I've done something similar but on a smaller scale. The workflow I came up with was:
First, designate a brain as the primary brain. In this case it would be the one with the IP information. I created the secondary brain by copying all the relevant thoughts from the primary brain to a new brain. In my case it was because there are a lot of additional thoughts in my primary that I didn't want copied over. If most of the thoughts needed to be copied I would have just restored a brainzip to a different location, changed the brain name and assigned it a new GUID (File > Utilities). I did this so the thoughts all have the same GUID, so when I copy thoughts that exist in both databases from the primary to the secondary brain they will overwrite the existing thought and keep the structure of the primary brain, rather than creating a separate thought. I make the majority of my changes or additional thoughts in the primary brain. I try and keep the information flow one way as much as possible, otherwise it gets more complex. When I go through adding or modifying thoughts in the primary brain that need to go to the secondary one I add a tag (I use copy). Then after there are enough changes or new thoughts I activate the tag, select all the children and copy the thoughts. I then open the secondary brain and paste the thoughts. I follow the prompts and overwrite thoughts/tags/links so the changes appear. Then I activate the copy tag in the secondary brain and go through each thought and remove any information I don't want in there. As I check the thoughts, I remove the tag until they are all verified. Finally I go back to my primary brain and remove the tag from all the thoughts and start again. I do have a few thoughts in the secondary brain that are unique, but since I don't need to port them over they never change. When overwriting thoughts, it does keep the same parent (as long as you started by duplicating the brain or copying the thoughts) and if a parent exists in the primary but not the secondary, if it's not copied it won't be added. However, if you copy the parent at a later time, it will be linked correctly in the new database (as far as I remember, I don't do this very often). Hope this helps.