I would use a fresh solution of 0.1 M sodium bicarbonate, which is mildly alkaline and does not interfere with the succinimidyl ester reaction.

In addition to Adam’s suggestion, I would recommend that you keep urea in the mixture, in order to keep the protein soluble (most likely, this is why it precipitates in PBS).

Thanks Adam and Pierre, this is very helpful.

I would be tempted to use urea, but I'm concerned that biotinylation of unfolded areas will impede protein refolding, as I will need the protein in its native state after the reaction for the purpose of follow-up experiments.

I'll be looking into other biotin derivatives.

I'm afraid you are trying to do the second step before the first. If you need the native protein and are concerned that biotinylation of unfolded regions might interfere with refolding, try to find conditions to keep the protein in solution (hopefully in the folded stated, note that solubility does not automatically indicate natively folded protein, you could have soluble aggregates with all sorts of intermediate folds). After that I would try biotinylation.

If you want to biotinylate anyway in the unfolded sample, I would recommend changing the buffer from 7M urea to 6 M guanidine (in this unfolded state it doesn't really matter what buffer you use, PBS or Carbonate as suggested above is equally working as long as you keep the pH in the alkaline range at >pH7.5-8.5).

Note that Urea could contain trace amounts of ammonia (which you can smell) that will inhibit the biotinylation of your protein. Guanidine usually contains no or much less free amines.

Last not least, you could use a biotinylation tag like Avitag, if you can produce a new version of the protein which could be biotinylated during expression and hopefully refolded afterward. Maybe adjusting expression conditions would give better yields of folded soluble protein.

Would a thiol reactive chemistry (i.e. Biotin-Maleimide) be a suitable alternative? Then traces of amines would not interfere.