This inaugural Woodward Wednesdays post will discuss the subject of Woodward's 1965 chemistry Nobel Prize lecture - work culminating in the synthesis of (+)-cephalosporin C. It was difficult to choose a synthesis to open the series with, as a lot of Woodward's papers are, quite rightly, considered classics and have been dissected elsewhere. Woodward's synthesis of strychnine, for example, crops up in a number of reviews, has a wikipedia page and is discussed at length in Nicolaou's Classics in Total Synthesis (Chapter 2) as well as T. Hud's Way of Synthesis (pages 803-808) and probably many other places besides; I'm not sure I can add much there that hasn't already been said! Woodward's reserpine synthesis, one of my top five syntheses of all time, is (unfortunately?) also covered in a similarly comprehensive fashion. Strangely, the cephalosporin synthesis remains much less well known, despite containing some excellent chemistry and a few 'Woodwardian' steps.
"Having here this morning the responsibility of delivering a lecture on a topic related to the work - for which the Prize was awarded, I have chosen to present an account of an entirely new and hitherto unreported investigation which, I hope, will illuminate many facets of the spirit of contemporary work in chemical synthesis" - R. B. Woodward, Nobel Prize Lecture, 11 December 1965
So began Woodward's Nobel lecture - in a departure from tradition, for he spoke for the entirety of his lecture solely on his thus far undisclosed work on the synthesis of (+)-cephalosporin C, unpublished until the following year (J. Am. Chem. Soc., 1966, 88, 852), in his characteristically intelligent and articulate manner.
"Often in the course of synthetic work one or two key ideas set the style, development, and outcome of the investigation, while providing the flexibility essential for any long journey through unknown territory, beset with perils which at best can be only dimly foreseen."
I'm reminded of this every time I see a new methodology shoehorned into a total synthesis, but here Woodward is referring to his choice of (+)-cysteine as the starting material for the route. Although it's not obvious at the outset quite how it'll all work out, it seems like a sensible place to begin. Besides, what other chiral, commercially available sulfur containing compounds can you name? Remember, this is 1965 - if you want to make a single enantiomer of a natural product then your options are either a dip in the chiral pool or a tricky chiral resolution, as other methods of introducing asymmetry along the way don't exist yet. With the starting material chosen, the first steps of the route were spent making it a bit easier to handle and, crucially, creating a cyclic system on which stereoselective operations might be carried out more easily. Surprisingly, the first reaction with acetone requires no additional reagents; the protected derivative can be formed just by heating cysteine in acetone, which is news to me! Also, notice the unusual conditions used for the Boc protection - phosgene and t-BuOH! That's right, this work predates the discovery of Boc2O, when introduction of a Boc group wasn't the stress free occurrence it has now been reduced to. Perhaps the diazomethane used in the next step makes this seem a bit less crazy. In this case, the Boc protection almost certainly occurs via the mixed anhydride of the free acid in an intramolecular sense, which makes it a bit easier - there's a reason you can't buy t-BuOCOCl (namely that it's not terribly stable). I should say now that for reasons I can't imagine, no yields are given for any of the reactions in this paper, so sorry about the lack of details in places.
Next came the most unusual step of the synthesis - the functionalisation of the remaining cysteine methylene group. Apparently, this proved a bit tricky, but eventually a solution was found:
"I shall not detail here the many weapons which were brought into play against that still expectedly recalcitrant methylene grouping. Suffice it to say that the protected ester reacted with excess dimethyl azodicarboxylate at 105 ºC during forty-five hours to give the hydrazodiester in almost quantitative yield."
Unfortunately, although nitrogen had been successfully introduced at the required position, the hydrazodiester formed possessed the wrong configuration at the new stereogenic center! The group therefore needed a way to formally invert this group. Eventually it was found that when the afore mentioned hydrazodiester was treated with 'somewhat more than two moles of lead tetraacetate' in refluxing benzene it was converted to the corresponding acetoxy compound, largely with retention of configuration. Although the transformation from hydrazodiester to acetoxy wasn't stereospecific (due to the intermediates involved) it was found to strongly favour the (incorrect) diastereomer in which the acetoxy group was trans to the nearby ester. It wasn't necessary to separate the trans diastereomer from the cis, as hydrolysis of the mixture yielded a single diastereomer of the alcohol. Presumably this is due to the greater thermodynamically stability of the trans product and equilibration occurring via ring-chain tautomerism of the hemithioacetal.
Unfortunately, as creative and original as the preceding sequence was, it had provided only compounds epimeric to those required to complete the synthesis. Thus, three steps (mesylation, displacement with azide and reduction using aluminium amalgam) were therefore required to convert, with inversion, the hydroxyl to the amine required for the β-lactam formation. Cyclisation was achieved with a little Lewis acid assistance, in the form of triisobutylaluminium.
Finally, with end in sight, the final ring was installed by a clever sequence beginning with 1,4-addition of the lactam nitrogen to the unsaturated dialdehyde shown below (which was made in three steps from tartaric acid). Don't forget that the small ring size in β-lactams minimises the resonance structure responsible for normal amide reactivity (i.e. through oxygen) so they are nucleophilic (and, incidentally, protonate) at nitrogen. Simultaneous cleavage of the acetonide and Boc groups with TFA also formed the second ring of the natural product by condensation of the thiol onto the nearby aldehyde in a cool one pot cascade. Finally, the known sidechain fragment was coupled on with DCC and its free acid was then protected as the 2,2,2-trichloroethyl ester by coupling with trichloroethanol, again using DCC. At this point, containing 9 chloride atoms, I imagine that the mass spec of this compound was pretty interesting. The enal was then reduced to the allyl alcohol using borane in THF, and this was then acetylated using acetic anhydride in pyridine. Further standing in pyridine for 3 days effected isomerisation of the double bond into conjugation with the nearby ester. Finally removal of all three protecting groups using zinc dust in 90% acetic acid gave the natural product which was 'identical with natural material in paper chromatographic behavior, and in antibacterial activity against Neisseria catarrhalis, Alcaligenes faecalis, Staphylococcus aureus, and Bacillus subtilis'. A great synthesis, conducted without NMR, or apparently even TLC!
1. A later post in this series will discuss the possibility of further, unreceived, Nobels and uncredited work.
2. Later in the series I might write something on reserpine anyway, because it's just brilliant and it couldn't hurt to have a more detailed free summary of it available online for those interested. Also, no dissection of the paper I've seen so far has really captured the incredible detail of the experimental section, which in places is equally breathtaking and hilarious. I doubt any paper published in Tetrahedron (!) in modern times will contain the phrases "six gallons of benzene", "1.2 moles diazomethane" or "4.5 litres of pyridine" in its experimental. In places yields are given as an average over 10 runs and 3 different solvent systems are offered for recrystallisations. It's so evocative compared to the blandness of some of today's papers; I can almost smell the cigarette smoke and benzene and feel the teak benches when I read it. It makes me want to do a steam distillation, a late stage enantiomeric resolution (after all, the best model system is the other enantiomer...), and a semicarbazone melting point, while drinking whiskey in the lab and weeping openly at the beauty of organic chemistry in the Woodwardian era. But we digress...
3. Or, "In sum, our initial decision placed us in the exhilarating position of having to make a discovery, and of being prepared to deal with substances of an especially precarious constitution."
4. Another defining feature of syntheses from the Woodward (1944-1970s) and early Corey eras was the use of rings to enable reactions to be conducted stereoselectively. Stereocontrol was (and still is) a great deal harder to achieve in loose open chain systems than nice, predictable small rings. I could write a whole other series just on this! See Corey's 1978 erythronolide B synthesis for one of the greatest examples of cyclic templated stereocontrol the world has ever seen. Who'd look at that molecule now and then start from 2,4,6-trimethylphenol? Conversely, Kishi's landmark synthesis of monensin the following year marked the beginning of freedom for chemists from cyclic templates in complex molecule synthesis. Oops, we're digressing again...
5. I plan to write a brief history of the Boc group when I get some time. It's more interesting than you might expect.
6. Something like this.
Not sure why they start from unnatural D-tartaric acid, though. It's quite a bit more expensive than the L-form (and it's not like it matters which you use here).