Tuesday, December 11, 2012

Diazomethane and the Arndt-Eistert Homologation

For the past year we have been starting peptidomimetic chemistry up as a new research area in our group. Many chemists believe that peptide chemistry is easy and that peptide chemists aren't "real chemists". However, let me tell you from personal experience that there is absolutely nothing trivial about peptide chemistry. Even short sequences with normal alpha amino acids can be a nightmare to make, troubleshooting is complicated, purification can be a major pain and yields that a small molecule chemist would consider a total fail is generally acceptable in this area of research. Some years ago I was working with a Post Doc that came from Dieter Seebach's lab at ETH. He introduced me to beta amino acids and ever since I have been fascinated by the use of these building blocks in peptidomimetic research. Inspired by the work of Samuel Gellman we are focusing on the use of beta-3 amino acids in combination with alpha amino acids. Consequently, we synthesise beta-3 amino acids to incorporate these in our peptides.
There is a number of ways to make beta-3 amino acids but from personal experience one method stands out as the best route to these molecules: the Arndt-Eistert homologation. In this classic approach an alpha amino acid is converted to a diazoketone followed by the Wolff rearrangement to provide beta-3 amino acids. The Arndt-Eistert homologation basically homologates a carboxylic acid with one methylene group as shown in the scheme below.
The last step, the Wolff rearrangement, is carried out by sonicating the diazoketone in the presence of a silver catalyst (in the dark). Because nitrogen is evolved during the course of the reaction we normally have an empty balloon fitted on the flask to avoid pressure build up. I rather like the feature that the balloon slowly gets inflated during the course of the reaction as shown in the picture below.
Silver catalysed Wolff rearragement in a sonicator. Left t = 0 hr; Right t = 2 hr.
However, as you may have noticed there is a down side to the Arndt-Eistert homologation: diazomethane! The reagent has a fearsome reputation and I have heard of a couple of guys who have managed to blow themselves up and gone deaf in the process. Allegedly, one chemist at our department even managed to set fire to himself! This was a long time a go when less attention was being paid to laboratory safety and the accidents were due to sloppiness and improper handling of diazomethane. If you are careful and use the correct glassware (with clear seal joints) there is (almost) nothing to worry about. We have purchased the setup shown on the picture below. This is a very nice diazomethane still consisting of only three pieces that will produce up to 40 mmol of diazomethane in approximately one hour. We only use hot water as the heating source and keep everything behind a blast shield just in case. Diazomethan is generated from Diazald  as shown in the scheme below and used immediately. The procedure it quite simple. In the separatory funnel you place a solution of Diazald in ether this is added dropwise to a heated mixture of aqueous potassium hydroxide, ether and a high boiling alcohol [commonly 2-(2-ethoxy-ethoxy)ethanol]. Diazald reacts with the base to produce diazomethane that is distilled with ether to the receiving flask.
Notice that diazomethane is always handled in solution. The neat stuff is known to explode unpredictably so don't even think about doing that. Because of the way that diazomethane is produced it is hard to add an exact number of equivalents to a reaction. For the synthesis of diazoketones we simply go for an excess of diazomethane (approximately 2-3 equivalents based on a 70% yield of diazomethane). We commonly distill the diazomethan directly into the reaction flask to minimise handling. For the synthesis of beta-3 amino acids the alpha amino acid is first transformed into a mixed anhydride which is exposed directly to an excess of diazomethane.
Diazoinsane clear seal distillation kit purchased from Sigma-Aldrich.
Unlike diazomethane, Diazald is reasonably stable and easy to handle yellow solid. Unfortunately, Diazald has obtained a rather bad reputation despite being relatively safe to deal with as long as you don't eat it, set fire to it, beat it with a hammer or something similarly stupid. Consequently, it can be rather hard to get hold of. When I worked in Australia it was particularly problematic as it can only be shipped by road and isn't produced in the country! Here in Denmark we get it from Germany but it does take a while because they don't send it with the regular shipments so you have to plan a bit ahead.
If you think that playing around with beta-3 amino acids could be fun I can recommend the company Anand Chem based in Slovakia. They produce almost all beta-3 amino acids with the proteinogenic side chains of excellent quality at a highly competitive price. Depending on what they have in stock you may have to wait a couple of weeks for the stuff but it is worth the wait considering the quality and the price. D!


Brandon said...


I've always thought using aspartate to make β3-peptide analogues was rather clever, as that sidesteps the need for diazomethane.

(ex. http://www.sciencedirect.com/science/article/pii/S1570963908000113)

Anonymous said...

Having done a few esterifications with diazomethane, I have to say I just love this reagent! You titrate your acid until the solution stays yellow (does not work if you have colored substance or impurities) and destroy the excess with AcOH.

There are procedures that do not require destillation, I used N-methyl-N-nitroso urea (add to ice cold aqueous KOH stirred with ether, the diazomethane goes into the ether and you simply remove the ether layer with a pipette).


Daniel Sejer said...

@ChristianPFC, Cool that your are using diazomethane but your really should not pipette the solution. The edge of a pipette is enough to make it go off. It's only a question of time before it happens. Instead you should use a clear seal joint sep. funnel. Remove the aqueous layer and then add the ether solution directly from the sep. funnel, please! I know of pasteur pipette induced diazomethan explosions so it is a real risk and not a theoretical one. Moreover, for the synthesis of methylesters you could consider using commercially available TMS-diazomethane that works really well for this transformation. D!

Anonymous said...

Do you have experience in using TMSCHN2 (conditions?) for diazoketone synthesis from the chloroformate activated carboxylic acids? I would love to plan an Arndt-Eistert homologation, and would like to use TMSCHN2 if possible. (just to be sure on the safety part :-) )
Some years ago I have used TMSCHN2 to make diazoketones from acid chlorides (works well!), but the current molecules won't allow such activation...

Jurgen C

Daniel Sejer said...

@Jurgen, Dammit for some reason Blogger stopped notifying me of comments so I'm only seeing your question now. We have very poor experience with TMS-diazomethane. Generally, we find that if it works it is very messy. The only smooth reaction with this stuff we had was in making methylesters. So we gave up on the stuff years ago and just use diazomethane instead. D!