Is this due to the barycenters for binary system etc.?
Since the stars etc. can orbit them, they probably have a non-zero mass, which would cause problems with principia.
Simply setting their mass to zero with a patch will change the initial position and velocity of the orbiting bodies since the mass determines the orbital period and the orbital parameters contain two values that depend on the period: [epoch] and [meanAnomlyAtEpoch].
I believe you can set the period to a fix value, if the orbit's mode is set to 0. (all configs I've come accross so far that had an explicit period value also had "mode=0".)
My idea of patch that could make GU principia compatible without needing to reassign parent bodies and recalculating orbital parameters:
