Thank you for all this information, and especially Paul for having a good search query - I had got nothing useful from any I tried at PDBe or RSCB.
No idea why it has been formed - the fold is identical to a high resolution example (1.3A) where the disulphide links the N & C termini within a single molecule.
Except for 2-3 residues which cause the domain swap the CA rmsd is < 2. It must be a crystallisation artefact I think,. or my mistake in interpretation... This new data is lower resolution but after omitting the hings region the density is pretty reliable..