It hasn't escaped my notice that today is not Wednesday, but this is just a follow up post, and you know what they say about gift horses and looking...
As we saw in the inaugural Woodward Wednesday post last week, the second step in Woodward's 1965 synthesis of cephalosporin C was the Boc protection of an amino acid derivative. Having chosen cysteine as the starting material, and performed the known reaction with acetone, the next transformation that the group needed to carry out was this was this:
What I wasn't aware of when I wrote that post was that one of the authors on the Woodward paper, Helmut Vorbrüggen, actually went on to publish a paper on the difficulties of this step and the group's eventual solution more than 40 years later (Synthesis, 2008, 3739-3740). It turns out that the nearby gem-dimethyl group made this protection unexpectedly challenging, and Vorbrüggen provides a good insight into the difficulties Boc protection used to entail, as well as the thought processes that lead to the final choice of reagents.
Normally, when I want to make a carbamate, I reach for the corresponding chloroformate (ROCOCl) or maybe the carbonate, but it turns out that neither is very useful for the introduction of Boc groups. BocCl is woefully unstable and decomposes rapidly if you handle it roughly, by, say, storing it in the fridge or showing it traces of air or water. Conversely, (t-BuO)2CO isn't so much unreactive as inert, remaining unchanged even under quite vigorous conditions such as heating to 150 ºC in concentrated sodium hydroxide solution, which greatly limits its synthetic usefulness. A few papers describing the use of the relatively stable yet reactive BocF also exist, but the main drawback of this reagent is the difficulty associated with its production.
The first thing the group tried was the rather exciting BocN3, a reagent not without its drawbacks that was popular in the 60s for this transformation. Discovered in the late 1950s by Capino and coworkers, BocN3 became a useful alternative to phosgene/t-BuOH for Boc protection. There are a couple of Organic Syntheses procedures for it, which capture some of the excitement accompanying its preparation.
"Caution! Tests conducted by the Eastman Kodak Company have shown that tert-butyl azidoformate, also known as tert-butoxy carbonyl azide and t-BOC azide, is a thermally unstable, shock-sensitive compound (TNT equivalence: 45%)."
I love the fact that this warning notice is found on the same page as the procedure describing the distillation of the compound. Unfortunately, although the reagent works pretty well for amines and hydrazines, it was found to be a bit ponderous for anilines. From the seminal 1957 Capino paper:
"The reaction with aniline is more sluggish, requiring about two weeks standing in pyridine in order to give a 50-60% conversion to the carbanilate. These results suggest that the azide will be generally useful for the acylation of amino derivatives which are at least as basic as aniline."
Woodward did attempt to apply it to the system above, but found it far too unreactive.
Although now the standard for Boc protection, this was only reported a year or two before Woodward began work and according to Vorbrüggen was dismissed because it wasn't commercially available. I think the problem was probably more likely that the seminal 1962 procedure for its preparation by Howe and Morris only proceeded in 5% yield, which might have been inconvenient for the scale on which Woodward and other chemists of that era usually worked. It wasn't until almost a decade later that a practical route to the compound was reported by Tarbell and coworkers who finally described the use of Boc2O as a reagent for protection of amines for the first time in a 1972 PNAS paper.
Earlier work on the penicillins had estabilished that the same system Woodward was working on could be protected with a Cbz group by treating the free acid with BnOCOCl in good yield, presumably via the intermediacy of the mixed anhydride. Intrigued by this precedent, but cautious of the instability of BocCl mentioned earlier, which had been known since the 1940s, Woodward and his team opted to generate the reagent in situ and perform the protection on the free acid. The original procedure for the preparation of this reagent involved the use of sodium t-butoxide, but after some experimentation it was found that use of t-butanol and pyridine was preferable on a large scale. BocCl was formed at -74 ºC, the cysteine derivative added and the mixture finally warmed to 0 ºC and then room temperature before workup. The Vorbrüggen paper contains the full the experimental details of this reaction expanded to give a general procedure for the large scale Boc protection of amino acids with BocCl via what RBW himself referred to as 'internal delivery'. A typically Woodwardian solution, which saddens me that rest of the experimental details for this synthesis were never published.
1. BocF is a stable and distillable liquid, readily prepared from ClCOF (carbonyl chlorofluoride). Unfortunately, ClCOF itself is not commercially available and not so much fun to make. From the 1948 paper:
"We have found that carbonyl chlorofluoride can be prepared readily by shaking a mixture of hydrogen fluoride and phosgene in a copper bomb at approximately 80 ºC and 280 pounds per square inch pressure."
The words 'bomb' and 'hydrogen fluoride' should never appear in the same sentence. You couldn't pay me to work with any amount of neat HF, and it's not like they did this on small scale - the procedure given uses 200 grams of the stuff. I do hope they cleaned what was left of the balance afterwards. But my favourite bit of the paper is from the end:
"The gas has an odour similar to but distinguishable from that of phosgene. It is readily adsorbed by sodium hydroxide or soda lime. It shows no tendency to react with glass."
Delightful. Also formed as a byproduct is fluorophosgene. We can only speculate as to why no description of its smell is given.