Activity restored by mixing recombinant domains?

[Lloyd Graham]

I began this thread by writing:
 > I'm looking for any precedents where the domains of a monofunctional
 > enzyme have been expressed separately as inactive proteins, but when mixed
 > together they associate to form a complex with significant enzymatic
 > activity. If anybody can provide instances where purified inactive domains
 > complement in vitro, I would be very grateful. 

Thanks for all the replies to date, and here is some feedback on them:

* RNAP large subunits... Yes, this is a good one.

* alpha peptide of beta-galactosidase... Sorry, the alpha peptide is too
short (39-50 aa) to be a true domain, and the 3D structure confirms that
it doesn't adopt a globular fold in the intact protein.

* ATCase from E. coli... Would be a good one if the fragments had
corresponded to the two domains but (as the contributor acknowledged) this
was not the case.

There must *surely* be several clean-cut examples of this entirely
expected phenomenon in the literature, but where are they?

Thanks again,

Lloyd.