Cnveniently, you can use an activity measurement to test your expressed protein for correct formation of the disulfide bond. Since there are only 2 cysteines, no incorrect bonds can form. Without reducing agent, there is a good chance the bond will form by itself, as log as the cysteines are not oxidized to cysteic acid. So there are at least 2 approaches you could take:

1. Express the protein, extract without reducing agent, and see whether the protein is active, indicating that the bond has formed.

2. Express the protein, extract with reducing agent, dialyze out the reducing agent, and measure activity.

Thank you for the proposal, Adam! So cytosolic expression should suffice and having the disulfuide formed during lysis (oxidative conditions)? No periplasmic signal sequence needed?

That assumes the protein is made in a soluble form. Since the cytosol is reducing, the disulfide will likely be reduced. I would keep it reduced during purification, then oxidize the purified protein, preferably in a relatively dilute form, then concentrate it afterwards. The idea is to avoid unwanted intermolecular crosslinking.

You migth try E. coli Origami (strain 1 or 2) from Novagen, whith trxB and gor deletion, resulting in an increased oxidation potential in cytoplasm, which allows formation of disulfid bounds.

Good luck!

You may also consider the additional use of mild oxidising agent(s) challenge during expression to tilt the balance of reducing equivalents.