Pseudocontact shifts (pcs) and paramagnetic residual dipolar couplings (rdc) provide structural information that can be used to assess the adequacy of a crystallographic structure to represent the solution structure of a protein. This can be done by attaching a lanthanide binding tag to the protein. There are cases in which only local rearrangements are sufficient to match the NMR data and cases where significant secondary structure or domain rearrangements from the solid state to the solution state are needed. We show that the two cases are easily distinguishable. Whereas the use of solution restraints in the latter case is described in the literature, here we deal with how to obtain a better model of the solution structure in a case (the catalytic domain of the matrix metalloproteinase MMP-1) of the former class.