to be really good at is picking up an idea from the past, re-examining the goals, and using modern techniques and reagents to solve problems. This appoach often allows him to disregard common retrosynthetic failings, such as recursive oxidation / reduction and protecting group abuse – but must come unstuck quite often. (I don’t know that it does come unstuck, but I rather hope that is the case for my own sanity…) Today’s target are terpenes, and the old-school technique I alluded to is remote oxidation. This one generally describes selective (and unactivated) C-H oxidation using a scaffold in which a remote group provides the impetus/selectivity.
Of course, the first requirement is a skeleton to work with, and rather than a semi-synthetic intermediate, Baran decided to make his own (and rather quickly too). First up is a little organocatalysis, using a standard proline-based catalyst to do an enantioselective Michael addition, which following a spot of base give the cyclohexenone in great e.e. and decent yield. Notable is the low catalyst loading (5% is pretty good for organocatalysis), and the use of a catechol to activate the enone for addition.
A few steps further on, and they’d bolted another ring to the cyclohexenone, with a cross-conjugated diene and single stereocenter in place. However, three simple reaction allowed the group to greatly enhance the stereo-complexity – cuperate addition and a pair of reductions – and now we have five stereocenters. Nice work! This trans-decalin is the least-oxidised (most reduced) carbon skeleton from which Baran works, and it’s a good thing he has such a concise route towards it, as he targets several natural products from here.
All of these targets share one further oxidation, so it makes sense to put it in first. This is done by firstly appending a controlling group for the remote oxidation – a carbamate featuring a trifluroethyl chain. This choice was directed by their methodology for 1,3-oxidation published last year in JACS, and is related to the Hofmann-Loffler-Freytag reaction (used to do a 1,3-bromination via atom transfer). Their development of this chemistry lead to the use of a trifluroethyl group to encourage N-centered radical formation, increasing selectivity. So we’re all-set to do a bit of 1,3-oxidation, but there’s a complication – as there are three possible tertiary centers avaliable. Their smart move was a bit of rationalisation using 13-C NMR to probe the electronegativity of the various carbons, confirming that they had a good chance of selectivity, which was bourne out. In this case, it works really quite well, using methyl(trifluoromethyl)dioxirane – a varient of DMDO in which is more active, and selective for equitorial C-H bonds.
The next most reactive position for this chemistry was dealt with accordingly, but this time doing a bromination to provide a common intermediate, 18. In the simplest case, treatment with silver carbonate resulted in the diol, which after clevage gave pygmol in cracking yield over four steps. It’d be interesting (but a lot of work) to see if this could’ve been stereoselecive, had the center been prochiral.
Using the same intermediate, Baran was able to complete two more targets, 11-epieudesmantetraol and eudesmantetraol, by firstly eliminating the bromide with tetra-methyl piperidine (can I assume that the methyl groups are bis-geminal, alpha to the nitrogen?), and then brominating with NBS to form a cyclic carbonate. This time we’re stereoselecitive, and it remained so when the opened the carbonate, which closed onto the primary bromide to give a terminal epoxide. Only one step was required to access each natural product, choosing either acidic or basic condition to open the epoxide (remember your first-year org-chem?).
This is a fantastic piece of work – not just smart, but well executed. However, it’s worth remembering that this approach isn’t entirely new, as remote oxidation was a very popular technique for steroid chemists. Use of transition-metal oxidants, such as iron coordination complexes was common, such as in this example by Paul Grieco. However, Baran had taken this idea and modernised it considerably, with the NMR/X-ray analysis of reactivity a critical step.