Today I will talk about microcystin and its mode of action. This molecule is one of my all-time favourites because of its unique reactivity. Below you can see a view I created using 1FJM entry from the Protein Databank. This picture shows the electrophilic warhead of microcystin and identifies the surface-exposed nucleophilic Cys-273 residue of PP1 phosphatase. This cysteine irreversibly interacts with the electrophilic acrylamide portion of microcystin.
My lab has been after some complex peptide macrocycles equipped with electrophilic aziridine residues. We do not yet have any significant stories to tell in our efforts to covalently inhibit cysteine-bearing protein targets, but we do have the methodological makings of an interesting approach in collaboration with Ben Cravatt of Scripps. So far our molecules seem to be inert against cysteines, which is somewhat of a surprise, yet gives us confidence that we might eventually find something really selective. Back to microcystin: this nanomolar phosphatase inhibitor is a nasty beast that was isolated from cyanobacteria. The corresponding blue-green algae contaminate drinking water and have long been known to be the cause of animal deaths.
amphoteros 
Your aziridines – are they N-acylated? Non-acylated aziridines are rather weak electrophiles, even when protonated (but N-sulfonylation or acylation boosts the reactivity). I was working with 1-azabicyclo[1.1.0]butane, I am still amazed how remarkably stable this strained aziridine is (it is isolated by distillation from aqueous reaction mix, at atmospheric pressure)
That is an interesting insight about the [1.1.0] system, didn’t know about the distillation trick. Ours are N-acyl aziridines (cyclic), which is why it is still a bit of a surprise that they can be stable to biological nucleophiles.
I am curious – if you have for example aziridine-2-carboxamide residue in the molecule (Nitrogen is in alpha, beta position to carbonyl), is there any good way of eliminating the aziridine in beta position to dehydroalanine, on a finished and deprotected cyclic peptide? There are probably easier ways of making dehydroalanine in the molecule, i.e. by derivatization of serine, but that would require separate sequence. I am thinking, since you are putting aziridines in your peptidesby default, if it would be possible for you to get two molecules from one…
Yes indeed, we have actually explored this recently. Check out: http://pubs.acs.org/doi/pdf/10.1021/ja412256f. You can definitely make dehydroalanines.