Amino acid side chains participate in a number of reactions that result in peptide modification. Some of these reactions are of fundamental significance and play an important role in folding processes. Disulfide formation by oxidation of two proximal cysteine residues is one such case. Apart from its relevance in protein biochemistry, this reversible process also ensures structural integrity of fairly small cyclic peptide systems. Ziconotide is an example of a disulphide-rich analgesic agent derived from a conotoxin.
There are also amino acid reactions one would want to avoid. Today I want to talk about threonine. I always viewed this amino acid with caution (and serine too). If you pay close attention, you will note that threonine looks like the product of an aldol reaction, in which the amide functionality acts as the enolate component, whereas acetaldehyde corresponds to the aldehyde partner. Whenever you see an adol process, you can be sure that its microscopic reverse is feasible. Writing in Organic Process Research and Development, a group from Astellas report a retro-aldol by-product in their synthesis of a cyclic peptide drug candidate. I think this is an important warning for those who work with serine and threonine containing peptides. Be careful with the base you use!
amphoteros 
I similar vein, phosphorylated threonines and serines are incompatible with Fmoc chemistry and eliminate to dehydronorvaline/dehydroalanine easily, even with a trace amount of base. So they have to be phosphorylated post-solid phase.
Yes, this is true. A good way to make the corresponding dehydro species when needed.
I remember, a long time ago, I was working with an Italian chemist named Fausto. He had a monstrous macrocycle peptide-based antibiotics (that was quite toxic and insoluble) available on 100g scale by fermentation, his group managed to get the material cleanly chopped in two pieces enzymatically and he was then trying re-build the excised part of the ring in a combichem fashion (to see if he could get better analogues). His biggest trouble was with the phenylserine residue that was essential for antimicrobial activity – phenylserine underwent facile retro-aldol even with few % of piperidine, so he ended up stuck with Alloc/silyl protection all the way. Even TBAF deprotection was difficult – unbuffered TBAF turned out too basic for phenylserine. Fausto finished his libraries of analogs though, for he was a very stubborn guy.
That’s really cool, thanks for this note – we really need to be careful with retro-aldol. I think it is important for the students to know these facts because every now and then we come across some unidentified LCMS peaks with strange masses. The more we know about these curious cases, the better. P.S. Did you call your friend “Dr Faustus”?
He was a tall, dark, lean and fierce fellow, with a prominent nose, flashing eyes and with disquieting, harsh inflection of voice – and his speech was animated by sudden hand gestures. And there was also a bit of a goat-like body odor about him. I kid you not. We actually called him Mephisto…
Classic!
I would have titled this one “Darkness on the Edge of Town”….
This is a great one! Ha…
This is very interesting…thanks for the post. I had never thought of threonine (and serine) like that before.
You mention retro-aldol under basic conditions. I was wondering if there are any examples of elimination of water under acidic conditions to form the a,b-unsaturated carbonyl. That sort of intermediate could cause all sorts of trouble.
Also, for either reaction, I imagine that even under relatively mild conditions an opportunistically-positioned intramolecular H-bond to the -OH might also catalyze the reaction. I don’t deal much in the way of cyclic peptides, but I suspect it’s difficult to account for this a priori.
Yes, there are indeed. I think this is in the literature. It is also definitely true that hydrogen bonds are going to contribute if the circumstances are right.